在图像去噪时,存在着一个如何兼顾降低噪声和保留细节的难题,为此本文提出边缘保持自适应平滑算法——灰度最小方差区域的均值去噪.该算法提高了图像平滑过程中边缘检测的方向性,在9个方向上对边缘细节进行检测类聚,具有较强的方向性和自适应性.利用该算法进行仿真试验,结果表明,与已有的去噪算法相比较,该算法较好地解决了边缘信息保持和去噪的矛盾,去噪效果也明显增强.
There exists a problem of how to give attention to retaining the details while image denoising. In this paper, a new adaptive algorithm of LVAM is proposed, which improves edge detecting orientation. It clusters image's edge from 9 directions and has stronger orientation and adaptability. By using the algorithm to simulate and comparing the results of eliminating noise, this algorithm can be used to solve the contradictions between retaining the edge information and denoising and the effect of image denoise will be greatly strengthened.