Research And Application Of Industrial Filter Cloth Breakage Inspection System

With the development of computer technology from the research phase to the application phase,enterprises also put forward new requirements for full automation of industrial sites.In terms of the food and beverage industry,industrial filter cloth is often used to filter out impurities to obtain high quality raw materials in the production process.Due to the multiple extrusion and folding of the filter cloth,it is inevitable that the filter cloth will be damaged,which will make the food not pure enough and the sediment is too much,resulting in a great drop in the quality of the food.Therefore,how to quickly and accurately detect the breakage in the filter cloth is a major problem in the food industry.In this paper,an industrial filter cloth breakage detection system is designed by combining machine vision,digital image processing and machine learning algorithm,which can meet the requirements of online detection,automatic control and intelligent management in automatic production process.The main work of this paper is as follows:Firstly,this paper analyzes the research background and significance of development of industrial filter cloth breakage detection system,introduces the development status of the fabric defect detection system at home and abroad,and the development process and research status of machine learning,and finally analyzes the shortcomings of the cloth defect detection system in China.Secondly,the detection process and principle of filter cloth are analyzed,and then the technical indicators required by the system are introduced,and the hardware structure and software framework of the system are designed.Finally,the breakage detection based on GLCM and the breakage classification based on machine learning algorithm are introduced.Thirdly,the texture feature extraction method of filter cloth is introduced in detail,and the GLCM feature extraction and ULBP feature extraction methods,as well as the important parameters affecting the extraction results,are analyzed with emphasis.Fourthly,the filter cloth breakage detection based on GLCM is designed,including the filtering and enhancement of the image in the early stage,the calculation of the filter cloth period,the confirmation of the gray level,the selection of the GLCM structure parameters,the determination of the adaptive threshold and extraction of feature areas and calculation of area.Summed-up distance matching superposition function is used to calculate the filter cloth cycle,and the optimal structural parameters are used to extract GLCM.Finally,the texture features of the filter cloth image are extracted by GLCM and ULBP operator,then the principle and application of the SVM learning algorithm are analyzed.,and the model is constructed by using the algorithm,then the training method and hyperparameter adjustment of the model are described in detail.Finally,the classification results using GLCM features,ULBP features and combined features are compared and the results are analyzed.