[15p-2F-1~15] 4.5 Information Photonics

▲ [15p-2F-8] [JSAP-OSA Joint Symposia 2015 Invited Talk] Ultrasensitive denoising imaging

〇Guang-Jie Zhai¹, Wen-Kai Yu^1,2, Qing Zhao² (1.CSSAR, CAS, 2.BIT)

キーワード：computational imaging and display,multispectral imaging

Traditional methods in remote spectral imaging usually cannot obtain images and spectra simultaneously, either rely on scene raster scanning or use an array sensor. Thus it is inevitable that the flux will be distributed evenly to the pixel dimension, resulting in a decline in the imaging sensitivity. To solve the problems, we proposed a compressive spectral imaging scheme with dual spatial light modulation. With the information of spectra and space both compressed, the spectral image of a colored object can be obtained by a single-photon point detector. Objects with both continuous and discrete spectrum are used to demonstrate the performance, which shows wide-spectrum and high efficiency in spectral imaging. Since this approach is based on the technique of single-pixel imaging, we have found a correlation between the samplings in both reflection arms of the digital micro-mirror, in view of which we propose a complementary way of double-pixel compressive imaging on denoising. In a telescopic system, this method can be used to retrieve images of a target with 20 cm resolution, obtaining an image in a higher quality of 1~2 orders than that acquired by the conventional single-pixel imaging scheme under the same sampling rate. Our results have shown that the above method has strong ability in producing high-quality images with fewer measurements, which leads to a broader prospect of applications in the remote-sensing area.