Research On Detection Technology For Surface Defects Of Wooden Boards

Wooden board is most needed in wood application, and the surface defect of wooden boards not only directly affects its strength, but also affects the level classification board. Therefore, researching on detection technology for surface defects of wooden boards has great significance for improving the value of wooden boards and speeding wood processing automation.On the basis of further study on status of detection technology for surface defects of wooden boards, this paper proposes an automatic detection system for surface defects of wooden boards based on machine vision. According to the surface defect detection requirements, the overall structure of the detection system is designed and technical specifications of the system are identified. Through analyzing the characteristics of surface defects of wooden boards, combined with the practical detection environment, the high-brightness linear light source, linear array CCD camera and High-speed image acquisition card with dual DMA buffer are selected and image acquisition lighting system is built.According to the characteristics of wooden boards surface defect images, the tendency item elimination method based on the curve fitting is proposed and removes the background trend effectively; the adaptive median filtering method is used to eliminate the noise and improve the signal-to-noise ratio of the defect feature; The image threshold segmentation method based on the mean and standard deviation to determine the threshold is proposed and obtains the binary image of defects, and the image denoising technology based on noise area is proposed and eliminates the noise which still exists in the binary image. Then according to the principle of feature extraction, defect features are extracted by seed filling method based on recursive. And the surface defects of wooden board are identified and classified depending on the feature set.The system realizes the automatic detection for surface defects of wooden board. The 200 samples with juncture, knot, sag and bubbling defects in four different defects types are tested and the results show that the average recognition rate can reach 93.5%.