实时彩色图像自适应中值滤波算法的FPGA实现

Real-time Color Image Adaptive Median Filtering Algorithm Based on FPGA

摘要:

图像在采集、压缩、传输和接收的过程中，会受到脉冲噪声的影响，给后续的边缘检测、图像分割或目标识别等造成干扰。传统中值滤波将图像的RGB分量转为YCBCR分量进行滤波，该转换方法会损失一定的色度信息，且未将彩色信息完全分离，导致滤波后的图像产生色彩偏移。针对上述问题，利用HSI彩色空间中的I和S色彩分量对椒盐噪声进行检测，并通过自适应中值滤波将椒盐噪声滤除，图像还原度更高；利用FPGA对算法进行了硬件加速，在xc7a100tfgg484-2芯片上运行频率达222MHz，可实现2k@60Hz视频的实时处理。

Abstract:

In the process of image acquisition, compression, transmission and reception, it will be affected by impulse noise, which will cause interference to subsequent edge detection, image segmentation or target recognition. Traditional median filtering converts the RGB components of the image to YCBCR components for filtering. This conversion method will lose a certain amount of chrominance information, and the color information is not completely separated, resulting in a color shift in the filtered image. To solve the above problems, the I and S color components in the HSI color space are used to detect the salt and pepper noise, and the adaptive median filter is used to filter the salt and pepper noise. The image reduction is higher. The algorithm is hardware accelerated by FPGA. The system runs at 222MHz on the xc7a100tfgg484-2 chip, which can realize real-time processing of 2k@60Hz video.