关键词: CMOS, 机器视觉, 边缘检测算子, 活立木, 边缘直线, 特征提取

Abstract:

Based on the image processing and machine vision technology,the edge feature of the tree trunk image was extracted by the edge detection operator,and the linear feature of the tree trunk edge was studied.The results show that:Prewitt operator and Canny operator retain more edge details on the trunk background,but the detection rate of the trunk edge is higher,and the trunk edge continuity is better.The detection error of the edge of the tree between the Robert operator and the Prewitt operator is large.The Sobel operator,LoG operator and Canny operator have higher detection precision for the edge of the trunk.The Gaussian filtered standing images of Sigma 0.1,0.3,1.4 and 3 are edge-detected and the diameter pixel distance extraction value is compared with the visual interpretation value.The results show that the recognition bias of Sigma is 0.3~1.4.

Key words: CMOS, machine vision, edge detection operator standing timber, edge straight, feature extraction