Single-pixel Imaging Based On LED White Light Spectrum

A new type of imaging technology called "ghost" imaging,which has the characteristics of resistance to atmospheric turbulence and non-field of view,has been widely concerned by the academic community in recent years.Why call it "ghost" imaging? This is because,compared with the traditional imaging technology,this imaging technology realizes the separation of objects and images-obviously no object information appears in the signal optical path containing the target object information,but after the correlation calculation between the idle optical path and the signal optical path,it actually shows Target object information,so it is also called associative imaging.In 2004,an image sampling technology that combines sampling and compression entered people's field of vision-compressed sensing algorithm.This algorithm breaks the limitation of signal sampling and performs sampling in a way of "sampling while compressing".At present,the most famous achievement of this theory is the single-pixel camera developed by Rice University.This provides a new idea for simplifying the experimental optical path for correlative imaging,and also greatly reduces the sampling rate of the image.So far,single-pixel imaging technology has emerged in correlative imaging.The imaging technology is divided into two types,active and passive.In this paper,a passive single-pixel imaging system based on LED is constructed,and the system is combined with spectroscopy to complete single-pixel imaging based on LED white light spectrum.When constructing a passive single-pixel imaging system,the system is roughly divided into an optical path modulation module,a data acquisition module and a signal reconstruction module.In the optical path modulation module,the digital micromirror device(DMD)is used as the main experimental instrument,and the Hadamard matrix after Haar wavelet transformation is set as the measurement matrix.Under the same experimental conditions,it is proved that the quality of the image reconstructed by the measurement matrix is better than that of the traditional Hadamard matrix and the random binary matrix,and the sampling rate is lower.In the data acquisition module,we used a free space bias detector,a benchtop photodiode amplifier,and a USB data acquisition card.Experiments show that the acquisition frequency of the acquisition module is three orders of magnitude higher than that of the acquisition module based on CCD camera.In the signal reconstruction module,we use a reconstruction algorithm based on total variational regularization.By comparing the change trend of structural similarity,signal-to-noise ratio and mean square error of reconstructed images using this algorithm and the correlation algorithm,it is concluded that the quality of the restored image of the algorithm is better than that of the correlation algorithm in the above imaging system.In summary,we obtain a single-pixel imaging system capable of high-speed data acquisition and low sampling rate.Finally,the single-pixel imaging system is combined with spectroscopy,so that ordinary two-dimensional correlation imaging can obtain three-dimensional data,that is,two-dimensional spatial coordinates and one-dimensional spectral coordinates.This will realize the imaging that can resist atmospheric turbulence during earth observation under the comparison of specific spectral characteristics,and has broad application prospects in the ocean,military,agriculture,etc.