Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2020 May 29;15(5):e0232778. doi: 10.1371/journal.pone.0232778

Method for extraction of airborne LiDAR point cloud buildings based on segmentation

Maohua Liu 1,2,#, Yue Shao 2,#, Ruren Li 2,*, Yan Wang 2,‡, Xiubo Sun 3,‡, Jingkuan Wang 1,*, Yingchun You 2,‡
Editor: Tayyab Ikram Shah4
PMCID: PMC7259685  PMID: 32469887

Abstract

The LiDAR technology is a means of urban 3D modeling in recent years, and the extraction of buildings is a key step in urban 3D modeling. In view of the complexity of most airborne LiDAR building point cloud extraction algorithms that need to combine multiple feature parameters, this study proposes a building point cloud extraction method based on the combination of the Point Cloud Library (PCL) region growth segmentation and the histogram. The filtered LiDAR point cloud is segmented by using the PCL region growth method, and then the local normal vector and direction cosine are calculated for each cluster after segmentation. Finally, the histogram is generated to effectively separate the building point cloud from the non-building.Two sets of airborne LiDAR data in the south and west parts of Tokushima, Japan, are used to test the feasibility of the proposed method. The results are compared with those of the commercial software TerraSolid and the K-means algorithm. Results show that the proposed extraction algorithm has lower type I and II errors and better extraction effect than that of the TerraSolid and the K-means algorithm.

Introduction

LiDAR (Light Detection and Ranging) is an active remote sensing technology that can provide dense sets of point cloud of a scanned target. This technology has been tagged as an emerging practical technique for 3D modeling of smart cities in recent year due to its non-invasive nature, high precision, high resolution, and rapid and flexible data acquisition. The extraction of buildings from the point cloud is a prerequisite in urban 3D modeling. Current building extraction methods can be divided into two categories. One category is to directly extract the building point cloud after classifying the LiDAR data according to some features. Rottensteiner and Briese [1] combined height difference thresholds, point cloud depths, and image texture features to extract buildings. Zhang et al. [2] used filtering algorithms to separate non-ground points first and then non-ground points and corresponding near infrared. The images were overlapped, and the NDVI value was calculated for each non-ground point. Most vegetation points were eliminated according to the NDVI value, and the building was extracted by the European clustering algorithm based on multi-echo and area features. Cheng et al. [3] used a mathematical iterative algorithm based on reverse iteration to extract building point cloud automatically. Cao et al. [4] used the region growing algorithm of gradient threshold and area feature to extract building point cloud.

The other category is the object-oriented classification method. First, the data points are divided into multiple objects, and the divided objects are classified according to the features for extracting building point cloud. A common regional growth segmentation algorithm divides a point cloud into multiple homogeneous regions. However, some of the homogeneous regions obtained will contain other landmarks when the vegetation is adjacent to the roof of the building. Niemeyer et al. [5] provided a powerful probabilistic framework for classification based on machine learning conditional random field model and realized the point cloud extraction of buildings using random forest classifiers. Richter et al. [6] segmented the point cloud by a smooth-constrained region growing segmentation algorithm, and then used the multichannel iterative algorithm (in combination with the height difference threshold and area features) to extract building point cloud. Awrangjeb and Fraser [7] combined area, height difference, spatial position, point cloud coplanarity, and other features to extract the point cloud of buildings. Zhang et al. [8] used a region growing algorithm that combines topological, geometric, echo, and radiation properties to segment point cloud, and utilized connected component analysis and support vector machines to extract building point cloud. Li Liang [9] adopted the layer-near method to extract building point cloud. First, the region growing algorithm was used to segment the point cloud. Then, the connected component analysis was utilized to carry out the European clustering on the initial building point cloud. Finally, buildings and vegetation were distinguished by combining several features, such as area and height difference.

However, the building extraction methods based on these two ideas are complicated and need to combine multiple feature parameters. This study proposes a new building extraction method. Since the building and vegetation areas are the main components of the non-ground point elements, the normal vector direction of the building roof is basically the same, and the vegetation surface normal vector varies greatly. According to the difference of normal vector characteristics of building roof and vegetation surface, PCL based Region growth algorithm is used to segment the 3D point cloud from non-ground points. The Region growth algorithm is a curvature-based clustering algorithm, and points with curvature values within a certain threshold range are determined as the same type of features. A novel histogram method is used to distinguish buildings from non-buildings, so as to accurately extract building point cloud data. Its principle is to calculate the local normal vector and the direction cosine of the normal vector of each clustered cluster surface after clustering and generate a histogram, and extract the building point cloud based on the distribution characteristics of the histogram. Compared with currently available methods for building extraction, the proposed method is simple, efficient, and accurate. In addition, the proposed method does not need to combine any characteristic parameters and shows superiority in filtering out most of the high vegetation points and effectively extracting the building point cloud.

Data structure and data processing methods

Data structure

Each point of the onboard LiDAR point cloud is stored in a file in the same organizational mode because of the sensor scanning mechanism. Points in the LiDAR data are usually spatially discrete and unorganized and are unsuitable for representation on the grid due to the geometry of the scanning device and the nature of the target interaction. Coding of each point cloud data requires all 3D coordinate values (X, Y, and Z), and the random point cloud becomes difficult to use when performing such operations as searching or performing interpolation operations. Thus, this study uses a K-D tree data structure. K-D tree is a way to organize point sets by using geometric index. It is a multi-dimensional generalization of spatial binary tree. When K-D tree method is used to divide point cloud data, the data on both sides of the node are usually divided into two parts. The number needs to be consistent, and the division ends when the number of nodes is less than the set threshold. The use of K-D tree method to organize point cloud data can greatly improve the efficiency of data search [10].

