Rapid Hyperspectral Imaging System via Sub-Sampling Coding

Bingchen Wei, Zhuang Zhao, Jing Han, Jun Lu, Haocun Qi

Abstract: In recent years, with the rapid advancement of hyperspectral imaging technology, a growing number of hyperspectral reconstruction approaches have arisen. However, most present technologies have the drawback of being unable to perform both quick spectrum measurement and high image resolution at the same time. In order to address this issue, this article proposes a rapid hyperspectral imaging system via subsampling coding. First, the hyperspectral scene is subsampled using the designed Hadamard coding matrix, then a suitable network structure for the hyperspectral image (HSI) super-resolution is investigated, and lastly high-resolution HSIs are obtained while the number of encodings is reduced. The experimental results reveal that this method is adaptable to a variety of imaging situations and performs well. At the same time, in the field of hyperspectral remote sensing, the problem of relatively low spatial resolution of HSI can be better solved. Furthermore, the method has unlimited image spectral dimensions, good generalization ability and noise reduction capabilities.