Research on Improved YOLOv7 Container Cargo Target Detection
Aiming at the problem that it is difficult to realize the precision target identification because of the centralized placement of box-type goods studied in this paper,an improved method based on YOLOv7 network is proposed and ap-plied to the detection in the centralized placement of box-type targets.Since box-type goods are usually stacked together and shielded from each other during unstacking,it is difficult for the target detection model to accurately identify the loca-tion and boundary box of each box.Coupled with the complex and harsh industrial environment,it is difficult to realize faster attention to the target.Therefore,CBAM attention mechanism is added to the original YOLOv7 network.Through the channel attention module and the spatial attention module,the complex scene information can be analyzed more quickly and efficiently,so as to achieve the real-time goal,so that the unpalletizing robot can identify and locate the target goods more quickly and accurately,and then perform the grasping action.
NETL
NSTL
万方数据
