Supervisor:
Southwest Ordnance Industry Bureau
Organizer:
Chongqing Ordnance Industry Society
Chongqing University of Technology
Chongqing University of Technology
Accurate visual detection and segmentation algorithm for special shaped parts assembly
| DOI: | 10.11809/bqzbgcxb2024.05.034 |
| Keywords: | small target; target detection; image segmentation; special shaped parts |
| Abstract: | In order to solve the problem of inaccuracy in the recognition of special shaped small targets in parts in industrial production, based on YOLOv7 visual detection method and U2Net image segmentation method, this paper designs YOLOv7 GTBM small target detection method and U2NET MGP pixel level image segmentation method for the detection of special shaped small targets. In this paper, the SIM card slot replaces the special shaped small parts in industrial parts as the experimental object. Firstly, the SIM card slot is roughly located by small target detection method, and then the edge contour is accurately segmented by pixel level image segmentation method. Finally, the subsequent operations are completed according to the actual assembly needs of the SIM card slot. The experimental results show that the scheme is superior to other algorithms in both detection effect and segmentation accuracy, and the success rate of capturing SIM card slot reaches 98%. |
| Published: | 2024-05-31 |