An Imaging Method for Spectral Imaging Systems Based on Optimization-Inspired Neural Networks

Abstract

The invention discloses an imaging method of a spectral imaging system based on an optimized heuristic neural network, and belongs to the field of computational photography. The implementation methodcomprises the following steps: establishing a forward propagation model of a spectral imaging system, realizing the forward propagation model by using a network, and constructing a coding aperture optimization network; constructing a hyperspectral image reconstruction network which is heuristic based on optimization and considers spatial correlation and spectral correlation of the hyperspectral image at the same time; making a training set; configuring parameters required by hyperspectral image reconstruction network training; training a hyperspectral image reconstruction network; establishinga connection between the coding aperture optimization network and the hyperspectral image reconstruction network, and constructing a joint network; configuring parameters required by joint network training; training the joint network; taking out a coding template obtained after training, and completing modulation from the hyperspectral image to the two-dimensional compressed image based on a CASIsystem imaging process; and reconstructing the target hyperspectral image block by block by using the trained hyperspectral image reconstruction network.

Description

technical field [0001] The invention relates to a hyperspectral image imaging method used in a spectral imaging system, in particular to a method capable of acquiring high-quality hyperspectral images, and belongs to the field of computational photography. Background technique [0002] Different from traditional RGB imaging or panchromatic imaging, spectral imaging captures the scene as a three-dimensional tensor, which samples the spectral information at each pixel position of the scene more finely in the spectral dimension. The hyperspectral image obtained by spectral imaging is rich in spectral information, which makes it more advantageous than traditional imaging technology in the fields of remote sensing, medical imaging, visual inspection, sewage detection, vegetation research, and atmospheric monitoring. are being used more and more widely. [0003] Since hyperspectral images are three-dimensional tensors, and currently used imaging sensors are two-dimensional, spect...

AI Technical Summary

Problems solved by technology

[0009] For the existing imaging methods, the imaging process and the computational reconstruction process are not considered at the same time, and the spatial prior and spectral prior of the hyperspectral image cannot be taken into account in the computational reconstruction process, and the reconstruction efficiency is low.

Images