基于YOLOv3算法的工业分拣目标检测

Industrial Sorting Object Detection Based on YOLOv3 Algorithm

在工业机器人视觉导引系统中,对于不规则的工业零部件,基于模板匹配的目标检测算法难以取得较好的定位效果和识别精度。在结构化环境下,对于传统视觉目标检测算法,利用YOLOv3算法开展视觉目标检测的实践。实验结果表明,该算法的定位精度和识别精度均取得了较高的水平,可以在工业现场进行部署。

In the vision guidance system of industrial robot,the object detection algorithm based on template matching is difficult to achieve good positioning effect and recognition accuracy for irregular industrial parts.This paper is a practice of visual object detection using YOLOv3 algorithm for traditional visual object detection algorithm under structured environment.The experimental results show that the localization accuracy and recognition accuracy of the algorithm have achieved a high level,and can be deployed in industrial sites.