摘要
本文主要研究中值滤波器和Sobel算子的滤波性能,并采用流水线方式硬件实现了伪中值滤波和Sobel边缘增强。两种方法均可以实时地滤除图象中的噪声,并勾画出边缘。处理后图象信噪比可提高约7dB,输出延迟小于200μs。此外,本文还提出一种新的噪声乎滑方法。该方法与中值滤波器一样可以保护图象边缘,滤除脉冲颗粒噪声,而且,它抑制随机噪声的效果优于中值滤波器。
In this paper the features of the median filter and the Sobel operator are mainly discussed. A pseudo median filter and a Sobel edge sharpening are realized in hardware by use of the water-flow working fashion. The results show that the two methods can filterout image noises and draw the outline real-timely. The filtered image's SNR increases about 7dB. The time delay is less than 200 μs. In addition, we put forward a new method for noise smoothing. It can protect the edge of images and filter out the pulse noises just as the median filter but better in filtering out the stochastic noises.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1989年第3期120-121,124,共3页
Acta Electronica Sinica
