结合Mahalanobis距离的SUSAN彩色边缘检测方法研究

针对HSI颜色空间在图像处理和分析方面更加符合人眼的视觉特性，提出了一种结合Mahalanobis距离的SUSAN彩色图像边缘检测方法．该方法首先将彩色图像从RGB空间转换到HSI空间，然后以H、S、I为变量计算空间中两种颜色的色差距离，结合SUSAN算子对边缘进行检测．实验结果表明：此方法可以有效地检测彩色图像的边缘．在保留图像边缘方面，性能优于传统的彩色边缘检测方法．

SUSAN color-edge detection method combined with Mahalanobis distance

This passage selected Hue Saturation Intensity（HSI） as image processing and anal- ysis tool for its suitability. A novel SUSAN color edge detection method originated in HSI space was designed, on the foundation of presented Mahalanobis color distance measure ment. SUSAN color edge detection method, integrated with SUSAN operator, converted RGB color space to HSI space to detect edge, which takes H,S and I as variables to calculat- ed the distance of the two colors. The experimental demonstrated that the proposed image detection method could detect the color edge effectively. The new method achieved better performance on edge preservation, in comparison of the traditional color edge detection.