一种基于视觉模型与连通域统计的阈值分割新算法

该文提出了一种新的基于视觉侧抑制模型和连通域统计的阈值图像分割算法.该方法模拟人眼处理图像机制,建立侧抑制模型,提取出图像中相对变化的信息,并计算变化量矩阵的直方图,然后对该直方图自下而上进行水平切割,并计算割线上的连通域数量,绘制连通域曲线图.仿真实验表明,该方法具有很好的亮度、对比度不变性和线条连通性,分割结果能更好的符合视觉感受.

A Novel Threshold Algorithm about Image Segmentation Based on Vision Model and Statistics of Consecutive Fields

This paper proposes a novel algorithm about image segmentation based on the vision lateral inhibition model and statistics of consecutive fields.The algorithm simulated the mechanism of the retina model of human eyes.First it picked up the variance information in the image,and then used level secant to cut the histogram from bottom to top.The algorithm counted the consecutive fields on every secant and compared each secant position and consecutive field numbers,and then made a consecutive curve.It can be proved that the secant corresponding to the curve's peak contains the boundary of the biggest consecutive field,which is the optimized threshold.The algorithm is very effective and robust.It is not only well adaptive to the change of brightness and contrast,and also can be applied to much complex background.It can improve the connectivity of strokes and lines.The result of segmentation is well coincident with the vision feeling of human eyes and the main reasons are discussed.