Research on spatial coding compressive spectral imaging and its applicability for rural survey

Abstract

Compressive spectral imaging combines traditional spectral imaging method with new concept of compressive sensing thus has the advantages such as reducing acquisition data amount, realizing snapshot imaging for large field of view and increasing image signal-to-noise and its preliminary application effectiveness has been explored by early usage on the occasions such as high-speed imaging and fluorescent imaging. In this paper, the application potentiality for spatial coding compressive spectral imaging technique on rural survey is revealed. The physical model for spatial coding compressive spectral imaging is built on which its data flow procession is analyzed and its data reconstruction issue is concluded. The existing sparse reconstruction methods are reviewed thus specific module based on the two-step iterative shrinkage/thresholding algorithm is built so as to execute the imaging data reconstruction. The simulating imaging experiment based on AVIRIS visible band data of a specific selected rural scene is carried out. The spatial identification and spectral featuring extraction capacity for different ground species are evaluated by visual judgment of both single band image and spectral curve. The data fidelity evaluation parameters (RMSE and PSNR) are put forward so as to verify the data fidelity maintaining ability of this compressive imaging method quantitatively. The application potentiality of spatial coding compressive spectral imaging on rural survey, crop monitoring, vegetation inspection and further agricultural development demand is verified in this paper.