»ã±¨±êÌâ (Title)£ºL_2-L_q reconstruction model for computational spectrometer£¨ÍÆËã¹âÆ×ÒǵÄL2-Lq³Á¹¹Ä£ÐÍ£©

»ã±¨ÈË (Speaker)£º ÍõÔöçù½ÌÊÚ£¨ÉϺ£½»Í¨´óѧ£©

»ã±¨¹¦·ò (Time)£º2023Äê10ÔÂ24ÈÕ(Öܶþ) 14£º30

»ã±¨µØÖ· (Place)£ºÐ£±¾²¿F309

Ô¼ÇëÈË(Inviter)£ºÁõÇÉ»ª

Ö÷°ì²¿ÃÅ£ºÀíѧԺÊýѧϵ

»ã±¨ÌáÒª£ºThe extensive application of computational strategies in spectral detection enables optics-free spectrometers, promoting the miniaturization of spectrometers. It reconstructs the spectral composition of the incident light based on the measurement results from a set of detectors and a known spectral response matrix. The spectral reconstruction model and its simulation are key to identifying the incident light components with high precision. In this study, we reconstructed the incident light function by discretizing the integral equation and applying L_2-L_q optimization approach to obtain the spectral reconstruction model. Simulation results show that L_2-L_q optimization has higher resolution and higher noise tolerance than certain other optimization approaches, which is advantageous for identifying unknown optical signals with a broad spectrum and high dynamic range. On this basis, we discuss the application of preconditioned L_(1/2) Regularization method with adaptive Regularization parameter in the reconstruction model of computational spectrometer.