基于多分辨率-分水岭算法的图像分割方法

提出了在小波多分辨率域中使用分水岭算法对图像进行区域分割和融合的新方法,该方法首先使用小波变换理论将原始图像变换为不同层次的金字塔多分辨率图像；然后通过分水岭算法获得最低分辨率下的分割图像；最后。利用逆小波变换将分割的图像映射到原始分辨率上．从而获得分割图像．实验结果表明：此方法可以大大减少噪声存在下的过分割现象．

Image Segmentation Based on Wavelet-Based Watershed

This paper presents a wavelet-based watershed image segmentation method for image segmentation,This method is based on a multi-resolutlon application of wavelet and watershed transform, followed by a wavelet coefflclents-based region merging procedure. First, pyramid presents is based on wavelet transform create several lower resolution images from original image; The lowest-resolution image then is segmented into many regions based on watershed algorithm; Finally, The region projection is used to recover a full-resolution image by inverse wavelet transform. Experimental result shows that this method can he applied to the segmentation of noise images and reduction of over-segmentation.