活立木真彩色图像降噪方法研究

Study on the Method of Noise Reduction on Color Images of Standing Trees

为了更好的消除真彩色图像噪声,提高图像的质量,本文提出了一种基于HSV空间状态下的图像降噪的方法。根据HSV各颜色分量相互独立的特性,将真彩色图像从RGB空间转换到HSV空间;通过真彩色图像的矢量值降噪技术与灰度图像的标量值降噪算法相结合的改进算法,分别对每一颜色分量进行处理。以峰值信噪比为判断依据进行对比实验,结果表明:相对于矢量中值滤波,改进后的算法对真彩色图像的降噪具有更好的效果。

In order to effectively eliminate the noises of true color images and improve image quality,the method based on the HSV color space for noise reduction on color images of standing trees was proposed.According to the characteristics that the each color components of HSV was independent of each other,the true images can be transformed from RGB to HSI.The each color component was processed separately in our approach through the improved algorithm of combining noise reduction based on the vector magnitude and the de-noising algorithm based on scalar magnitude.Taking peak signal-to-noise ratio （PSNR） as the basis for comparison experiments.The results showed that our proposed algorithm has superior effects on noise reduction of color images compared to vector median filtering.