基于深度学习的铸铁件识别与定位算法研究

Research on Iron Casting Recognition and Positioning Algorithm Based on Deep Learning

对铸铁件识别与定位系统进行了研究,利用YOLOv5s算法对不同铸铁件的正反面放置情况进行训练与检测,利用图像预处理算法,主要包括灰度处理、图像增强、滤波处理等算法,之后又进一步改进边缘检测算法和模板匹配算法,并结合以上算法对铸铁件进行定位处理。实验结果表明,采用YOLOv5算法可以快速、准确地检测出不同铸铁件的正反面放置情况,并且平均置信度可达到95%。结合图像预处理算法、改进后的Canny边缘检测算法以及SUFT模板匹配算法,能够实现铸铁件的定位过程。因此,所采用的铸铁件检测算法与定位算法可行,并且检测效果较好,为后续深入研究铸铁件定位与抓取技术提供了一定的理论基础。

The recognition and positioning system of iron castings was studied,the YOLOv5s algorithm was used to train and detect the placement of the front and back sides of different iron castings,and using image preprocessing algorithm,mainly including grayscale processing,image enhancement,filtering processing and other algorithms,and then the edge detection algorithm and template matching algorithm were further improved,and the positioning processing of iron castings was carried out in combination with the above algorithms.Experimental results show that the YOLOv5 algorithm can quickly and accurately detect the front and back placement of different iron castings,and the average confidence level reaches 95%.In this paper,the image preprocessing algorithm,the improved Canny edge detection algorithm and the SUFT template matching algorithm are combined to realize the positioning process of iron castings.Therefore,the cast iron detection algorithm and location algorithm adopted in this paper are feasible,and the detection effect is good,which provides a certain theoretical basis for further research on the positioning and grasping technology of cast iron.