The Design Of Bamboo Surface Defect Detection And Color Classification System Based On Machine Vision

In China,bamboo is rich in resources and has renewable ability,so bamboo has been widely applied in many aspects,especially in bamboo furniture,bamboo cutting board and so on.bamboo will be damaged by worms and mildewed in the procedure of growth and storage.Bamboo can easily split in the process of breaking.Because bamboo has different water content and drying time,it shows different color shade.The integrity and the color consistency of bamboo surface are important measure standards for product quality.So it is necessary to pick out the defecting bamboo and classify them according to the color shades.Currently,the industry of bamboo products selects the bamboo defects and classifies them by manual.This method has low efficiency,and will lead to visual fatigue and low product quality in a long time.This paper aims to design a system of the defect detection of bamboo surface and color classification based on machine vision.This system can achieve automatic picking of bamboo strips instead of manual picking.This paper mainly studied the methods of surface defect detection and color classification based on machine vision,and the overall schemes of surface defect detection and color classification of bamboo were designed.This paper mainly discussed the disfigurement detection of wormhole,detection of Crack,defect detection of green bamboo surface and the procedure of realization of the.color classification algorithm that based on CIEDE2000 color difference,and then we analyzes the test results.The main content of this paper are as follows:1)This paper mainly studied the light source type and the lighting scheme of the light source,and designed the light source illumination system for the defect detection of bamboo surface and color classification.There are kinds of combined manners of the light source and light source irradiation.The suitable illumination system can highlight the defect characteristics.Considering the needs of a wide range of lighting and bamboo color classification,D65 standard light source is chosen,which can simulate artificial daylight and has high color rendering index.2)In this paper,it is studied on the surface defect detection algorithm and designed the surface defect detection system based on machine vision.The types of defects mainly include wormholes,green bamboo surface,cracks and et.al.The characteristics of the defects are firstly analyzed in the process of algorithm design.Due to the fact that the wormhole area is similar to the background,Otsu threshold,flood fill and mathematical morphology are utilized to detect wormholes.According to the abrupt changes of the gray values of the crack defect,the improved Canny operator is used to detect the crack by adaptively adjusting the upper and lower threshold.According to the projection distribution of green bamboo surface in the X direction,the variances of bamboo strip width are regarded as the detection features.3)In this paper,there are studies about several kinds of color difference formulas which are commonly used in industry,and the color classification system based on CIEDE2000 color difference formula is designed.In this system,comparing the color difference between experimental and control group,the minimum class of experimental is as the similar.Comparison of CIE94,CMC,CIELAB and CIEDE2000 color difference formulas,we choose the CIEDE2000 color difference formula which is more close to the human visual perception system to calculate the color difference.between bamboo bars.In order to improve the robustness of color classification and prevent the abnormal samples from affecting color classification,in this paper,the Gauss coefficient is used to adjust the color difference between experimental and control group.4)In this paper,it is designed for the software system of real-time detection of bamboo strips.The system includes the user interface,multithread image processing,and computer communication.