基于OTSU准则及图像熵的局部递归分割算法研究

在图像分割中，门限的选取至关重要，本文在Otsu门限选择准则的基础上，研究了一种基于图像熵的局部递归的门限选择及分割算法，该算法在图像背景不均或图像不是简单的单峰、双峰图像的情况下可以有效的分割，分割后图像的细节更加丰富，有利于分割后的特征提取。对几种不同的目标进行了实验，获得了较好的实验结果。

Partial Recursive Segmentation Algorithm Based on Otsu and Image Entropy

In image segmentation, the threshold selection is very important. In this thesis, a partial recursive algorithm of threshold selection and segmentation is put forward, which is based on the Otsu threshold selecting method. With this algorithm we can segment the image effectively even if the image is uneven and is not the single-modal or bimodal one. The segmentation result has more detail, which is good to the feature-extraction. Finally we make experiment with some sorts of images and obtain good result.