Research Article

Yolo-Based Improvements in Remote Sensing Image Applications

Table 5

Comparison experiment on the NWPU VHR-10 dataset.

ā€‰Fast RCNNGACL
fast RCNN
Faster RCNNGACL
faster RCNN
YOLOv3YOLOv3-DEYOLOv4YOLOv5YOLOv7

Ship62.863.589.989.883.390.2590.590.193.5
Bridge79.272.780.986.178.1489.6180.186.292.3
Ground track field99.399.4100.099.899.8599.6799.799.199.8
Storage tank44.553.967.368.579.2787.3580.6785.286.4
Basketball court80.782.287.588.482.7790.4690.185.388.6
Tennis court76.980.978.679.789.1690.4190.892.195.4
Airplane90.090.390.790.990.8699.8792.897.399
Baseball diamond90.096.989.288.890.7186.9298.394.398.9
Harbor90.490.989.890.082.5590.6989.188.592.1
Vehicle49.757.988.088.579.3581.3288.891.196.7
Mean AP76.478.986.286.685.59690.64690.08790.9294.27