CSF algorithm

Point cloud filtering separates the set of terrain surface points from the airborne LiDAR point cloud data. Its essence is the initial classification of point cloud data. The original airborne LiDAR point cloud is divided into ground and non-ground points, which is the key step of subsequent point cloud data processing [11]. The CSF [12,13] method is different from the traditional point cloud data filtering algorithm and it is a 3D computer graphic algorithm based on cloth simulation. The principle of the algorithm is that a piece of virtual cloth sinks on the surface of the terrain due to gravity. The cloth is assumed to be sufficiently soft enough to adhere to the surface, and the shape of the cloth is the digital surface model. The terrain is flipped over. If the cloth has rigidness and is defined by rigidness, then the final shape of the cloth is the digital elevation model. The cloth is essentially a mass spring model. By analyzing the interaction between the cloth node and the corresponding point cloud data, the shape of the cloth is finally determined to realize data filtering and obtain ground and non-ground points. The specific process of the filtering method is as follows:

(1) Invert the original point cloud data;

(2) Place the cloth above the original laser point cloud to be processed and set cloth simulation parameters;

(3) Perform grid division on the original point cloud data, and then search for the nearest laser point matching each cloth grid point to determine whether it is a movable point. If it cannot be moved, calculate the elevation of this point as a collision point.

(4) If it is a movable point, calculate the position of the point under external force and calculate the height difference between the point and the collision point. If the height difference is less than or equal to 0, the elevation of the point is set as the collision point elevation value, and it is set as the immovable point;

(5) Calculate the displacement that needs to be adjusted for each movable cloth grid point under the internal force;

(6) Repeat (4) and (5) until the number of iterations is not less than the maximum number of iterations, and the iteration ends;

(7) Determine whether the height difference between the cloth grid point and the original laser point cloud matching it is less than the threshold. If it is less than the height difference threshold, it is a ground point, otherwise it is a non-ground point.

PCL-based region growing segmentation algorithm

PCL is an open source point cloud library that involves the acquisition, filtering, segmentation, registration, and feature extraction of point clouds. This study is based on the PCL to implement the segmentation algorithm for region growing. The segmentation algorithm is applicable to the point cloud data with uniform features. By setting constraints and combining the fusion requirements of the segmentation data, different objects are segmented from the scene by using the features of the objects in the scene. In the PCL, the region growing segmentation class pcl::Region Growing is adopted to implement the region growing segmentation algorithm. The goal of the algorithm is to merge adjacent points satisfying the smoothing constraint, and each cluster point set belongs to the same smooth surface. The algorithm is a segmentation based on the angle difference of the normal angle. By comparing the angle between the seed point and its neighbor point, the point cloud smaller than the smoothing threshold is used as the part of the same smooth surface [14]. The points are sorted from small to large according to the curvature value of the point, and the initial seed point is selected as the point with the smallest curvature to start growing. Growing from the smoothest region reduces the total number of segments, which in turn increases segmentation efficiency. The region where the initial seed point is located is the smoothest region. The point cloud in the adjacent range of the seed point is searched, and the curvature between the seed point and each adjacent point is calculated. If the curvature is less than the set threshold, then the neighborhood point is regarded as the homogeneous region of the seed point, and the current seed point is deleted. The aforementioned steps are performed cyclically until the points in the point set are all processed, and the growth of the region ends.

Histogram-based building detection method

The building cluster must be separated from the tall vegetation to extract the buildings in the scene. A simple method is to extract the building point cloud through LiDAR data visualization software manually, but this method is inefficient and unsuitable for large areas. Existing building extraction algorithms need to combine multiple characteristic parameters and the calculation process is complex. For example, it is necessary to combine characteristic parameters such as the height difference threshold, the point cloud depth, image texture, echo, and area. In this study, a simple and efficient histogram building detection method was proposed. The local normal vector of point cloud computing in each cluster was obtained after region growth segmentation and the direction cosine in the X, Y, and Z directions were calculated. The number of point cloud with cosine value was counted and the histogram was generated. According to the difference of histogram distribution between vegetation and building clusters, the architectural complex and non-building cluster are separated.

Commercial software TerraSolid detection

The filtering function of TerraSolid software is based on the irregular triangulation encryption algorithm proposed by Axelsson [15]. The original point cloud is filtered to separate non-ground points, including buildings, vegetation, artificial facilities, and vehicles. The ground point elevation is used as a reference to classify non-ground points by comparing the elevation information of the point cloud and setting a reasonable elevation threshold [16]. Point clouds from 0 m to 0.5 m above ground are generally classified as low vegetation, point clouds from 0.5 m to 2 m above ground are classified as medium vegetation, and point clouds above 2 m are classified as tall vegetation and buildings.

