Target Recognition and Grab Location Based on Machine Vision and Deep Learning

The application of target recognition and sorting has greatly improved the level of industrial automation, liberated human hands, and effectively saved resources and improved efficiency. The purpose of this paper is to study target recognition and location based on machine vision and deep learning. Firstly, the development of machine vision is summarized, the algorithm of spatial structure feature extractor is proposed, and the overall architecture of fusion model is explained simply. After comparing the network model of this paper with that of SSD, the network positioning accuracy is analyzed. The experimental results show that when the threshold IoU is less than 0.5, the detection performance of faster R-CNN is better than that of SSD network, and when the threshold IoU is more than 0.5, the situation is the opposite. The experimental results show that the positioning accuracy of SSD network is higher than that of faster R-CNN, and the comprehensive detection performance gap between the two network models will reverse with the improvement of task requirements for positioning accuracy.