Mango Target-Detection Algorithm Based on YOLOv8 Integrated with Adaptive Spatial Pyramid

Dataset	Number of image	Average resolution /（pixel ×pixel）	Average count	Maximum count	Number of annotation
With mango of self-made dataset	3058	4032×3024	8.0	25	24464
Without mango of self-made dataset	302	4032×3024	0	0	0
MangoSet-A	7379	640×640	1.2	4	7351
MangoSet-B	50	640×640	2.7	12	134
MangoSet-C	1129	224×224	1.6	19	1250

Table 2. Comparison of CA for different algorithms on different datasets
View table
Table 2. Comparison of CA for different algorithms on different datasets
Algorithm MangoSet-A MangoSet-B MangoSet-C Self-made dataset
Faster R-CNN 86.45 81.23 89.10 84.56
RetinaNet 87.98 85.50 90.25 86.30
YOLOv8 88.67 84.20 89.90 85.00
Proposed 90.12 86.47 92.35 87.80

Table 3. Comparison of CA between YOLOv8 and the proposed network at different resolutions
View table
Table 3. Comparison of CA between YOLOv8 and the proposed network at different resolutions
Resolution /（pixel×pixel） YOLOv8 Proposed
1280×1280 89.46 91.69
640×640 87.32 90.83
320×320 85.12 87.07

Tools

Get Citation

Copy Citation Text

Baoyu Wang, Hantang Li, Xiyong Chen, Wei Yao. Mango Target-Detection Algorithm Based on YOLOv8 Integrated with Adaptive Spatial Pyramid[J]. Laser & Optoelectronics Progress, 2025, 62(14): 1415002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites