Table 3.

Comparison of the accuracy of different object detection networks

Models	R2	MAE	MAPE	RMSE
Mask R-CNN	0.824	16.0	29.6%	25.0
Swin-S Mask R-CNN	0.894	10.1	18.3%	18.0
YOLO V4	0.914	11.5	23.1%	17.5
YOLOV5-L	0.921	11.1	17.8%	14.0
YOLOX-L	0.918	8.72	17.5%	14.4
YOLO POD	0.967	4.18	10.0%	6.48

All the methods are tested at 1024 × 1024 resolution. Bold text indicates the best results.