Position: Home > Articles > Study on the Denoising from a Mine Image of Adaptive Non-local Mean Filtering Algorithm Based on Gradient Information
Journal of Shandong Agricultural University(Natural Science Edition)
2018
(6)
1063-1067
梯度信息自适应的非局部均值滤波算法对矿井图像降噪研究
作 者：
张雷
单 位：
广东科贸职业学院
关键词：
八方向Prewitt算子;自适应性;非局部均值滤波算法;矿井图像处理
摘 要：
为了提高非局部均值滤波算法对矿井图像的降噪效果,本文将改进的非局部均值滤波算法与八方向Prewitt算子相结合,通过梯度信息对滤波参数进行优化,赋予了该非局部均值滤波算法较强的自适应性。另外,八方向Prewitt算子增加了非局部均值滤波算法搜索窗口的移动方向,提高了搜索相似像素点的效率。本文算法在对人工加噪的图片进行处理时,所需时长仅为2.38 s,与经典的非局部均值滤波算法相比缩短了40%;本文算法处理所得图像的信噪比为26.473,是经典的非局部均值滤波算法的1.2倍。通过对处理所得图像质量进行对比发现,本文算法对矿井图像的降噪效果更为明显,且图像中边缘和纹理的细节信息更为丰富,可用于矿井检测图像的处理。
译 名：
Study on the Denoising from a Mine Image of Adaptive Non-local Mean Filtering Algorithm Based on Gradient Information
作 者：
ZHANG Lei;Guangdong Polytechnic of Science and Trade;
单 位：
ZHANG Lei%Guangdong Polytechnic of Science and Trade
关键词：
Eight direction Prewitt operator;;adaptability;;non-local mean filter algorithm;;mine image processing
摘 要：
In order to improve the denoising effect of non-local mean filtering algorithm for mine images, an improved non-local mean filter algorithm was employed combined with eight direction Prewitt operator. The improved non-local mean filter algorithm showed excellent adaptability originated from the gradient information. In addition, the moving directions of the searching window of the non-local mean filtering algorithm were added by the eight direction Prewitt operator, which improved the efficiency of searching similar pixels. When the improved algorithm was employed to deal with artificial noised picture, the required time was only 2.38 s, which was about 60% of that of the classical non-local mean filtering algorithm;The PSNR of the resultant picture obtained by the improved non-local mean filter algorithm was 26.473, which was 1.2 times of that obtained by the classical non-local mean the filtering algorithm. By comparing the obtained image quality, the improved algorithm showed better denoising effect, and maintained more detail information of edge and texture. It could be used for the treatment of mine detection image.
导出引用
计量
文章访问数: 7
HTML全文浏览量: 0
PDF下载量: 0
所属期刊