Filter experiment

This study selects two datasets in the south and west parts of Tokushima, Japan. Dataset 1 has a high terrain in the west and a low terrain in the east. The main ground features include buildings, tall trees, artificial facilities, low and medium vegetation, bridges, and rivers. The buildings are scattered, and the surrounding vegetation is sparse. A total of 2663447 laser foot points are present in the survey area. The point cloud density is 31 points/m2.The minimum and maximum elevations are +100.80 m and +308.09 m, respectively. The original point cloud of dataset 1 colored by elevation is shown in Fig 1(A). In dataset 2, the terrain in the survey area is low on the west and high on the east. A few types of ground objects, mainly buildings and vegetation, are concentrated and surrounded by green vegetation. A total of 2064906 laser foot points are present in the survey area. The point cloud density is 20 points/ m2. The minimum and maximum elevations are +144.22 m and +355.19 m, respectively. Each color in the color legend of Figs 1 and 2 corresponds to the corresponding elevation value. The original point cloud of dataset 2 colored by elevation is shown in Fig 2(A). Filtering experiments on the point cloud data in datasets 1 and 2 are performed by the CSF algorithm to obtain non-ground points. The non-ground points of datasets 1 and 2 are shown in Figs 1(B) and 2(B), respectively. The number of points in datasets 1 and 2 before and after filtering is counted. The statistical results are shown in Table 1.

Fig 1. Comparison of dataset 1 before and after filtering.

Fig 1

Open in a new tab

1(A) Raw data before filtering. 1(B) Non-ground points after filtering.

Fig 2. Comparison of dataset 2 before and after filtering.

Fig 2

Open in a new tab

2(A) Raw data before filtering. 2(B) Non-ground points after filtering.

Table 1. Statistics of cloud points before and after filtering.

Parameter Dataset 1 Dataset 2
Total points 2663447 2064906
Point cloud density (P/m2) 31 20
Ground point 1476770 896180
Non-ground point 1186519 1708640
Noise point 158 86
Open in a new tab

Results and analysis

From the above-mentioned filtering experiments, the non-ground points of datasets 1 and 2 are obtained. The filtered dataset is organized by a K-D tree, and the region growing algorithm is implemented on the basis of an open-source PCL. The segmentation algorithm divides these points into different clusters. The region growing algorithm needs to set the minimum and maximum numbers of clustering points for datasets 1 and 2. According to the prior knowledge of the data, the minimum number of clustering points in dataset 1 is set to 20 and the maximum number of clustering points is set to 7200. The minimum number of clustering points in dataset 2 is set to 40 and the maximum number of clustering points is set to 5000. The normal vector of each point cloud from the obtained clusters is calculated, and the normal vector and the cosine of X, Y, and Z directions are calculated. The histograms is generated, and the building and non-building point groups are separated according to the change characteristics of the histogram. This study enumerates the normal vector cosine of some sampling points. The histograms generated by datasets 1 and 2 are shown in Figs 4 and 5, respectively.

Fig 4. Non-building and building normal histograms for dataset 2.

Fig 4

Open in a new tab

(a) X-tree normal histogram (b) Y-tree normal histogram (c) Z-tree normal histogram (d) X-building normal histogram (e) Y-building normal histogram (f) Z-building normal histogram.

Fig 5. Building extraction results.

Fig 5

Open in a new tab

5(A) Dataset 1 building extraction results. 5(B) Dataset 2 building extraction results.

As shown in the histograms in Figs 3 and 4, the histogram of normal lines on the tree surface is scattered with many peaks either for datasets 1 or 2. Moreover, the normal cosine value of the maximum frequency in the X and Y directions tends to be 0, and the normal cosine value of the maximum frequency in the Z direction tends to be 1. However, the histogram of normal building surface shows few peaks and has a concentrated distribution, which can distinguish buildings from non-buildings.

Fig 3. Non-building and building normal histograms for dataset 1.

Fig 3

Open in a new tab

(a) X-tree normal histogram (b) Y-tree normal histogram (c) Z-tree normal histogram (d) X-building normal histogram (e) Y-building normal histogram (f) Z-building normal histogram.

Fig 5 shows the building point cloud extracted from datasets 1 and 2 by using the proposed method. A total of 65 and 105 buildings are extracted from datasets 1 and 2, respectively.

The marked area in the figure refers to the point cloud where non-buildings are mistakenly divided into buildings or incomplete point clouds are extracted. Notably, two large vehicles are mistakenly divided into buildings in dataset 1, which may be caused by that the top of the car is similar to the surface of the building and has a relatively close direction cosine value. Many data are mistakenly extracted in dataset 2, among which three are incomplete and caused by tall vegetation covering the top of buildings. In the five other places, tall vegetation adjacent to buildings is mistakenly divided into buildings. The reason is that the surface of these vegetation areas is relatively smooth. Most of the external contours are also regular, which is similar to the direction cosine of the buildings.

The building extraction based on TerraSolid software needs to set the minimum size of the building and the elevation threshold from the ground point. The minimum size of buildings in dataset 1 is 20, and the minimum size of buildings in dataset 2 is 40. Therefore, the parameters in dataset 1 are set to 20 and 2.2 m, and those in dataset 2 are 40 and 2.2 m. The building extraction results for datasets 1 and 2 are shown in Fig 6. A total of 74 and 107 buildings are extracted from datasets 1 and 2, respectively. The marked areas in the figure are all points where the vegetation points are calculated as buildings. Notably, the TerraSolid software is not ideal for extracting buildings from datasets 1 and 2.

Fig 6. TerraSolid building extraction results.

