基于自适应加权中值滤波的二维Otsu图像分割算法

Two-dimensional Otsu image segmentation algorithm based on adaptive weighted median filter

为解决传统二维Otsu算法在含噪声较多的图像应用中分割效果较差这一问题,提出一种基于自适应加权中值滤波的二维Otsu图像分割算法。该算法首先利用一种新的自适应加权中值滤波对噪声图像中值滤波;然后将中值图像的二维直方图区域划分由四分法改为二分法;最后利用改进的二维Otsu算法对图像作精确分割。实验结果表明,该算法对灰度噪声图像具有更强的抗噪性且分割效果也更为理想。

The traditional two-dimensional Otsu image segmentation algorithm got ineffectively result in the image segmentation while the image contains lots of noise. To solve the problem, this paper proposed a new two-dimensional Otsu image segmenta- tion algorithm based on adaptive weighted median filter. This method firstly preprocessed the image with noise by a new adap- tive weighted median filter, then transferred the region division in two-dimensional histogram of median image from four parti- tions into two partitions, finally segmented the image accurately by an improved two-dimensional Otsu algorithm. The experi- mental results show that the proposed algorithm has better anti-noise performance and better segmentation result for image with