Research On Inspection Method Of Steel Wire Rope Surface Defects Based On IWOA-SVM

As a commonly used load-bearing component in logistics equipment,steel wire rope plays an important role in the use of logistics operations.In the actual use process,the steel wire rope working environment is relatively bad,coupled with a long-term high load state,it is inevitable that there will be wear,broken wire,deformation and other defects,and even gradually evolved into the steel wire rope fracture,threatening the safety of life and property of staff.Therefore,in order to accurately and timely find the surface defects of wire rope,and carry out risk assessment and maintenance,it is necessary to carry out nondestructive testing research on wire rope.In recent years,machine vision technology provides a new idea for wire rope defect detection.After studying the nondestructive testing of steel wire rope using machine vision technology at home and abroad,it is found that the testing method has many problems,such as single feature extraction method,insufficient use of feature information,lack of optimization of classifier,poor generalization performance,etc.,which affect the testing effect of steel wire rope.Focusing on the problems existing in wire rope defect detection,this paper proposes a wire rope defect detection method based on image feature fusion and IWOA-SVM.The main research contents are as follows.(1)the construction of the image acquisition system hardware,wire rope under the condition of existing laboratory chose can satisfy the requirement of the industrial camera lens and supporting industry,and to build the hardware part of the collection system of wire rope image acquisition,according to the collected the data sets of wire rope image produced experiment.(2)Research on wire rope image preprocessing methods.In order to improve the image quality of wire rope and improve the robustness of wire rope defect detection algorithm,CLAHE algorithm was used for image preprocessing of collected wire rope images before feature extraction,and other image enhancement algorithms were combined for comparison.The applicability of CLAHE algorithm in this paper is proved.(3)Research on wire rope image feature extraction method.In order to make full use of wire rope image information and improve image feature description ability,this paper based on CPICS-LBP combined with HOG wire rope image feature extraction method.Firstly,the texture features and gradient features of the wire rope image were extracted by using CPICS-LBP and HOG algorithm,and the new fusion features were fused to form new fusion features.Secondly,PCA is used to reduce the dimensions of the fusion features to reduce the noise and redundancy.The experimental results show that the combination of CPICS-LBP HOG feature has stronger feature description ability than the single feature,and can improve the identification accuracy of wire rope defect detection.(4)Research on the feature recognition method of wire rope image.Aiming at the problem of poor generalization of classifier,an improved whale optimization algorithm(IWOA)is proposed to optimize the classifier model of SVM.Firstly,the control factors of whale optimization algorithm(WOA)are improved,and nonlinear control factors and adaptive inertia weights are introduced to improve the search ability of WOA algorithm.Secondly,IWOA is applied to SVM to optimize its parameters and improve SVM performance.Finally,the CPICS-LBP and HOG features extracted from the experimental data set were input to the IWOA-SVM classifier for recognition and classification.The experimental results show that the SVM classifier optimized by IWOA can improve the identification accuracy of wire rope defect detection,and the algorithm has high stability.