Fig 6

Open in a new tab

6(A) Dataset 1 extraction results. 6(B) Dataset 2 extraction results.

In order to objectively evaluate the method of this paper, this paper also uses the K-means clustering algorithm based on Python to extract buildings. The K-means clustering algorithm is an iterative clustering analysis algorithm. The step is to randomly select K objects as the initial cluster center, then calculate the distance between each object and each seed cluster center, and assign each object to the cluster center closest to it. The cluster centers and the objects assigned to them represent a cluster. For each sample assigned, the clustering center of the cluster is recalculated based on the existing objects in the cluster. This process is repeated until a certain termination condition is met. In this paper, a total of 4 cluster classifications are set up, and the classification is based on Z value and echo intensity. The extraction results of buildings in datasets 1 and 2 are shown in Fig 7. Among them, 56 buildings are extracted in datasets 1, and 114 buildings are extracted in datasets 2. It can be seen from the figure that the buildings in the datasets 1 are incompletely extracted and only the outer contours are extracted for most of the buildings. There are many cases where non-buildings are misclassified into buildings in dataset 2. The finishing effect is not satisfactory.

Fig 7. K-means building extraction results.

Fig 7

Open in a new tab

7(A) Dataset 1 extraction results. 7(B) Dataset 2 extraction results.

In order to verify the extraction accuracy of this method, quantitative analysis was performed on the method in this paper, the method based on TerraSolid software extraction, and the K-Means method. The reference for accuracy evaluation is the results of manually extracted buildings. Three evaluation variables are defined as type I errors, type II errors and type Total errors. The type I error indicates the probability of misclassifying a building point cloud into a non-building point cloud. The type II error indicates the probability of misclassifying non-building point clouds into building point clouds. Define N as the actual number of point clouds in the experimental area, N1 as the number of actual building point clouds in the experimental area, N2 as the number of building point clouds extracted through experiments, and n as the number of point clouds for correctly judging buildings. The calculation formulas of the type I error, the type II error and the type Total error are shown in formula (1), formula (2) and formula (3), and the specific calculation results are shown in Table 2.

Table 2. Error analysis of point cloud extraction method for buildings.

Error type Type Ⅰ Type Ⅱ Type Total
Method Dataset 1 Dataset 2 Dataset 1 Dataset 2 Dataset 1 Dataset 2
Proposed method 5.68% 8.56% 0.36% 1.18% 3.66% 5.31%
TerraSolid 7.64% 8.89% 1.27% 1.21% 5.46% 5.89%
K-means 10.42% 15.39% 13.67 16.79 10.36% 14.71%
Open in a new tab
TypeI=N1−nN1 (1)
TypeII=N2−nN−N1 (2)
TypeTotal=N1+N2−2nN (3)

It can be seen that the type I errors, type II errors, and type Total errors extracted by datasets 1 and 2 using TerraSolid software and K-means algorithm are higher than that of the method in this paper. The K-means algorithm has the largest type I error, type II error and type Total error in both datasets 1 and 2. Generally, the method of this paper has the highest accuracy in extracting buildings.

Conclusions

This study proposes a building point cloud extraction method based on PCL region growth algorithm and histogram. First, the segmentation algorithm of Region growth is used to segment non-ground points into different clusters. Then, the normal vector of point cloud in each cluster and the normal vector cosine of the X, Y, and Z directions are calculated and the histogram is generated. Experiments show that the algorithm is simple in principle and does not need to combine any characteristic parameters. Additionally, the proposed algorithm can filter out most of the high vegetation points and effectively extract the building point cloud to meet the precision requirements of building point cloud extraction. The method proposed in this paper is compared with TerraSolid, a commercial software, and the open-source algorithm K-means. For the two datasets, the mean type I error extracted by the proposed method is 7.12%, the mean type II error is 0.77%, and the total error is 4.99%. The mean values for the three types of errors for using the TerraSolid are is 8.27%, 1.24%, and 5.68%, whereas those values for the K-means algorithm is 12.91%, 15.23%, and 12.54%. In summary, the building extraction results based on the algorithm proposed in this paper have less error and higher accuracy.

Acknowledgments

Thanks to the data provided by the Shenyang Geological Survey of the China Geological Survey and SHEN YANG KIMOTO INDUSTRIES CO.,LTD.

Data Availability

The data in the article can be downloaded at the following DOI:10.6084/m9.figshare.12227945.

