利用物体的归一化偏差本体反射特征进行彩色图像的分割

SEGMENTATION OF COLOR IMAGES USING NORMALIZED DEVIATION BODY REFLECTION

本文提出了一种基于物体的归一化偏差本体反射特征进行彩色图像分割的方法。通过对成像过程的分析,我们应用了光照和表面反射率的有限维线性模型(Finite-Dimensional Linear Model)和双色反射模型(Dichromatic Reflection Model),将光源因素以及几何因素分离出来,经过做归一化处理后的偏差光谱反射率已消除了光照和几何条件的影响,代表物体固有的颜色特性,因此可做为分割和识别物体的可靠依据。在实验图像中同时存在着耀斑、影调和暗区等区域,为此采用初始分割对这些区域分别处理。此外,本文还通过采用将区域生长与边缘约束相结合的方法,进一步提高了分割的可靠性。

This paper presents a segmentation method of color images, which is based on the normalized deviation body reflectance of the objects. Normalized deviation reflectance eliminates the influence of the illumination and the geometry conditions. So far it represents the intrinsic color property of the object, and can be utilized as the reliable evidences for segmentation and recognition. There exists highlights and shading areas and shadows in the experiment images. Therefore we process these areas respectively through the initial segmentation. In addition, in order to improve the reliability of the segmentation, we combine the region-growing method with the edge constrained method.