基于MAP的高光谱影像增强方法

MAP Based Hyperspectral Image Enhancement Method

介绍了MAP方法的基本原理,建立了适用于高光谱影像增强的观测模型,给出了模型的代价函数。在假设高光谱影像像点相互独立的条件下,给出了代价函数的求解方法,分析并给出了基于MAP的高光谱影像增强流程。利用PHI高光谱影像及其对应的高分辨率全色影像数据,进行了MAP高光谱影像增强实验。实验结果的定量分析表明,本文方法能够在保持影像光谱信息的同时,利用高分辨率全色影像,提高增强影像的信息内容和清晰度。

In this paper,basic principle of MAP method was introduced,observation models that were appropriate for hyperspectral images were constructed,and a cost function of the model was given.Assuming that hyperspectral image pixels were independent,a solution method for the cost function was given,and then,the MAP based hyperspectral image enhancement process was analyzed.A PHI hyperspectral image dataset and its corresponding high resolution panchromatic image were used to conduct the MAP based hyperspectral image enhancement experiments.Experimental results proved that the method proposed in this paper could reserve the spectral information of original hyperspectral images,and meanwhile,this method could use the high resolution panchromatic image to improve the information and clearness of the enhanced images.