Research On The Defect Detection Technology Of Linear Array Based On Machine Vision

As the core component of the internal combustion engine,the cylinder liner's quality directly affects the overall performance.In the production process of the cylinder liner,due to its own internal pressure,clamping force,etc.,it is easy to cause defects such as air holes,cracks and shrinkage holes on the surface of the cylinder liner.The occurrence of these defects will affect the performance of the product.At present,domestic defect detection for casting surfaces is mostly done manually.As the market demand increases,this method has certain limitations.Not only is it inefficient,it is also easily affected by subjective factors.Based on machine vision,it can not only avoid the problems of human eye recognition,but also has the advantages of robustness and efficiency.Therefore,this paper presents a linear array cylinder liner defect detection and identification research based on machine vision,mainly to complete the following research.(1)Overall research of the system platform.Firstly,a complete set of image acquisition hardware platform is built,which can realize automatic detection of cylinder liner defects.Secondly,the quality of the acquired image is improved through the optimized design of the X-ray imaging system.(2)Research on image processing of cylinder liner defects.First,a comparative analysis of the frame integral superposition method and the spatial filtering method,combined with the spatial filtering algorithm,this paper uses a combination of median filtering and Wiener filtering to reduce image noise.Secondly,in order to increase the contrast in light and dark areas,histogram equalization is used to enhance the image.Then,through the exhaustive method,the threshold algorithm is compared one by one,and an algorithm based on the combination of the maximum correlation improvement method and morphological filtering is proposed to segment the suspected surface of the cylinder sleeve.The experimental resultsshow that the suspected defects on the surface of the cylinder liner are completed.(3)Research on extraction and recognition classification of cylinder liner defect surface features.First,according to the characteristics of cylinder liner surface defects in the image,geometric features,Hu invariant moments and improved LBP texture features are extracted as feature vectors,and PCA principal component analysis is used to reduce the dimensionality of the fused high-dimensional data.Then,the related theory of support vector machine SVM is studied,and the radial basis is selected as the kernel function and related parameters to classify the defects.Experimental results show that it meets the requirements of cylinder liner defect classification.The linear array cylinder liner defect detection research based on machine vision proposed in this paper can quickly and accurately detect defects while also realizing the rapid classification of defect types.It meets the needs of enterprises for automatic detection of defects and has important academic research value.