最大熵原理在非结构化道路图像识别中的应用

Image recognition-based maximum entropy on unstructured road

根据非结构化道路环境的特点,结合最大熵理论在图像处理中的运用,提出了一种基于最大熵理论的非结构化道路图像识别的算法。对于以往划分道路时,阴影区域和道路与非路交界区域容易出现分类错误的现象,通过两次最大熵分类,把这些区域重新划分,最终将道路与周围的环境区分开来。同时满足实时性的要求。

Combining the features of unstructured road and the maximum entropy theory used in image processing, an algorithm based on the maximum entropy theory is proposed to classify the unstructured road. In the former algorithms, it is difficult to distinguish shadows from the unstructured road, and there are always some mistakes at the boundary of the road and surroundings. So by using maximum entropy classification twice and combining these regions, roads and surroundings can be recognized successfully. Meanwhile, it satisfies the real-time demand.