【作者】 陈旭； 杨义校； 陶然；

【Author】 CHEN Xu;YANG Yixiao;TAO Ran;School of Information and Electronics,Beijing Institute of Technology;Beijing Key Laboratory of Fractional Signals and Systems;

【机构】 北京理工大学信息与电子学院； 分数域信号与系统北京市重点实验室；

【摘要】 主要研究了基于物理驱动的快照压缩光谱成像技术,并结合遥感图像语义分割任务,将快照压缩光谱成像过程与语义分割任务统一起来,实现了对压缩高光谱图像的语义分割技术。提出采用基于广义交替投影的总体变差重建方法在公开天宫数据集WHU-OHS上对高光谱数据进行模拟压缩与重建,并对压缩重建后的高光谱图像使用四种常用神经网络进行语义分割,从而实现压缩高光谱成像的语义分割技术。更多还原

【Abstract】 This paper mainly studies the snapshot compressed spectral imaging technology based on physics drive,and combines it with the remote sensing image semantic segmentation task to unify the snapshot compressed spectral imaging process and the semantic segmentation task,thus realizing the semantic segmentation technology of compressed hyperspectral images.This paper proposes to use the total variation reconstruction method based on generalized alternating projection to simulate the compression and reconstruction of hyperspectral data on the public Tiangong dataset WHU-OHS,and uses four common neural networks to perform semantic segmentation on the compressed and reconstructed hyperspectral images,thereby realizing the semantic segmentation technology of compressed hyperspectral imaging.更多还原