A Machine Vision System For Detection Of Bamboo Block

With the rise of the global low carbon economy and the idea of "green,environmental protection and sustainable development".Bamboo products more and more favored by the world's consumers.It is inefficient and denormalized to use workers' eyes to filter bamboo in many manufacturers who produced mahjong mat in China.So an automatic defect detection system is designed by combining with machine vision technology to improve production efficiency,increase the income of enterprises.The paper's main work as following:(1)Design of bamboo block screening device.The device is improved based on the research of machine vision system and the existing mechanical transmission system.The experimental results show that the device has the advantages of fast response speed,and can be used for the screening of defective bamboo blocks.(2)Research and design of image acquisition system.The image acquisition system is designed for the surface defect detection of bamboo block based on the research of illumination and imaging system and the selection of related equipment is completed.(3)A new image enhancement algorithm is proposed based on the difference of HE and HSV.In this method,The weighted difference of H,S space is enhanced by the histogram equalization and then using median filtering algorithm to reduce the contrast enhancement effect caused by histogram equalization.Because the brown spot defect is high contrast only in the H space,white spot defect is high contrast only in S space,but the new space has high contrast both in white spot defects and brown spot defects.Therefore,the subsequent defect detection algorithm only needs to deal with the new space,so as to reduce the running time of the detection algorithm and meet the real-time requirements of the system.(4)Research of positive and negative identification of bamboo block.Methods compared with the existing literature on the application of machine learning bamboo block identification of positive and negative,this paper designs a simple and efficient bamboo block identification algorithm,the two types are the opposite of bamboo blocks using S threshold and Laws horizontal projection method based on energy judgment.The algorithm can accurately identify the positive and negative sides of the bamboo block,and reduce the influence of the texture characteristics of the negative side of the bamboo block on the surface defect detection of the bamboo block.(5)Study on bamboo block face surface texture features,respectively in the spatial region growth method and application in frequency domain using Fourier transform method and the use of support vector machine learning method to detect the surface defects of the bamboo blocks,analyzed and compared three kinds of methods.(6)The software platform of mixed programming based on MFC and Opencv library was completed,and the whole surface defect detection system of C++ was tested.The experimental results show that the system can run stably for a long time and the detection rate can be up to 10,000 hours per hour,and the defect detection rate of bamboo block can reach more than 97%.The system has reached the technical standard of practical industrial application and has a good prospect.