基于自适应预处理的图像分割方法ImageSegment-电子与信息学报.PDFVIP

第 29 卷第 1 期 电 子 与 信 息 学 报 Vol.29No.1 2007 年 1 月 Journal of Electronics Information Technology Jan. 2007 基于自适应预处理的图像分割方法 沃 焱 韩国强 张见威 (华南理工大学计算机科学与工程学院 广州 510641) 摘 要: 为了防止分水岭算法过分割问题,该文提出了一种基于自适应预处理的图像分割算法。该方法在分水岭算 法的基础上，首先结合像素点亮度特征和空间分布特性应用自适应方法对梯度图像进行预处理。通过考察各像素 点邻域中像素分类后的分布情况，来判断考察点是处于区域中心还是处于边界，并据此对考察点的梯度值进行调 节。然后在预处理后的梯度图像上选定标记，将预处理后的梯度图像中大于200个像素的连通区域标定为标记。最 后用分水岭分割方法对带标记的参考图像进行分割。试验结果表明，该分割方法具有良好的分割效果。 关键词: 图像预处理；图像分割；分水岭变换 中图分类号：TP391 文献标识码：A 文章编号：1009-5896(2007)01-0087-05 Image Segment Algorithm Based on Adaptive Image Preprocessing Wo Yan Han Guo-qiang Zhang Jian-wei (College of Computer Science and Engineering, South China University of Technology, Guangzhou 510641, China) Abstract: To avoid over segmentation produced by watershed segment algorithm, an image segment algorithm based on the adaptive preprocessing is proposed in this paper. This algorithm is based on the watershed transform. First the gradient image is preprocessed according to the pixels intensity and spatial information. After the points in the neighborhood of each image pixel are classified，their distribution situations are used to estimate the location of the image pixel, in the center or near the boundary of a region. Each pixel’s gradient magnitude is adaptive adjusted according to its location. Then the flat regions larger than 200 pixels in the preprocessed gradient magnitude image are extracted as markers. According to the markers, the preprocessed gradient magnitude image is segmented by the watershed transform. Experimental results indicate that this algorithm provides a good segmentation for segment-based image coding. Key words: Image preprocessing; Image seg