复杂背景下的遥感图像特征识别算法

Remote Sensing Image Feature Recognition Algorithm under Complex Background Algorithm

针对复杂背景和可变光照下的静态彩色图像人脸检测,提出一种基于多目标优化和蚁群算法的遥感图像特征识别方法。首先将遥感图像特征选择转化成多目标优化问题,然后利用特征子集维数和识别率加权构造了目标函数,最后利用蚁群算法的全局寻优和正负反馈投机实现特征子集搜索,从而找到遥感图像中的最优特征子集,实现准确识别。仿真实验结果表明,算法能够很快的找到最优特征子集,消除无用和冗余特征,降低了特征维数,识别率高。

For static color images under complex background and variable light face detection, a kind of multi- objective optimization and ant colony algorithm based on feature recognition methods of remote sensing images are put forward. First, the remote sensing image feature selection is converted into multiobjective optimization prob- lems, and then use dimension feature subset and recognition rate weighted objective function is constructed, finally using the ant colony algorithm of global optimization and the positive and negative feedback speculative feature sub- set search, to find the optimal feature subset in the remote sensing image, achieve accurate recognition. Simulation experimental results show that the algorithm can quickly find the optimal feature subset, and eliminate the useless and redundant features, reduce the feature dimension, high recognition rate.