摘要： 提出一种新的图像空间特征提取方法，采用基于颜色向量角和Euclid距离的CCVAE算法，检测原始图像的彩色边缘，分别对边缘像素、非边缘像素建立边缘直方图描述子和空间颜色直方图，描述颜色分布信息。仿真实验结果表明，该方法具有较高的准确率和召 回率。

关键词: 空间特征, 彩色图像边缘检测, 空间颜色直方图, 颜色向量角

Abstract: This paper proposes a novel approach for image spatial feature extraction. It uses CCVAE algorithm based on color vector angle and Euclid distance to detect original image color edge. For the edge pixels, the Edge Histogram Descriptor(EHD) is built. For the non-edge pixels, the Spatial Chromatic Histogram(SCH) is built. Experimental results show that the approach has high precision rate and recall rate.

Key words: spatial feature, color image edge detection, Spatial Chromatic Histogram(SCH), color vector angle

中图分类号: 

• TP391