Abstract
Limited by the existing imagery sensors, hyperspectral images are characterized by high spectral resolution but low spatial resolution. The super-resolution (SR) technique aiming at enhancing the spatial resolution of the input image is a hot topic in computer vision. In this paper, we present a hyperspectral image (HSI) SR method based on a deep information distillation network (IDN) and an intra-fusion operation. Specifically, bands are firstly selected by a certain distance and super-resolved by an IDN. The IDN employs distillation blocks to gradually extract abundant and efficient features for reconstructing the selected bands. Second, the unselected bands are obtained via spectral correlation, yielding a coarse high-resolution (HR) HSI. Finally, the spectral-interpolated coarse HR HSI is intra-fused with the input HSI to achieve a finer HR HSI, making further use of the spatial-spectral information these unselected bands convey. Different from most existing fusion-based HSI SR methods, the proposed intra-fusion operation does not require any auxiliary co-registered image as the input, which makes this method more practical. Moreover, contrary to most single-based HSI SR methods whose performance decreases significantly as the image quality gets worse, the proposal deeply utilizes the spatial-spectral information and the mapping knowledge provided by the IDN, which achieves more robust performance. Experimental data and comparative analysis have demonstrated the effectiveness of this method.
Highlights
Hyperspectral imagery sensors usually collect reflectance information of objects in hundreds of contiguous bands over a certain electromagnetic spectrum [1], and the hyperspectral image (HSI) can simultaneously obtain a set of two-dimensional images [2]
HSI SR has been studied for a long time in remote sensing and many methods have been proposed to improve the spatial resolution of the HSIs
We propose an HSI SR method by combining an information distillation network (IDN) [19] with an intra-fusion operation to make a deep exploitation of the spatial-spectral information
Summary
Hyperspectral imagery sensors usually collect reflectance information of objects in hundreds of contiguous bands over a certain electromagnetic spectrum [1], and the hyperspectral image (HSI) can simultaneously obtain a set of two-dimensional images (or bands) [2]. An IDN is used for super-resolving the interval-selected bands individually, a process exploiting their spatial information and the mapping learned by the IDN. The unselected bands are fast interpolated via cubic Hermit spline method, which uses the high spectral correlation in the HSI to obtain a coarse HR HSI Both spatial and spectral information is utilized. Most existing single-based methods super-resolve bands in the HSI individually, which neglects the spectral information In this way, their performance is highly correlated to the images’ spatial quality. Different from most fusion methods, which require another co-registered image as the input, the other input of the proposed intra-fusion is an intermediate outcome of the SR processing, which fully exploits the information the LR HSI conveys in a subtle way.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
