Investigation On Long Distance Active Three-dimensional Imaging Based On Compressed Sensing

Long distance active three-dimensional(3D) imaging technic is an information acquiring method which can obtain3D information of long distance target. Compared with other3D information acquiring methods, long distance active3D imaging technic has many merits of non-contact, high speed and high resolution, etc. Long distance active3D imaging technic has been widely employed in remote sensing imaging, terrain reconnaissance, medical imaging and radar imaging, etc. Therefore, long distance active3D imaging technic has great applications and research value in the civilian and military.Compressed sensing(CS) is an emerging theory of signal processing theory, it points out that under certain conditions, with much fewer numbers of sampling than that required by the Nyquist sampling theorem, the original signal can be accurately reconstructed. In this thesis, compressed sensing is introduced to long distance active3D imaging technic. With the merit of simultaneous signal sampling and compression, the long distance active3D imaging technic based on CS can reduce the number of the huge data obtained by traditional3D imaging technic.In this thesis, the research status of active3D imaging are introduced, and the basic principle of compressed sensing theory is described firstly. Then, the long distance active two-dimensional(2D) and3D imaging system is designed and constructed. The system is composed by using a high-power pulsed laser, a Cassegrain telescope, a digital micro-mirror array(DMD), a photomultiplier tube(PMT) and a capture card(A/D). This system is applied to the imaging of the long-range2D and3D object. Through the experimental system built, the measurement to the absolute distance of the object at a distance of230m and4.5km, respectively, is implemented and the64pixels X64pixels3D images have been realized. Sparsity of the target and the sampling rate on the reconstructed image quality also has been analyzed. It is demonstrated that for the image reconstruct of long distance using CS, with the increase of sampling rate, the quality and contrast of the reconstructed image is improved to some extent. The sparser the image of the object, the less the number of sampling required for image reconstruction. Since the long distance imaging system works in the atmospheric environment, it will be inevitably affected by the atmospheric turbulence. In this part, an image affected by the atmospheric turbulence is obtained in simulating. The affected image by the atmospheric turbulence is reconstructed by using compressed sensing algorithm. The quality of the reconstructed image is also analyzed. Then a new3D image reconstruction method based on distance-dimensional coding modulation is proposed. By including the distance-dimensional coding information in the compressed sensing algorithm, this new method can further reduce the compression ratio. Using this method, a simulation is completed, verifying that the new method works well and it is feasible.Finally, the present work is summarized and the future work is prospected.