基于图像处理技术的绝缘子裂纹识别研究农业电气化与自动化专业论文.docxVIP

优秀毕业论文 精品参考文献资料 江苏大学高校教师硕士学位论文ABSTRACT 江苏大学高校教师硕士学位论文 ABSTRACT Development of the patrol robot for automatic transmission line is to ease people’S work condition,improve the patrol efficiency and the safe operation of the power supply．The patrol robot’s vision simulation systems task for HV Power Line inspection was presented in the paper．The insulator images were obtained by simulating vision system of the robot．The images Were processed based on image processing and recognition，then the cracks can be detected and the classifications of the cracks were finally achieved．The main work has been done as follows： Firstly,a system platform of the automatic checking and recognition of the insulator crack image 011 thcoretics was brought．The hardware was designed to acquire the insulator image． Secondly,three methodsre studied on image filter,including mean filtering， median filtering and adaptive filtering．They a陀used to deal诵th the image in order to strengthen the image．The appropriate selection of image smoothing method is based on image features and characteristics of the noise． Thirdly,the edge detection methods were adopted in order to obtain the image edge．It brings forward an image subtraction extraction method based Oil the crack’S edge characteristics．The edge of the image WaS detected based on a variety of methods，including gradient-based edge detection,LoG operator,Canny edge detection and wavelet multiscale edge detection．The results of the experiment were compared．The crack images were obtained by using subtraction method．Then，the original image is separated into an image of the 1 6 1 6 binary features image，which is eliminated”isolated”point by the disrepair extent technology． Lastly,the three—layer BP neural network was constructed to recognition crack image on the insulator．Four features were extracted to form the feature vector,which includes the number of the disrepair sub—image，connection,horizontal projection variance，vertical projection variance，and