基于改进人工蜂群算法的高光谱图像端元提取方法

An endmember extraction method for hyperspectral remote sensing imagery based on improved artificial bee colony algorithm

针对高光谱图像中端元提取的问题,提出了一种基于改进人工蜂群算法的提取方法。首先,为平衡人工蜂群算法全局和局部搜索能力,研究了加权构造蜂引导的搜索策略,构造了改进人工蜂群算法。在8个基准测试函数中进行实验,验证了新算法的性能有明显提升。然后,介绍了基于IABC端元提取的核心思想与主要步骤,与ABC和常规提取算法在模拟和真实高光谱遥感数据中进行实验对比,结果表明了新算法具有更好的适用性。

To solve the problem of endmember extraction for hyperspectral remote sensing imagery, a new endmember extraction method based on improved artificial bee colony algorithm is proposed. First, the weighted generated bee guided search strategy is used to balance the exploration and exploitation in ABC, and a new algorithm named IABC is proposed. Experiments are carried out on 8 benchmark functions, and the results show that the performance of the new and the main steps of the IABC-based extraction has better applicability compared with ABC and real hyperspectral data. algorithm is significantly improved. Then, the core idea are introduced. The results show that the new algorithm conventional extraction algorithm in the simulation and