图象分割算法的改进研究

针对传统基于图论的图象分割方法对噪声敏感以及计算复杂度大的问题,对传统算法进行了相应的改进,综合考虑像素的灰度信息和空间位置信息,计算权函数表达式时充分考虑到节点之间及节点与区域间的空间近邻关系.对比实验表明,该算法能够有效地从背景中把目标物体分割出来,并且当目标物和背景相近时,相比其他两种算法能够去除更多背景;该算法分割结果更接近于人眼视觉特征.

Improvement Researches on Image Segmentation Algorithm

In image segmentation, the traditional method of graph theory clustering is sensitive to noise and has computational complexity. So, an improved approach is proposed in this study, combine the information of pixel gray scale and spatial position, Spatial relations between two pixels and between pixel and area were considered when calculating weight function. Experiment with matched control showed that, this algorithm could distinguish an object from the background more effectively than traditional techniques; when the object is close to background, this algorithm can remove more background compare with the other two algorithms; The segmentation from this algorithm is more close to the visual characteristic of human being.