摘要
金相组织图像分析是显微图像分析中的一个重要内容。由于噪声的存在会对图像分析带来很大的误差,因此进行图像分析前需要对金相图像进行滤噪处理,从而使图像的细节更加突出,便于目标识别。针对标准均值滤波方法存在的不足,提出自适应加权均值滤波方法。该方法通过检测确定图像中的脉冲噪声点,并用改进的均值滤波方法对检测出的噪声点进行滤波。实验结果表明,自适应加权均值滤波能在有效地去除噪声的同时,较好地保护图像细节,较标准均值滤波具有更优良的滤波效果,而且可与更大窗口的中值滤波效果相媲美,其处理速度比大窗口的中值滤波快。
Image analysis of metallographic plays an important role in micro image analysis. Due to the existence of the noises, it brings some errors of image analysis. Therefore, the noises in the metallographic must be deleted before analyzing the image, which will not only make the details of the image obvious, but also make the target identified more easily. To overcome the drawbacks of the standard average filter method, an improved method of the adaptive, weighted and average filter was proposed. The impulse noises of image can be detected by applying this method, and tee found noise is filtered by means of the improved method of average filter. Experimental results show that the method can remove noise effectively while preserving fine details well and has better filter performance than the standard average filter method. It has the same performance as bigger region median filter effect and can obtain the faster process speed than big region median filter.
出处
《辽宁石油化工大学学报》
CAS
2005年第4期58-61,共4页
Journal of Liaoning Petrochemical University
关键词
噪声检测
自适应加权均值滤波
中值滤波
Noise detection
Adaptive, weighted and average filter
Median filter