中国计量大学学报2017，Vol.28Issue(1)：29-34,6.DOI:10.3969/j.issn.2096-2835.2017.01.005
一种小模数齿轮边缘检测效果评价方法
A feature image-based method for evaluating small modulus gear sub-pixel edge-detection effectiveness
摘要
Abstract
When machine vision is used to evaluate the accuracy of a small-modulus gear,the edge feature information extracted from the whole gear image cannot directly describe an individual target in the image.An identification algorithm is needed to adapt to the variable local features subsequently.We presented a feature image-based method for receiving the effectiveness of edge detection to fetch abundant local image information and evaluate the profile extraction accuracy of the small-modulus involute gear vision measurement system.Firstly,a feature image model based on the functional features of the invoIute tooth profile edges in the gear image was created.Then,the Zernike-moment sub-pixel edge detection algorithm was used to fetch the edges of the small-modulus involute gear feature image.At last,based on the standard function for constructing the feature image,the deviation between the edge detection result of the feature image and that of the standard function was quantified so as to evaluate the edge detection effectiveness.The experiment shows that,when the feature image of the small-modulus gear is used to evaluate the Zernike-moment sub-pixel edge detection algorithm,the detection accuracy of the involute tooth profile is better than 0.58 pixels.关键词
Zernike矩/边缘检测/特征图像/边缘评价/小模数齿轮Key words
Zernike moment/edge detection/feature image/edge evaluation/small-modulus gear分类
信息技术与安全科学引用本文复制引用
周泽恒,叶树亮,朱维斌..一种小模数齿轮边缘检测效果评价方法[J].中国计量大学学报,2017,28(1):29-34,6.基金项目
国家质检总局公益性行业科研专项(No.201210001-3). （No.201210001-3）
