一种自适应的图像双边滤波方法

提出一种利用双边滤波的图像平滑滤波方法,即在滤除图像中高频噪声的同时,按照图像亮度变化保持图像中处于高频部分的边缘信息的自适应滤波过程。该滤波方法将传统的Gauss滤波器的权系数优化成Gauss函数和图像的亮度信息乘积的形式,优化后的权系数再与图像作卷积运算。这样,滤波时就可以考虑到图像的亮度信息,在滤除图像噪声的同时尽量保持了图像的边缘。由于双边滤波的方法可以使滤波器的权系数随着图像的亮度变化而改变,所以在滤波过程中能达到自适应滤波的目的。

An adaptive bilateral filtering method for image processing

A bilateral filtering method for image processing, i.e. an adaptive filtering process that filters high-frequency noise of image and simultaneously maintains the high-frequency edge information of image based on image brightness variations is proposed. The filtering method optimizes the weight coefficient of Gaussian filter into the product form of Gaussian function and image brightness information. The optimized weight coefficient will be applied to convolution of image. In this way, image brightness will be taken into account and the image edge should be simultaneously maintained as possible during filtering image noise. Since bilateral filtering method can make the weight coefficient of filter vary with image brightness, so the adaptive filtering effect can be attained during the filtering process.