Table 1.

Comparison of YOLO with related works.

Reference	Dataset Used	Algorithms	Findings
Li et al., 2021 [26]	Remote sensing images collected from GF-1 and GF-2 satellites. Training: 826 images. Testing: 275 images. Resolution: 300 × 300, 416 × 416, 500 × 500, 800 × 800, 1000 × 1000	Faster R-CNN YOLO v3 SSD	YOLOv3 has higher mAP and FPS than SSD and Faster R-CNN algorithms.
Benjdira et al., 2019 [12]	UAV dataset Training: 218 Images Test: 52 Images Resolution: 600 × 600 to 1024 × 1024	Faster R-CNN YOLOv3	YOLOv3 has higher F1 score and FPS than Faster R-CNN.
Zhao et al., 2019 [27]	Google Earth and DOTA datasetTraining: 224 Images Test: 56 Images Resolution: 600 × 600 to 1500 × 1500	SSD Faster R-CNN YOLOv3	YOLOv3 has higher mAP and FPS than Faster R-CNN and SSD.
Kim et al., 2020 [29]	Korea expressway dataset Training: 2620 Test: 568 Resolution: NA	YOLOv4 SSD Faster R-CNN	YOLOv4 has higher accuracy SSD has higher detection speed
Dorrer et al., [28]	Custom Refrigerator images Training: 800 Images Test: 70 Images Resolution: NA	Mask RCNN YOLOv3	The detection of YOLOv3 was 3 times higher but the accuracy of Mask RCNN was higher.
Rahman et al., [13]	Custom Electrical dataset Training: 5939 Test: 1400 Resolution: NA	YOLOv4 YOLOv5l	YOLOv4 has higher mAP compared to YOLOv5l algorithms
Long et al., [30]	MS COCO dataset Training: 118,000 Test: 5000 Resolution: NA	YOLOv3 YOLOv4	YOLOv4 has higher mAP compared to YOLOv3
Bochkovskiy et al., [7]	MS COCO dataset Training: 118,000 Test: 5000 Resolution: NA	YOLOv3 YOLOv4	YOLOv4 has higher mAP and fps than YOLOv3
Ge et al., [14]	MS COCO dataset Training: 118,000 Test: 5000 Resolution: NA	YOLOv3 YOLOv4 YOLOv5	YOLOv5 has higher mAP than YOLOv3 and YOLOv5l YOLOv3 has higher FPS than YOLOv4 and YOLOv5l