Research On The Method Of Mesh Quality Detection Based On Machine Vision

With the development of machine vision theory and electronic information technology,machine vision technology has been widely used in various fields.Machine vision detection technology has the advantages of non-contact,high precision,high efficiency and real-time.At present,the quality detection of the mesh is manual detection,this method has low detection accuracy and low efficiency,which greatly reduces the automation level of mesh production.In this paper,machine vision detection technology is applied to the mesh quality detection,and the method of mesh quality detection based on machine vision is deeply studied.The main work and conclusions are as follows:Firstly,this paper studies the three aspects of image filtering,image segmentation and edge detection,and discusses the theory of several common algorithms and the methods of parameter selection.Through experimental comparison,the processing results and time of the algorithm are considered comprehensively,the median filtering algorithm,Otsu algorithm and Canny algorithm are finally selected to achieve image preprocessing.Secondly,in order to complete the dimension measurement for each mesh,each mesh region is segmented by the connected domain analysis method.Based on the research of Hough Transform,according to the characteristics of the research object,an improved Probability Hough Transform algorithm is proposed,by segmenting the mesh image and using the gradient direction information of the pixel to limit the parameter range of the line detection.And the improved Probabilistic Hough Transform algorithm is used to detect the line of each mesh,the detected linear information is used to calculate the mesh size.The experimental results show that the improved Probability Hough Transform algorithm proposed in this paper improves the linear detection speed and dimensional measurement accuracy.Then,according to the defect characteristics,the mesh defects are divided into two categories: mesh defect and node defect,and the mesh defect detection method and mesh node segmentation method are respectively proposed.The geometric features and texture features of the two types of defects are extracted respectively,and the method of Principal Component Analysis is used to select features and reduce the dimension of defect features.The two pattern recognition algorithms of BP Neural Network and SVM are studied in this paper,the improved learning algorithms of BP Neural Network and the method of parameter selection are discussed in depth.The BP Neural Network classifier based on improved algorithm and SVM classifier are designed to conduct a comparative experiment.The experimental results show that the BP Neural Network classifier based on Levenberg-Marquardt algorithm has higher classification accuracy and stronger generalization ability.Finally,the design and implementation of the vision inspection system of mesh quality is completed,including the selection of key hardware and the construction of the detection platform,and the design of the software modules and system construction.The experiment has verified that the system has greatly improved in measurement accuracy,detection accuracy and detection speed,and meets the actual requirements of the project.In summary,this paper studies the mesh quality detection technology based on machine vision,and deeply discusses the mesh size measurement method and the mesh defect detection and classification method,and proposes a detection algorithm with fast detection speed and high accuracy.The set of mesh quality visual inspection system realizes the automatic detection of the mesh quality and improves the intelligent level of mesh production.