木材彩色图像缺陷分割-计算机工程与应用.PDF

Computer Engineering and Applications 计算机工程与应用 2013 ，49 （18） 153 木材彩色图像缺陷分割 ——基于Gabor 滤波的改进C-V 彩色模型 仇逊超，王阿川，曹 军 QIU Xunchao, WANG Achuan, CAO Jun 东北林业大学 信息与计算机工程学院，哈尔滨 150040 Information and Computer Engineering College, Northeast Forestry University, Harbin 150040, China QIU Xunchao, WANG Achuan, CAO Jun. Segmentation for colored image of wood defect by improved C-V colored model of Gabor filter. Computer Engineering and Applications, 2013, 49 （18）：153-158. Abstract ：Through analyzing the features of wood defects and the features of veneer knots, this paper presents an identification algorithm of wood defects, which is based on improved C-V colored model of multi-channel Gabor filters. This algorithm, which protects the colored information of the image, makes the colored image be an integral image. The algorithm applies multi-channel Gabor filters and K -means clustering to getting colored image, which is the discrimination of the defective target and the back- ground. The algorithm applies the improved C-V colored model on the new image to abstracting the new image ’s edge. To com- pare with the identification algorithm of gray image defects, which is based on improved C-V model and wavelet transformation, the results show that this method can segment the colored image of wood defects and the colored image of the multi-object veneer knot fast and accurately. Key words ：colored image of veneer knot; colored extension of C-V model; multi-channel Gabor filters; multi-object segmentation 摘 要：分析了木材节子缺陷、单板节子的特点，提出了一种基于多通道Gabor 滤波的改进C-V 彩色模型的木材缺陷识别 算法。该算法将彩色图像作为一个整体的图像，保留了图像的彩色信息。该算法利用多通道Gabor 滤波器、K -均值聚类算 法得到缺陷目标与背景的彩色区分图像；利用改进的彩色C-V 模型对新图像进行边缘提取，得到理想的实验结果。与采 用基于改进C-V 模型与小波变换的灰度图像缺陷识别算法相对比，结果表明该方法可快速、准确地实现对木材节子缺陷 彩色图像及单板多节子彩色图像的分割。 关键词：单板节子彩色图像；改进的彩色C-V 模型；多通道Gabor 滤波器；多目标分割 文献标志码：A 中图分类号：TP391.41 doi ：10.3778/j.issn. 1002-8331.1112-0357 1 引言