一种新的彩色图像特征检测中相似度计算方法

A New Method to Calculate Similarity in Color Image Feature Detection

现代数字图像处理技术中处理的大多是彩色图像。很多处理场合都需要直接实时地对彩色图像进行处理,这样就可以避免彩色图像灰度化过程中的细节损失,也有利于提高算法速度。在基于色彩分布的彩色图像特征检测过程中,需要对色彩空间进行颜色的量化,也需要对不同像素点的颜色值进行比较,并作出合适的相似性和异质性计算。论述了一种新的色彩距离计算方法,大大提高了彩色图像特征提取的准确率与速度。

Color images is always processed in digital image processing, and in most cases have to process color images then can avoid to change a color image into a gray one and also can avoid the error in that process. The color values are need to quantization and calculated the distance of different colors. This is the most important work, without it can not compare the differences between pixels. A good method to describe the distance between colors based on feature detection in color images is given. It enhances the accuracy in feature detection and also can be used in detecting the similarity of color images.