Machine Vision Based Veneer Online Sorting Detection System

With the rapid development of China's manufacturing industry,plywood is widely used in various fields of social production and plays an important role in the national economy.As the raw material of plywood,the dimension error and surface defect of veneer will reduce the quality of plywood products.In order to avoid the quality problems of plywood caused by dimension errors and surface defects,this paper relies on a project of a timber machinery enterprise in Shandong Province,based on machine vision inspection technology and the advantages of fast and accurate machine vision inspection technology,taking the dimension and surface defects of veneer as the research object,the construction of the inspection system,the detection of the dimension of veneer and the recognition and classification of the surface defects of veneer were studied in depth.A good detection system is the basis of recognition and detection.Aiming at the dimension and surface defect types of veneer,the overall structure of the system is designed,and the hardware platform of the system is built.The combination of dark field illumination and backlight illumination assists the CCD camera in capturing the veneer image.This method can highlight the outline and surface features of the veneer and contribute to the detection and identification of the veneer detection system.In order to reduce the interference of environmental factors on the quality of acquired images,a histogram equalization method is used to highlight the required image features.On the basis of image histogram equalization,combined with several commonly used image edge segmentation algorithms for edge detection,the best Canny segmentation algorithm is selected according to the actual segmentation effect,and the veneer image circumscribing polygon method and edge fitting extraction method are adopted.The combined method is used to process and calculate the segmented image,and the accurate detection of the veneer dimension is realized.Aiming at the recognition and classification of veneer surface defects,the whole classification framework was established by using the bag of words model theory.The improved FREAK algorithm and the improved SURF algorithm were used to extract the veneer defect features respectively.The effect of rotation angle on feature extraction isexplored and the effectiveness of the improved algorithm is verified.The extracted features are clustered using the K-means++ clustering method to construct a visual dictionary.In order to obtain the optimal detection results,the parameters of the kernel function in support vector machine(SVM)were optimized,and the surface defects of veneer were recognize and classified by using support vector machine classifier.The classification results of the two improved algorithms under Gauss radial basis function were obtained.According to the classification results,it is found that the two improved algorithms can effectively recognize and classify the veneer image.The improved SURF algorithm has higher average recognition rate and the improved FREAK algorithm has faster average running speed.According to the selected algorithm and parameters,the software part of the online sorting and detection system is designed.The test results show that the design scheme and algorithm can effectively solve the problem of online recognition and sorting of veneer,meet the needs of actual production,and have a certain application value.