Research on surface defect intelligent detection technology of non-woven fabric based on support vector machine

In order to realize the on-line detection of non-woven fabric defects in the industrial environment, this paper first designs a set of stable non-woven fabric defect image acquisition system according to the requirements of the industrial site, and selects the linear array with linear light source, the camera adopts backside lighting system to collect cloth image. In this paper, a defect detection method based on support vector machine (SVM) is proposed for non-woven fabric turning mesh defect detection. Four eigenvalues of gray co-occurrence matrix describing texture parameters are extracted. The training is carried out in the classifier of kernel function backward guide support vector machine, and the abnormal area is located by SVM. Experimental results show that this algorithm can correctly identify the network turning defects, and can take into account high accuracy and real-time; At the same time, it can effectively solve the problem that it is difficult to obtain defect training samples in production practice, and has good promotion for unknown samples to be tested, which can meet the requirements of industrial testing.