On Sampling Efficiency Of Single-Pixel Imaging Technology

Single-pixel imaging technology can obtain the spatial information of object and reconstruct its image by using single-pixel detectors.Because the production of single-pixel detectors is generally easier and more cost-effective than that of pixelated detectors,especially for working at invisible wavelengths,so single-pixel imaging is of great significance in the field of imaging at invisible wavelengths.And as single-pixel imaging only need to get intensity value to reconstruct,so it has lower requirement of imaging environment,when the traditional imaging has to keep there's no block between pixelated detectors and object.Therefore,single-pixel imaging occupies an important position in the field of low light level imaging and imaging through scattering media.However,quality of single-pixel imaging is still inevitably lower than the traditional imaging technology.And the greatest disadvantage of single-pixel imaging detectors is that it cannot obtain the information of location with only one measurement.To make up for this defect,single-pixel imaging cost a huge quantity of measurements,especially when reconstructing the high-resolution image.Inefficiency of single-pixel imaging significantly affects its applicability and prevents the development of relevant researches.The means to improve efficiency of single-pixel imaging by improving sampling efficiency is researched in this paper.The conclusion that Fourier single-pixel imaging(FSI)is the most high-sampling-efficiency is proposed by comparing different common single-pixel imaging techniques.On the basis of this,two methods of improving sampling-efficiency of FSI are proposed for the object with sparse space and frequency-domain respective.By sampling according to the features of object,sampling numbers can be slashed when the quality of result is barely influenced,successfully improving the sampling-efficiency of FSI.