In response to the problem that traditional spectral signal processing models are affected by noise and difficult to fully capture spectral information, resulting in insufficient spectral reconstruction accuracy, an innovative spectral signal processing model based on super‐resolution theory fused with a hypernetwork architecture has been proposed. The core is to establish a wideband spectral reconstruction method through super‐resolution theory and introduce a super network structure to further optimize the spectral reconstruction process. Finally, performance verification and analysis are conducted on the dataset. Experimental verification showed that the research model had the smallest reconstruction error in reconstructing spectra of acid, alcohol, and benzene substances, with a mean square error value reduced by 85.64% compared to the other three models and a spectral reconstruction time reduced by 80.55%. In the simulation test, the model performed optimally in both 20‐dB noise level and Δλ ~ N (0, 100 2 ) environments, with reconstruction errors reduced by 44.22% and 77.21% compared to other models. The classification accuracy has improved by 6.92% and 10.12%, and the spectral reconstruction time has been reduced by 72.91% and 80.26%, respectively. It maintained optimal performance at two noise levels of 30 and 40 dB, as well as other offset levels. The results indicate that the proposed spectral signal processing model has good noise resistance and can effectively improve the accuracy of spectral reconstruction, providing a new approach for the development of spectral analysis technology in complex environments.
Zhang et al. (Thu,) studied this question.