关键词: 木材缺陷检测, 小波, 数学形态学, 区域生长

Abstract: Wood defect detection is a very important step in wood processing.In order to implement automatic detection，a method based on wavelet transform and mathematical morphology is presented in this paper.The method uses wavelet transform to suppress the interference information.Then bottom-hat transform of morphology and threshold processing are applied to get seeds in the wavelet reconstructed image.Eventually the wood defect can be segmented through region-growth algorithm.Experiments prove that this method can detect wood defect effectively and accurately and meet the requirement of wood processing.

Key words: wood defect detection, wavelet transform, mathematical morphology, region-growth