基于HSI颜色空间噪声检测的彩色图像去噪算法

Color Image Denoising Algorithm Based on HSI Color Space Noise Detection

目的为了有效去除彩色图像中的椒盐噪声,提高彩色图像质量。方法采用椒盐噪声检测和中值滤波相结合的方法,提出一种基于HSI颜色空间噪声检测的彩色图像去噪算法。将图像转换到HSI颜色空间,根据椒盐噪声在S通道具有极大值或极小值的特点判断出可疑椒盐噪声的位置,在H通道、I通道将可疑椒盐噪声分为噪声点和有用信号,对检测出的噪声像素进行中值滤波去噪。结果采用文中算法去噪后,验证图像主观评价值(Z)为1.30,平均PSNR为37.54,SSIM为0.99,Entropy为7.31,在主客观评价上优于现在常用算法。结论文中提出算法可以为彩色图像椒盐噪声的去噪提供理论基础,具有一定的实际应用价值。

The work aims to effectively remove salt and pepper noise in color image and improve the quality of color image.A color image denoising algorithm was proposed based on HSI color space noise detection by combining salt and pepper noise detection with median filtering.The image was transformed into HSI color space,and the position of the suspicious salt and pepper noise was judged according to the characteristics of noise having maximum or minimum value in S channel.In H channel and I channel,the suspicious noise was divided into noise points and useful signals,and the detected noise pixels were denoised by median filtering.The subjective Z-score of denoising verification image was 1.30,the average PSNR was 37.54,SSIM was 0.99 and Entropy was 0.73,which was better than the commonly used algorithm in subjective and objective evaluation.The algorithm proposed can provide theoretical basis for the removal of salt and pepper noise in color images,and has certain practical value.