Funding Statement

National Natural Science Foundation of China (Award Number: 51774204). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.Rottensteiner F.; Briese C. A new method for building extraction in urban areas from high-resolution LiDAR data. ISPRS Archives. 2012, 34(3/A), 295–301. [Google Scholar]
  • 2.Zhang W.M.; Wang H.T; Chen Y.M.; Yan K.; Chen M. 3D Building roof modeling by optimizing primitive’s parameters using constraints from Lidar data and aerial imagery. Remote Sens. 2014, 6(9), 8107–8133. 10.3390/rs6098107 [DOI] [Google Scholar]
  • 3.Cheng L.; Zhao W.; Zhang W.; Liu Y.X.; Li M.C. Building region derivation from LiDAR data using a reversed iterative mathematic morphological algorithm. Optics Communications. 2013,286, 244–250. 10.1016/j.optcom.2012.08.028 [DOI] [Google Scholar]
  • 4.Cao H.; Li Y.Q.; Niu L.B.; Mao J.; Yan Y.Y. Point cloud extraction of buildings based on airborne LiDAR data. Journal of Henan University of Urban Construction. 2014, 23(1): 59–62. (in Chinese) [Google Scholar]
  • 5.Niemeyer, J.; Rottensteiner, F.; Soergel, U. Classification of urban LiDAR data using conditional random field and random forests. Urban Remote Sensing Event, 2013 Joint. Sao Paulo: IEEE, Brazil, 2013. 10.1109/JURSE.2013.6550685
  • 6.Richter R.; Behrens M.; Döllner J. Object class segmentation of massive 3D point clouds of urban areas using point cloud topology. International Journal of Remote Sensing. 2013, 34(23), 8408–8424. 10.1080/01431161.2013.838710 [DOI] [Google Scholar]
  • 7.Awrangjeb M.; Fraser C. Automatic segmentation of raw LiDAR data for extraction of building roofs. Remote Sens. 2014, 6(5), 3716–3751. 10.3390/rs6053716 [DOI] [Google Scholar]
  • 8.Zhang J.X.; Lin X.G.; Ning X.G. SVM-based classification of segmented airborne LiDAR point clouds in urban areas. Remote Sens. 2013, 5(8), 3749–3775. 10.3390/rs5083749 [DOI] [Google Scholar]
  • 9.Li, L. Fast 3D modeling of buildings based on airborne LiDAR data. Master, University of Electronic Science and Technology of China, Chengdu, China. May 2016. (in Chinese)
  • 10.Yang, P. Building extraction and modeling based on airborne LiDAR point cloud data. Master, Liaoning University of Engineering and Technology, Fuxin, China, Dec. 2012. (in Chinese)
  • 11.Yang, N. Research on classification and building reconstruction technology of airborne LiDAR point cloud based on segmentation.Doctor, Information Engineering University, Zhengzhou, China, April 2014. (in Chinese)
  • 12.Zhang W.M.; Qi J.; Wan P.; Wang H.; Xie D.; Wang X. Y, et al. An Easy-to-Use Airborne LiDAR Data Filtering Method Based on Cloth Simulation. Remote Sens. 2016, 8(6), 501 10.3390/rs8060501 [DOI] [Google Scholar]
  • 13.Zhang W.M; Chen Y.M.; Wang H.T; Chen M.; Wang X.Y.; Yan G.J. Efficient registration of terrestrial LiDAR scans using a coarse-to-fine strategy for forestry applications. Agricultural and Forest Meteorology. 2016,225, 8–23. 10.1016/j.agrformet.2016.05.005 [DOI] [Google Scholar]
  • 14.Zhou B.N.; Min H.S.; Kang Y.W. Research on 3D Point Cloud Segmentation Algorithm in PCL Environment. Microelectronics Computer.2018, 35(6), 101–105. (in Chinese) [Google Scholar]
  • 15.Axelsson P. DEM generation from laser scanner data using adaptive TIN models. International Archives of Photogrammetry and Remote Sensing. 2000, 33, 110–117. [Google Scholar]
  • 16.Zhang, S.Y. Research on Data Filtering and Building Extraction of Airborne LiDAR Data. Master, Chang'an University, Xi'an, China, April 2018. (in Chinese)

Decision Letter 0

Tayyab Ikram Shah

10 Jan 2020

PONE-D-19-29944

Method for extraction of airborne LiDAR point cloud buildings based on segmentation

PLOS ONE

Dear Mr. LI,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Feb 24 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Tayyab Ikram Shah, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please amend your Data availability statement to provide direct links/URLs or contact details to where other researchers can obtain the datasets. Please also provide in your methods section enough detail about the datasets so that another researcher can replicate the datasets.

3.In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

4. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

5. We note that Figures 2, 3, 6 and 7 in your submission contain satellite images which may be copyrighted.

All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (a) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (b) remove the figures from your submission:

a. You may seek permission from the original copyright holder of Figures 2, 3, 6 and 7 to publish the content specifically under the CC BY 4.0 license. 

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

b. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

6. Thank you for stating the following financial disclosure:

 'Author Contributions: LIU Maohua and LI Ruren conceived the research concept and designed it. SHAO Yue implemented the experiment and produced and analyzed the results. WANG Yan and SUN Xiubo supervised. SHAO Yue drafted the manuscript. WANG Jingkuan and YOU Yingchun reviewed the manuscript'

  1. Please provide an amended Funding Statement that declares *all* the funding or sources of support received during this specific study (whether external or internal to your organization) as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now 

  1. Please state what role the funders took in the study.  If any authors received a salary from any of your funders, please state which authors and which funder. If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

c. Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Good research. However, the manuscript does not sufficiently explain the scientific significance of the research presented in the manuscript. Some of my observations are given below:

• The research presented in the manuscripts does not seems to be adding any value as the results are not much improved from the commercially available TerraSolid.

• Line 197-198: No reason explained for setting different minimum – maximum clustering points for the two data sets.

• Line 238-240: No reason explained for setting different parameters for the two data sets.

• Figure 1 does not add any significant value to the manuscript and may not be required.

• Figure 2 and 3 does not have essential cartographic (Scale and north arrow) elements shown on them. No significant improvement in building is seen because of the filtering process and thus does not seems to be adding any value in the manuscript.

