Time Of Flight Imaging Via Compressed Sensing

With the rapid development of laser ranging technology,because of its non-contact,non-destructive,high speed,long distance and other characteristics,it has been gradually applied to industry and daily life,and widely used in remote sensing,machine vision,three-dimensional imaging and other fields.At present,three-dimensional imaging technology mainly includes: structured light three-dimensional imaging,binocular vision three-dimensional imaging and time of flight three-dimensional imaging technology.Time of Flight 3D imaging uses the round-trip time of a photon traveling in space to measure the depth of field of an object.In recent years,time-of-flight technology has significant advantages,and has been widely used in disaster rescue,remote sensing,communication,gesture recognition,game interaction,face recognition,three-dimensional scene measurement and other fields.At present,most of the time-of-flight 3D imaging technologies have no obvious reconstruction effect on high-quality scenes.A large number of scanning lighting points in the scene will limit the acquisition speed,and focal plane arrays with high spatial and temporal resolutions are not widely used.Compared with traditional color cameras,time-of-flight sensors have low resolution and low signal-to-noise ratio.The research content of this paper is to introduce the compressed sensing theory into the time-of-flight three-dimensional scene measurement,which can reduce the image noise and effectively improve the image signal-to-noise ratio.The principle of time-of-flight imaging is to transmit and receive near-infrared light and obtain the scene depth by calculating the phase difference of the optical signal,which can reduce the interference of background light to the imaging to a certain extent.However,due to the characteristics of complementary metal oxide semiconductor,only 10% of the filling factor,so the reflected signal accounted for a small part of the detector,resulting in the image signal-to-noise ratio,resolution reduction.In this paper,the iterative soft threshold algorithm is used to reconstruct low signal-to-noise ratio images from time-of-flight sensors to obtain high signal-to-noise ratio images,and high quality images can be obtained when the sampling rate is reduced.In this paper,numerical simulation and time of flight based compressed sensing 3D imaging simulation experiments are carried out.The details of the reconstructed image are clearer than the original image,and the signal-to-noise ratio of the reconstructed image can be improved effectively.In the case of under-sampling,the image with high signal-to-noise ratio can be obtained.Compared with 3D objects,the noise of reconstructed images is reduced during full sampling,and the difference of details between the two images is not obvious.The restoration of details of 3D object reconstruction is not as obvious as that of binary images.Through the evaluation of image quality indicators,such as structural similarity,signal-to-noise ratio,peak signal-to-noise ratio and root mean square error,etc.,through simulation,the reconstructed image is improved compared with the original image quality,the detail part is recovered obviously,the image noise is reduced effectively,the signal-to-noise ratio is increased and so on.