Research On Automatic Detection Technology Of Rotor Winding Based On Machine Vision

The motor rotor,as a product of huge demand in the industrial field,still needs the manual calibration for the detection of its winding quality.And it turns out to be low efficiency sometimes.Taking the common TT720 rotor as a sample,this paper is devoted to the study of an online detecting system for the qualification of the workpiece winding patterns.The main research contents are as follows:(1)The module architecture and process scheme of online detection system are studied.The software and hardware system is designed for the module functional requirements of image acquisition,system control,image processing and recognition and the automated execution of the inspection process is realized then.(2)A double-template progressive matching localization algorithm is proposed.It is a time-consuming process that each source image needs to be relocated based on the template.To solve this problem,the size of source image is optimized based on the the historical location information.The experiments show that the proposed algorithm can effectively reduce the time cost and improve the positioning stability.(3)A classification algorithm for image feature extraction and differential expression is studied.The similarity mathching mechanism is established according to differences between the normal and flaw image by the extracting the features of contour edge,texture and morphology.Experiments show that this method can accurately detect the eligibility of rotor image.(4)A template searching algorithm is proposed based on the contour distribution descriptor.In the process of image feature matching,the value of the similarity is fluctuated when the image of the sample is matched with the different template.The suitable template is selected from the standard template library by using this descriptor and the similarity distribution of feature matching becomes stable after selecting.Based on the above-mentioned algorithm,an on-line detection system is developed,and the detection capability of this system is verified through on-site experimental testing.