Table 3.
Examples of vision-based datasets for DL-based SHM.
| Reference(s) | Goal | Dataset |
|---|---|---|
| Gulgec et al. [210] | Robust damage detection and localization of steel connections | 30,000 damaged and 30,000 healthy strain distribution matrices in 28 × 56 dimension |
| Ye et al. [211] | Concrete crack detection | 14,000 concrete crack images with 80 × 80 pixel resolutions obtained from a concrete beam test |
| Xu et al. [131] | Multi-type seismic damage identification and localization | 2400 images with 640 × 640 pixel resolution of concrete cracking, concrete spalling, rebar exposure, and rebar buckling |
| Nahata et al. [193] | Post-earthquake damage extent identification of buildings | 1200 RGB images with 224 × 224 × 3 pixel resolution |
| Beckman et al. [167] | Concrete spalling damage detection and quantification | 444 concrete spalling images with the resolution of 853 × 1440 pixels |
| Jang et al. [157] | Detection of micro and macro concrete cracks | 20,000 hybrid combining vision and infrared thermography of concrete crack and intact images with 224 × 224 pixel resolution |
| Dung and Anh [33] | Concrete crack detection, segmentation, and density evaluation | A public dataset of 40,000 concrete crack 227 × 227 pixel images |
| Zhang et al. [166] | Real-time autonomous bolt loosening detection | 300 tight and loosened bolt images with 224 × 224 pixels |
| Wang et al. [171,196] | Spalling detection for historic masonry structures | 500 images with 500 × 500 pixel resolutions |
| A vision-based [189] | Crack detection of gusset plate welding in steel bridges | 12,896 images with 64 × 64 pixels of cracks and the same number of non-cracks |
| Liu and Zhang [192] | Image-driven low cycle fatigue-induced damage identification for post-hazard inspection | 8259 images with 224 × 224 pixels extracted from a the video was taken during the experimental test |
| Ni et al. [198] | Concrete thin crack identification and width measurement | 65,319 crack and 64,681 non-crack 224 × 224 RGB images for GoogleNet and 60,000 images for ResNet |
| Hoskere et al. [140] | Rapid and autonomous post-earthquake inspections including identification of damage presence and damage type | A set of 665 images with 288 × 288 pixels containing post-earthquake damage scenarios such as concrete cracks, spalling and exposed rebar |
| Chen [113] | Inspection of nuclear power plants and detection of cracks in video frames | A total 147,344 crack and 149,460 non-crack 120 × 120 image patches |
| Liang [143] | Post-disaster system-level failure analysis of bridges, the structural component-level identification, and local-level damage localization | 492 number of 224 × 224 RGB images |
| Zhao et al. [202] | Classification the types of bridges, recognition of bridge components and crack detection | 3832 RGB images of an arch, suspension and cable-stayed bridges with 227 × 227 pixel resolution and 60,000 intact and cracked concrete RGB images with 224 × 224 pixel resolution |
| Gao et al. [142,212] | Component type, spalling condition, damage level, and damage type determination. | 2000 images with a size of 224 × 224 RGB (Structural ImageNet) |
| Dorafshan et al. [203] | Autonomous inspection of concrete structures using Unmanned Aerial Systems (UASs) | 9011 227 × 227 pixel images of lab-made bridge decks including 1471 cracked and 7540 intact cases taken by Nikon camera |
| Suh and Cha [204] | Damage type detection and localization | 2366 images of concrete cracks, steel delimitation, corrosion, and bolt loosening with a size of 500 × 375 pixel |
| Wang et al. [196] | Damage type identification (intact, crack, efflorescence, and spall) and localization for masonry historic structures | 5145 stretcher and header brick images with 480 × 105 and 210 × 105 pixel resolutions respectively |
| Kim et al. [72] | Determining the existence and location of cracks from surface images considering crack-like noise patterns | 3186 images crack and intact surfaces with different distances between the camera and 227 × 227 pixel |
| Silva et al. [213] | Automated inspection of concrete structures and crack detection | 3500 sample of concrete surface images of 256 × 256 pixels with and without cracks |
| Dorafshan et al. [96] | Image-based crack detection in concrete structures | 18,000 concrete panel images with the size of 256 × 256 pixel simulating reinforced concrete bridge decks |
| Sharma et al. [214] | Image-based detection of the crack presence | 15,600 crack and non-crack 28 × 28 RGB image patches |
| Li and Zhao [169] | Crack detection of concrete surfaces | real concrete surface RGB images with 224 × 224 pixel resolution and taken by a smartphone |
| Kang and Cha [14] | Autonomous UAV method using ultrasonic beacons for damage detection and localization | 40,000 cracked and intact concrete surface images with 256 × 256 pixel |
| Yang et al. [101] | Semantically identification and pixel-wise segmentation | A collection of 800 images of various cracks with 224 × 224 pixel |
| Kim and Cho [122] | Crack detection on concrete surfaces | 7195 images of cracks, joints, edges, plants, and intact surfaces with 227 × 227 pixel scraped from the Internet |
| Beckman et al. [205] | Volumetric damage detection and quantification | 444 images of concrete spalling with the resolution of 853 × 1440 pixels |
| Kumar et al. [215] | Automated CCTV inspection of sewer pipelines | 12,000 images of cracks, root intrusions, deposits, and in pipelines with the dimension of 256 × 256 |
| Narazaki et al. [144,190,199] | Structural bridge components recognition | 39,081 images of a concrete girder bridge with a size of 240 × 320 |
| Karaaslan et al. [186] | Mixed reality inspection including detection and segmentation of cracks and spalls | 51,300 concrete crack, road damage and bridge inspection images with the size of 300 × 300 |
| Oliveira et al. [216] | Damage classification in an aluminum plate | 720 frames of 128 × 128 greyscale representation of electromechanical impedance |
| Li et al. [217] | Concrete bridge inspection and estimating the probability of being cracked | 326,000 samples of 18 × 18 one channel greyscale patches of concrete cracks and non-cracks |
| An et al. [124] | Autonomous detection of macro- and micro-cracks | A set of 20,000 images of crack and intact images 227 × 227 RGB images |
| Duarte et al. [163] | Building damage (rubble piles, debris) classification and assessment from images such as | A total of 12,973 of the satellite and airborne 224 × 224 pixel images |
| Ji et al. [218] | Identification of collapsed buildings from post-event satellite images | 613 collapsed and 1857 non-collapsed buildings Post-Earthquake Satellite images with the resolution of 96 × 96 |
| Kang and Cha [159] | Intact and cracked concrete area classification | A broad variation of 2304 × 1280 raw images of concrete surfaces |
| Modarres et al. [219] | Crack detection of concrete bridges and composite panels | 2400 number of real concrete crack and intact surfaces and 6000 debonded and intact sandwich panels with a resolution of 96 × 96 pixel |
| Yeum et al. [141] | Classification and localization for visual assessment of connections in a full-scale truss structure | 100,000 images of welded joints with a dimension of 256 × 256 |
| Hoskere [139] | Damage localization and classification for post-earthquake structural assessment | 1695 images of concrete spalling, exposed rebar, steel corrosion, concrete cracks, fatigue cracks and asphalt cracks with a resolution of 288 × 288 |
| Nazaraki et al. [191] | Pixel-wise bridge component recognition | 11,897 urban, bridge and general 180 × 180 images |
| Atha and Jahanshahi [31] | Assessment and corrosion detection on metallic surfaces | 67,187 images of regions with and without corrosions with 128 × 128 pixel resolution |
| Cha et al. [25] | Automatic concrete crack detection | 40,000 images of cracks on concrete images with a dimension of 256 × 256 pixel |