Reviewer #2: The authors:

This is an important work underlying methods development using LeDAR. The method established here is compared with a commercially available algorithm. I would recommend a comparison to an open source method in addition to the commercial software. I have my comments as a word document attached please.

Regards,

The reviewer

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: Plos_one_review_AK.docx

Click here for additional data file. (16.9KB, docx)
PLoS One. 2020 May 29;15(5):e0232778. doi: 10.1371/journal.pone.0232778.r002

Author response to Decision Letter 0

3 Apr 2020

Dear Editors and Reviewers,

Thanks very much for the time and effort you spent in reviewing our manuscript entitled “Method for extraction of airborne LiDAR point cloud buildings based on segmentation” (ID: PONE-D-19-29944R1). We would like to thank the Editors-in-Chief for giving us a chance to resubmit the paper, and also thank the Associate Editor and the two reviewers for giving us constructive and insightful suggestions. These comments would be very helpful in improving the quality of our paper and provide important guidance to our future research.

We studied the comments carefully and revised the manuscript in accordance with the comments of the editor and reviewers exactly. Revised portions were marked in red in the revised manuscript. The following is a point-to-point response to the comments of the Associate Editor and the two reviewers.

………………………………………………………………………………………

Academic Editor:

Comment 1: Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Response: Thank the Academic Editor very much for his/her constructive comments. We have ensured that the manuscript meets the journal style requirements.

Comment 2: Please amend your Data availability statement to provide direct links/URLs or contact details to where other researchers can obtain the datasets. Please also provide in your methods section enough detail about the datasets so that another researcher can replicate the datasets.

Response: Thank the Academic Editor very much for his/her constructive comments. We have modified the data availability statement in the system and provided the URL in the cover letter.

Comment 3: In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

Response: Thank the Academic Editor very much for his/her constructive comments. We have resubmitted the minimal dataset for the study and provided the URL in the cover letter.

Comment 4: PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

Response: Thank the Academic Editor very much for her constructive comments. We have ensured that the corresponding author have an ORCID iD (0000-0002-8732-3413)and verified it in the Editorial Manager.

Comment 5: We note that Figures 2, 3, 6 and 7 in your submission contain satellite images which may be copyrighted.

All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

Response: Thank the Academic Editor very much for his/her constructive comments. Figures 2, 3, 6 and 7 are not satellite images. They are 3D point clouds and we took a screenshot from a certain perspective for easy reading. Therefore, these figures do not involve copyright issues. Figures 2 and 3 show the results before and after data set 1 and data set 2 filtering. Figure 6 is the result of building extraction by the method in this paper. Figure 7 shows the results of building extraction based on TerraSolid.

Comment 6: Thank you for stating the following financial disclosure:

'Author Contributions: LIU Maohua and LI Ruren conceived the research concept and designed it. SHAO Yue implemented the experiment and produced and analyzed the results. WANG Yan and SUN Xiubo supervised. SHAO Yue drafted the manuscript. WANG Jingkuan and YOU Yingchun reviewed the manuscript'

Response: Thank the Academic Editor very much for his/her constructive comments. We have provided a revised Statement of Funding and have added a revised statement to the cover letter (Highlighted part).

………………………………………………………………………………………

Reviewer #1:

Comment 1: Line 197-198: No reason explained for setting different minimum – maximum clustering points for the two data sets.

Response: Thanks a lot for the very insightful comment. Because the data of data set 1 and data set 2 are different, the minimum and maximum number of clustering points should be set according to the specific conditions of the data. We have explained in the revised manuscript why the minimum and maximum clustering point settings in data set 1 and data set 2 are different, line 237-240.

Comment 2: Line 238-240: No reason explained for setting different parameters for the two data sets.

Response: Thanks very much for the valuable comment. TerraSolid software needs to know the minimum size of the buildings in the dataset to extract the buildings. According to prior knowledge, the minimum size of buildings in dataset 1 is 20 square meters, and the minimum size of buildings in dataset 2 is 40 square meters. We have added explanations in the revised draft, line 279-282.

Comment 3: Figure 1 does not add any significant value to the manuscript and may not be required.

Response: Thanks very much for the valuable suggestion. According to the comment of Reviewer 1, we deleted figure 1.

Comment 4: Figure 2 and 3 does not have essential cartographic (Scale and north arrow) elements shown on them. No significant improvement in building is seen because of the filtering process and thus does not seems to be adding any value in the manuscript.

Response: Thanks a lot for the very constructive and insightful comment. Based on the comment of Reviewer 1, we have added a north arrow and a scale on Figures 2 and 3, line 224-228(Since Figure 1 has been deleted, the original Figures 2 and 3 are the current Figures 1 and 2). I want to explain here, Figures 1 and 2 here are the results before and after filtering, and the purpose is to separate ground points and non-ground points rather than extract the buildings. Non-ground points include buildings, trees, and other targets. As shown in Figures 1 and 2, most ground points are removed after filtering. Figure 5 is the result of extracting buildings (line 276) and we can clearly see the buildings in Figure 5.

………………………………………………………………………………………

Reviewer #2:

This is an important work underlying methods development using LiDAR. The method established here is compared with a commercially available algorithm. I would recommend a comparison to an open source method in addition to the commercial software. I have my comments as a word document attached please.

