基于色彩滤镜阵列的织物图像莫尔条纹消除算法

发布时间：2018-08-25 13:31

【摘要】：针对织物普遍具有的细密栅栏特性的规则性干扰,以致图像中产生明显的波纹状彩色条纹图案的困扰,给服装的展览、宣传带来不便的难题,研究了织物图像中莫尔条纹的消除问题。而传统的基于JPEG图像的重构算法,在去除条纹时会降低图像的色彩质量,影响服装的整体美感。根据对织物图像中莫尔条纹的形成机理和数码相机成像过程中Bayer CFA色彩滤镜阵列的排列特性进行分析,提出了一种先检测莫尔条纹区域,进而采用色彩插值补偿的一种莫尔条纹消除算法。算法首先对织物数码图片的原始数据G分量进行水平、竖直方向的高通滤波,模拟CFA采样过程的混叠现象,从而检测出图片中含有莫尔条纹的区域。然后在莫尔区域分别再对R、B分量进行处理,最终实现了莫尔条纹的消除。实验结果表明,与传统方法相比,通过提取图像的G分量信息进行处理,能在不影响图像色彩质量的同时去除莫尔条纹,更具实用性。
[Abstract]:In view of the regular interference of the fine fence characteristic of the fabric, there are obvious corrugated color stripes in the image, which brings inconvenient problems to the clothing exhibition and propaganda. The elimination of moire fringes in fabric images is studied. The traditional reconstruction algorithm based on JPEG image will reduce the color quality of the image and affect the overall beauty of clothing. Based on the analysis of the formation mechanism of moire fringes in fabric images and the arrangement characteristics of Bayer CFA color filter array in digital camera imaging, a new method for detecting moire fringe region is proposed. Then a moire fringe elimination algorithm based on color interpolation compensation is proposed. The algorithm first performs horizontal and vertical high-pass filtering on the G component of the original data of the fabric digital image, and simulates the aliasing of the CFA sampling process, so as to detect the region containing moire fringes in the image. Then the R _ (B) component is processed in the moire region, and finally the moire fringe is eliminated. The experimental results show that, compared with the traditional method, the moire fringe can be removed without affecting the color quality of the image by extracting the G component information of the image, which is more practical.
【作者单位】：辽宁公安司法管理干部学院;
【分类号】：TP391.41

【相似文献】