Table 2.
Summary of CNN-based data analysis approach in imaging-based plant phenotyping.
| Phenotyping category | Phenotyping task | Main processing approach | Particular improvement strategy | References |
|---|---|---|---|---|
| Plant stress | Stress detection and classification | Image classification | NA | [43, 44, 47–53] |
| Sliding window | [45, 57] | |||
| Explainable visualization | [55, 56] | |||
| Advanced imaging | [58] | |||
| Synthetic data augmentation | [54] | |||
| Object detection | NA | [46] | ||
|
| ||||
| Plant development | Plant lodging | Image classification | NA | [68] |
| Canopy morphology measurement | Object detection | NA | [61, 65] | |
| Semantic segmentation | NA | [59, 63, 64] | ||
| Leaf morphology measurement | Instance segmentation | NA | [60, 62] | |
| Characterization of plant growth pattern | Combination of CNN and other DL methods | NA | [66, 67] | |
|
| ||||
| Plant development | Counting plant/plant organs in still images | Regression | NA | [69, 70, 72, 79] |
| Synthetic data augmentation | [71, 73] | |||
| Multiscale and multimodal data fusion | [74, 75] | |||
| Nonsupervised learning mode | [76, 78, 80] | |||
| Explainable visualization | [77] | |||
| Image classification | NA | [82, 83] | ||
| Object detection | NA | [84–91, 93, 94] | ||
| Sliding window | [92, 95] | |||
| Synthetic data augmentation | [95] | |||
| Semantic segmentation | NA | [96–101] | ||
| Sliding window | [101] | |||
| Instance segmentation | NA | [102–106] | ||
| Synthetic data augmentation | [102, 103, 105, 106] | |||
|
| ||||
| Plant development | Counting plant/plant organs in image sequences and videos | Object detection | 2D orthoimage reconstruction | [111–113] |
| 3D structure reconstruction | [107, 115–119] | |||
| Video tracking | [108–110] | |||
| Semantic segmentation | Movement encoding | [114] | ||
|
| ||||
| Plant development | Counting root tips | Regression | NA | [120] |
| Root system architecture segmentation | Semantic segmentation | NA | [120–124] | |
| Inpainting for oversegmentation correction | [125, 127] | |||
| Advanced imaging | [126, 128] | |||
| Synthetic data augmentation | [125, 126] | |||
|
| ||||
| Postharvest quality | Fruit chemical composition measurement | Regression | NA | [138] |
| Fruit defect detection | Image classification | NA | [131, 132, 135, 136, 140] | |
| Advanced imaging | [134] | |||
| Sliding window | [129, 137] | |||
| Fruit defect quantification | Semantic segmentation | NA | [141] | |
| Advanced imaging | [139] | |||