Response: Thank Reviewer 2 a lot for his/her very constructive and insightful comments and approval of our study. Reviewer 2 pointed out the major and minor problems of our study. We revised the manuscript in accordance with his/her comments exactly. According to the comment of Reviewer 2, we compared the proposed method with the K-means algorithm in the revised manuscript in addition to the commercial software (TerraSolid).

Comment 1: Page 2.Line 21-22, Suggestion: Delete “The airborne ………..to …… ground object. It is also

Response: Thanks very much for the helpful comment. We have deleted “The airborne LiDAR technology is a new type of mapping for quickly acquiring 3D data of ground and ground objects. It is also an important”.

Comment 2: Page 2.Line 22, Change suggested: The LiDAR technology is a means of …………..

Response: Thanks a lot for the very insightful and helpful comment. We have changed "It is also an important "to "The LiDAR technology is a" based on the comment of Reviewer 2, line 30.

Comment 3: Page 2.Line 35, Cloud has lower type I …. Instead of, less type I.

Response: Thanks very much for the valuable comment. In the revised manuscript, we have changed “less type I” to “lower type I”, line 42.

Comment 4: Page 2.Line 35, Terra Solid, a commercial software.

Response: Thanks very much for the valuable comment. Since we have introduced that Terra Solid is a commercial software in the abstract, we changed the sentence to “Results show that the proposed extraction algorithm has lower type I and II errors and better extraction effect than that of the TerraSolid and the K-means algorithm.” based on the comment of Reviewer 2.

Comment 5: Page 2-4. Introduction needs strengthening, particularly on why this method is needed, and its comparative advantages to currently available open source methods.

Response: Thanks very much for the helpful comment. According to the comment of Reviewer 2, we added some contents in the first and last paragraphs in the Introduction, line 92-94, 96-103(Highlighted content).The major superiority method over currently available methods are that the proposed method is simple, efficient, and accurate. In addition, the proposed method does not need to combine any characteristic parameters and shows superiority in filtering out most of the high vegetation points.

Comment 6: Suggestion: Include a section to describe the study area, so as to have a clear view of the test sites and location of these two.

Response: Thanks very much for the valuable comment. The location and data characteristics of the study area have been described in the original text, lines 204-210.

Comment 7: Points in the LiDAR …… instead of The points in the LiDAR

Response: Thanks a lot for the very insightful comment. We have changed “The points in the LiDAR” to “Points in the LiDAR”, line 107.

Comment 8: ….. and unorganized. Suggestion: delete “….. and area unsuitable for …..to ….. interaction.

Response: Thanks a lot for the very insightful comment. Considering the comment of Reviewer 2, we have deleted “and are unsuitable for representation on the grid due to the geometry of the scanning device and the nature of the target interaction”.

Comment 9: This section (Data structure) needs improvement for clear understanding of the data used. It has missing links.

Response: Thanks a lot for the very insightful comment. This section writes the organization of the point cloud rather than a detailed description of the experimental data. The specific description of the data used is in the article 204-210.

Comment 10: Suggestion: A comparison with other methods, specifically open source would be nice.

Response: Thanks a lot for the very insightful comment. In order to better verify the usability of the method in this paper, we have added the open-source algorithm K-means algorithm to the paper. The algorithm principle and experimental results are shown in lines 290-331(Highlighted content).

Comment 11: Page 5-6.Line 97, this section needs explanation and clarity, particularly of the principle of the method.

Response: Thanks very much for the helpful comment. We added the flow of the cloth simulation filtering algorithm, lines 133-153.

Comment 12: Page 6.Line 115-117, Common ……… to region growing segmentation, etc. is an incomplete sentence. Re-write please.

Response: Thanks very much for the helpful comment. We deleted “Common segmentation algorithms such as RANSAC (Random Sampling Consistency), Euclidean distance segmentation, region growing segmentation, etc.” And the first sentence is modified as “PCL is an open source point cloud library that involves the acquisition, filtering, segmentation, registration, and feature extraction of point clouds”, line 155-156.

Comment 13: Page 7. Line 144, Suggestion: provide examples and nature of complexity.

Response: Thanks a lot for the very helpful comment. We have provided examples and nature of complexity, line 183-185.

Comment 14: Page 7.Line 147,….in each cluster was obtained after…….

Response: Thanks a lot for the very helpful comment. We have changed “….in each cluster obtained after…….” to “….in each cluster was obtained after…….”, line 187.

Comment 15: Page 9.Line 166, Confusing sentence, revise please. The data …… to …. low right.

Response: Thanks a lot for the very helpful comment. We have changed “The data set 1 shows the terrain is left high and low right.” to “Data set 1 has a high terrain in the west and a low terrain in the east”, line 205.

Comment 16: Page 9.Line 170, 31 points/m2; write and else where as points/m2.

Response: Thanks a lot for the very helpful comment. We have checked the full text and changed all points / m2 to points /m2.

Comment 17: Page 9.Line 174, Change: lush to green… surrounded by green vegetation……

Response: Thanks very much for the valuable comment. Considering the comment of Reviewer 2, we have changed “are concentrated and surrounded by lush vegetation” to “are concentrated and surrounded by green vegetation”, line 213.

Comment 18: Page 9.Line 163. As the heading of the section is “Experiment and result analysis”, I would suggest to remove this heading here.

Response: Thanks a lot for the very helpful comment. We have deleted this heading.

Comment 19: Page 11.Line 217. … to be 0…. To be 1…..

