In order to solve the problem of Riveting structure damage detection in weapon equipment,support vector machines (SVM) method of computer visual recognition was proposed.First,the image of rivet was collected and image preprocessing was carried out to extract the four relevant characteristic parameters.Secondly,under the condition of small sample,the SVM algorithm was used to train and establish a visual recognition model,and the relationship between the characteristic parameters and the final model was analyzed.Finally,the SVM vision recognition method,Hough transform method and independent judgment method were compared and analyzed.The comparison results showed that the recognition rate of the SVM vision recognition method is faster and the recognition accuracy is higher,and the feasibility and effectiveness of the proposed method was proved.
NETL
NSTL
维普
万方数据
