Improved YOLOV3 target recognition algorithm for adaptive edge optimization

In order to accurately identify the target part in the image, an adaptive edge optimization YOLOV3 target recognition algorithm was proposed. Firstly, the K-means++ algorithm is used to calculate the anchor box suitable for the data set of this paper. Then the adaptive edge error function is designed and combined with the improved PSO algorithm to obtain the YOLOV3-AEEF algorithm. Then collect the part image and enhance the data, label the picture, and get the sample set. The network is trained after loading the pre-training weights and tested on the test set. The experimental results show that when there are more incomplete parts interference in the sample picture, YOLOV3 identifies the background as a part, and YOLOV3-AEEF can accurately identify the target part, and in the case of a higher recall, the precision is 21% higher than YOLOV3, which improves the overall performance of the network.