摘要
采用基于Python的Opencv计算机图像处理技术对CCD相机提取到的齿轮图像信息进行处理。通过对完好齿轮图像做灰度化处理以及阈值处理,再经过形态学操作之后,分别通过Canny和Sobel算子对图像轮廓信息进行比较,提取较为理想的轮廓信息。将有损伤缺陷的齿轮经过同样的方法提取轮廓边缘,由于采集的齿轮每次的位置以及角度不一致,采用具有旋转平移缩放不变性的Hu矩进行形状匹配,能快速简易地对齿轮缺陷与损伤作出识别与量化评价。
Opencv computer image processing technology based on Python was used to process the gear image information extracted from CCD camera. After graying and thresholding the intact gear image,and morphological operation, the image contour information was compared by Canny and Sobel operators,and the ideal contour information was extracted. The gear with damage defect was extracted by the same method for contour edge. Since the position and angle of the gear collected are inconsistent each time, Hu moment with invariant of rotation, translation and scaling was adopted for shape matching,which can quickly and easily identify and quantify the gear defect and damage.
作者
寇皓为
苏燕辰
陈国俊
罗莉
Kou Haowei;Su Yanchen;Chen Guojun;Luo Li(School of Mechanical Engineering,Southwest Jiaotong University,Chengdu 610031,China;School of Literature,Sichuan Normal University,Chengdu 610066,China)
出处
《煤矿机械》
2021年第5期44-46,共3页
Coal Mine Machinery
关键词
图像处理
齿轮损伤检测
阈值处理
边缘检测
HU矩
形状匹配
image processing
gear damage detection
threshold processing
edge detection
Hu moment
shape matching