【作者】 赵彦玲； 王鹏； 张国亮； 刘献礼；

【Author】 ZHAO Yan-ling WANG Peng ZHENG Guo-liang LIU Xian-li Mechanical & Power Engineering College , Harbin Univ. Sci. Tech.,Harbin 150080,China

【机构】 哈尔滨理工大学机械动力工程学院；

【摘要】 针对刀具监测中对破损区域图像边缘检测与分割效果差的问题,在使用Robert、Sobel、Prewitt几种经典分割算子检测比较后,提出了减少平滑边缘信息的Canny算子的改进算法,并以VC6.0为平台,开发了相应的分割与识别软件。经实际检验证明,本算法对刀具破损的典型形态识别具有可行性,正确率达到96%。更多还原

【Abstract】 According to recognition of the image dilapidation area problem in the TCM system,the effect of present method of edge detection and image segmentation is not satisfaction.We have used sutra arithmetic operators that Robert, Sobel and Prewitt has been researched for our examination.We improved on Canny arithmetic operators ,which can reduce smoothness information by contrast with those arithmetic operators. We has used VC6.0 as development platform.and developed segmentation and recognition program. The real operating indicates that the new method has a good feasibility and precision reaching to 96% for representative tool dilapidation shape.更多还原