基于能量最小的拉普拉斯流域分割算法

针对拉普拉斯流域分割边界处分割效果不好的缺点，根据能量区域集中性原理,定义了一个差异度函数。当两个区域间的差异度大于它们其中任何一个,则认为它们之间有边缘存在,否则它们属于同一个区域,应进行合并。通过在vs2008环境进行仿真，对蝴蝶图像进行分割,最终分割成蝴蝶和三个背景区域,不同的区域用不同的颜色表示。实验表明该算法分割效果较好,速度也比较快。

Laplace watershed segmentation based on minimum energy

Because laplace watershed segmentation on image edge is not very effective. Based on the energy regional concentration theory, the authors defined a function to measure area differences. When the difference between two regions is greater than any one of them, this is a border lie between the two regions; otherwise they belong to the same region, they should merge into one. Through the experiment under vs2008, the authors split butterfly specimens into four parts, marked different regions by using different colors. Experiment demonstrates the method has good effect, and also less time-consuming.