The problems concerning veneer knot defect repair in China’s veneer manufacturing process and the important role of wood texture in industrial production are analyzed. In view of the problems concerning defect image repair of veneer with texture, a new repair method based on image decomposition is put forward. First, a denoising algorithm based on minimum total variation is used for veneer image decomposition to extract the structure and texture parts of veneer with texture defects;then an advanced BSCB algorithm is used to repair the structural image of veneer and a Criminisi algorithm with extension based on sample blocks is used to repair the texture part. Finally, superposing and synthesis of the repaired images are conducted. The effectiveness of this method is verified through experiment demonstration, with the experimental result showing that repair result of the defect images of veneer with texture with this method is superior the repair result with a single method.% 分析了我国单板生产过程中单板节子缺陷修补存在的问题和木材纹理在工业生产中的重要作用，针对带纹理的单板缺陷图像修补问题，提出了一种新的基于图像分解的修复方法。首先采用基于最小化总体变分的去噪算法对单板图像进行分解，提取出带纹理缺陷单板的结构部分与纹理部分；然后对单板结构图像采用改进的BSCB算法进行修复，对纹理部分采用扩展基于样本块的Criminisi算法进行修复，最后将修复好的图像叠加合成，通过实验验证了该方法的有效性。实验结果表明，该方法对带纹理单板缺陷图像的修复效果要优于采用单一方法的修复效果。
展开▼
