Research On Polarization Imaging Technology Of Laser Active Illumination And Image Fusion Algorithm

In this dissertation,laser active illumination and polarization imaging technology are combined to achieve high quality image acquisition under complex weather conditions or harsh environment,some advanced image fusion algorithms are developed to improve image quality.The research work of this technology has been carried out step by step from several aspects,such as laser band selection method,laser beam transmission transformation and quality evaluation,laser link transmission model construction,experimental analysis and verification of imaging mechanism and advantages,image fusion and so on,in order to lay a theoretical and technical foundation for the follow-up work and accelerate the engineering application of laser active illumination polarization imaging technology.Firstly,the laser band for illumination,the beam propagation quality evaluation system and the mathematical model of laser power and illumination distance are determined.The laser band for illumination is determined according to many coupling factors such as atmospheric window,target characteristics,absorption effect,etc.And the beam propagation quality evaluation system is determined on the basis of analyzing the mismatch loss between laser beam expansion and optical axis,effect of atmospheric turbulence on light intensity distribution,beam drift and expansion.The mathematical model of laser power and illumination distance is established on the basis of analyzing the requirement of laser power such as the sensitivity of detector end,image contrast and signal-to-noise ratio.Secondly,a platform is built to study static active polarization imaging technology in smoke and dust environment.The 532 nm CW laser which was collimated and expanded is used as illumination source,metal bullet model is used as detection target,smoke environment is created by smoke generator in organic glass hood,polarization direction is rotated by adjustable polarizer,image is collected by LM135 camera,and image processing is carried out.While the results of laser illumination and polarization imaging are quantified by the image information entropy,average gradient,spatial frequency and standard deviation,and the imaging advantages of this technology in smoke and dust environment and dark conditions are determined.Thirdly,on the basis of the original experimental results and the improvement and application of the original three image fusion algorithms,such as image fusion algorithm based on feature extraction and visual information preservation,image fusion algorithm based on potential low rank representation,image fusion algorithm based on discrete cosine transform and local spatial frequency,then the performance of the algorithm is compared by the fusion of experimental images.Finally,proposes a new fuzzy adaptive polarization image fusion algorithm to further improve the image quality.The experimental results of the new algorithm show that the algorithm can further enhance the detection distance,image contrast and signal-to-noise ratio,and expand the digestible information of the image,especially for the dim target under low visibility strip.Compared with the visible image,the entropy of fusion image was increased by 1.75 times,and the average gradient was increased by 6.2 times.The new algorithm provides important data reference and technology foundation for the application of laser active illumination and polarization imaging technology.The polarization imaging technology under active laser illumination studied in this dissertation can suppress background noise,improve detection distance,acquire target details,identify target camouflage and distinguish attractants.It has guiding significance and wide application prospects in precision guidance,battlefield reconnaissance,security control,intelligent transportation,medical and other fields.