边缘宽度细化与交叉皮层模型结合的图像分割算法

An image segmentation algorithm based on combination of ramp width reduction and intersecting cortical model

针对弱边缘图像分割中交叉皮层模型产生目标几何尺寸失真等问题,提出了一种基于边缘宽度细化与交叉皮层模型结合的图像分割算法。该算法借助prewitt边缘算子和边缘斜坡模型,细化图像边缘宽度。在此基础上,利用二维直方图将交叉熵扩展至二维空间,以获得交叉皮层模型的最优分割阈值。实验结果表明,该算法不仅能克服边缘模糊的影响,对弱边缘图像进行精确分割,而且处理速度也大幅提升了。

Weak-edge image segmentation based on intersecting cortical model （ ICM ） always leads to targets ’ geometric dimension. In order to solve the problem, an image segmentation algorithm based on combination of ramp width reduction and ICM is proposed. In the algorithm, ramp width is reduced with the help of prewitt operator and ramp edge model. Besides, the 1-D cross entropy is made into the 2-D by the 2-D histogram, so as to gain the opti_mal threshold of ICM. Results show that not only can the novel algorithm conquer blur contour and get precise segmen_tation, but also the algorithm efficiency is improved dramatically.