Preprocessing of color image based on digital camera

Abstract

To realize color image segmentation,target identification,feature extraction,and other image processing projects,the image preprocessing methods are researched.Based on analysing characteristic of digital camera and human visual error and understanding data formats of color image,an octree arithmetic is applied to color quantization and a vector median filter is used to eliminate noise to realize preprocessing for color images.In combination with a experiment,this paper discusses and analyzes window parameters of vector median filter,and presents an effective evaluation methods by combining color information similarity theory.Finally,a reasonable color quantization method and window parameters are obtained for digital camera,and the chromatism can be controlled within 3 units.Experimental results show that octree color quantization and vector median filter can perform well in preprocessing of color image for digital camera.