Spectral-attention transformer for single-pixel multispectral reconstruction

Abstract

Compared with array detectors (such as CCD or CMOS), single pixel detectors have potential in invisible band and weak light applications to broaden the spectrum of spectral imaging. Existing leading methods for spectral reconstruction (SR) focus on designing deeper or wider convolutional neural networks (CNNs) to learn the end-to-end mapping from the RGB image to its hyperspectral image (HSI). These CNN-based methods achieve impressive restoration performance while showing limitations in capturing the long-range dependencies and self-similarity prior. To cope with this problem, we propose a novel spectral-attention transformer(SAT-net) method for single-pixel multispectral reconstruction. In addition, we introduce total variation (TV) to maintain the smooth structure of HSI. The experimental results of simulation and real data show that the proposed SAT-net is superior to other traditional algorithms based on compressive sensing(CS) methods.

Optimization and Performance Evaluation of an Atomized Acid System for the Expansion of Carbonate Gas Injection

“…Foamed acid is formed by adding gas (nitrogen or carbon dioxide gas) to hydrochloric acid and then adding a foaming agent to form the foamed acid. The presence of foam reduces the contact area between the rock and acid, thus reducing the reaction rate of acid with rock; the high viscosity can control the acid fluid filtration, thereby increasing the effective working distance of acid fluid, and the residual acid is easy to be removed after the construction [10,11]. However, it has the disadvantage of large consumption and high requirements for construction equipment.…”