Response: Thanks a lot for the very helpful comment. We have changed this sentence in the text to “Moreover, the normal cosine value of the maximum frequency in the X and Y directions tends to be 0, and the normal cosine value of the maximum frequency in the Z direction tends to be 1”,line 258-260.

Comment 20: Page 12.Line 226. Vehicles instead of cars be more appropriate here.

Response: Thanks a lot for the very helpful comment. We have changed “cars” to “vehicles”, line 267.

Comment 21: Page 12.Line 239-240. Why the thresholds are different between set 1 and set 2.

Response: Thanks a lot for the very insightful comment. TerraSolid software needs to know the minimum size of the buildings in the dataset to extract the buildings. According to prior knowledge, the minimum size of buildings in dataset 1 is 20 square meters, and the minimum size of buildings in dataset 2 is 40 square meters. We have added explanations in the revised draft, line 279-282.

Comment 22: Page 12.Line 242. Points are calculated as buildings. Instead of mistakenly divided into buildings…..

Response: Thanks a lot for the very insightful comment. We have changed “points are mistakenly divided into buildings” to “points are calculated as buildings”, line 285.

Comment 23: Page 13.Line 249-258. Reference data used? Or how the errors were calculated and what comparison were made?

Response: Thanks a lot for the very insightful comment. The reference data used in this paper for accuracy assessment are manually extracted building results, line 311-312.

Comment 24: Page 13.Line 261. Table 2: Also include total error in addition to Type I and Type II errors.

Response: Thanks a lot for the very insightful comment. We have added a calculation method for the Total error, as shown in equation (3). And re-calculate the method of this paper, based on TerraSolid software and K-means algorithm, Type I errors, Type II errors and Type Total errors, as shown in Table 2.

Comment 25: Page 14.Line 270. Region growth…..

Response: Thanks very much for the valuable suggestion. We have changed “region growing” to “Region growth”, line 334-335.

Comment 26: Page 14.Line 278. Is compared with TerraSolid, a commercial software.

Response: Thanks very much for the valuable suggestion. We have changed “The proposed method is compared with the commercial software TerraSolid” to “The proposed method is compared with TerraSolid, a commercial software”, line 342.

Comment 27: Figure 1.Needs clarity. It doesn’t reflect the phenomena.

Response: Thanks very much for the valuable suggestion. Taking into account the comments of Review 1 and Review 2, we decided to delete Figure 1.

Comment 28: Figure 2. Doesn’t explain …

Response: Thanks very much for the valuable suggestion. Data set 1 in Figure 2(Original Figure 2 is now Figure 1) is the original point cloud data before filtering and the non-ground point data after filtering. We have already written in the article, lines 218-220. Colors in the figure represents the size of the elevation value corresponding to each color, and we have added an explanation of Colors in the text, line 215-217.

Comment 29: Figure 4 and 5. These needs explanation or reflection of building and non-building segments. The color or pattern may be changed.

Response: Thanks very much for the insightful suggestion. We have changed the colors of the normal vector histograms of the buildings in the original Figures 4-5, see Figures 3-4 after the changes.

………………………………………………………………………………………

In addition, we proofread the manuscript to minimize the typographical, grammatical, and bibliographical errors. Therefore, some sentences and words were rewritten for a native expression, but the meanings did not change. These revisions were not included in the point-to-point response to the comments of the reviewers. However, all revised portions were marked in red in the revised manuscript.

We appreciate for the warm work of the Editors & Reviewers earnestly, and hope that the corrections will meet with approval. Once again, thank you very much for your thoughtful comments and suggestions. Should you have any questions, please contact us without hesitation. Thanks and best regards!

Yours Sincerely,

Ruren LI

February 20, 2020

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file. (35.4KB, docx)

Decision Letter 1

Tayyab Ikram Shah

22 Apr 2020

Method for extraction of airborne LiDAR point cloud buildings based on segmentation

PONE-D-19-29944R1

Dear Dr. LI,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Tayyab Ikram Shah, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: Thank you for agreeing to the suggestions and incorporating most of the comments. Though I have no more comments, yet a few minor changes will further improve this manuscript. I have these as:

Introduction: non-invasive instead of noninvasive

Line 204, 218-219: datasets instead of data sets

Line 205, 211, 217: dataset instead of data set

211, 216, 218, 220 and elsewhere: figure instead of Fig; Check figure numbers 1(A) and then 1-2. The later be either 1(B) or the earlier be 1-1; be consistent in numberingfigure instead of Fig; check figure numbers 1(A) and then

Line 225, 226: (A) Raw data before filtering

(B) Non-ground points after filtering

Line 228: (B) Non-ground points after filtering

Line 302: dataset 1 not data sets 1. Check for this typo elsewhere in the manuscript.

Wish the authors best of luck

The Reviewer

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Acceptance letter

Tayyab Ikram Shah

4 May 2020

PONE-D-19-29944R1

Method for extraction of airborne LiDAR point cloud buildings based on segmentation

Dear Dr. Li:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Tayyab Ikram Shah

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Attachment

    Submitted filename: Plos_one_review_AK.docx

    Click here for additional data file. (16.9KB, docx)
    Attachment

    Submitted filename: Response to Reviewers.docx

    Click here for additional data file. (35.4KB, docx)

    Data Availability Statement

    The data in the article can be downloaded at the following DOI:10.6084/m9.figshare.12227945.

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES