| With the rapid development of the information age,requirements of image quality become increasingly urged to present target information,and more and more imaging techniques have been proposed correspondingly.Optical imaging detection techniques normally achieve imaging by obtaining the intensity,spectrum and spatial state of the target,then the properties of the target are acquired.However,these techniques are highly environmental dependent.If the target is exposed in the low visibility environment,or the target is camouflaged,covered or hidden in a complex background,the reflected light will become weaker with the increasing of transmission distance due to the strong scattering of the scattering medium(such as smoke,dust,etc.);while,the scattering light becomes stronger and stronger,which leads to the decrease of signal to noise ratio.So,the target information is submerged in the background noise.Polarimetric optical imaging technique is a novel optical imaging detection technique based on the polarization property of the reflected light from target.By acquiring several images of an object with different polarization orientations,the polarization information of the object can be obtained by pixel-to-pixel calculations based on Stokes vector.Polarimetric imaging reflects the polarization information of the target,which can be used to suppress the background noise,to improve the detection distance(visibility),and to get the detailed characteristics of the target,thus to realize the purpose of identifying the target.Active polarimetric imaging has become an important research direction of polarimetric imaging since it is suitable for all-weather detection,especially for the detection distance improvement of the dark,weak and small targets.In this dissertation,both theoretical and experimental researches on polarimetric imaging of Mueller matrix,imaging enhancement of target by polarimetric descattering,and three-dimensional polarimetric imaging have been studied.The main research contents and innovative achievements can be listed as follows:(1)A fast Mueller matrix measurement method based on a division of aperture simultaneous polarimetric imaging technique is proposed.This method can overcome the drawbacks of multiple operating steps and the large time cost for traditional Mueller matrix measurement method by mechanically rotating polarizer or wave-plate.The innovations of this technique include: the liquid crystal variable retarders(LCVRs)are used as the polarization state generator;a division-of-aperture polarimetric camera with the full-Stokes polarization state simultaneous detection is not only used as the polarization state analyzer,but also used as the imaging detector.Four polarized images(including three linearly polarized images and one circularly polarized image)can be obtained simultaneously by the polarimetric camera,so 16 polarized images can be obtained by 4 times adjustments of the driving voltages of the LCVRs,so that the 16 elements of the Mueller matrix can be solved.Compared with the traditional method,which requires at least 16 or more times independent measurement,the proposed method is simple in calculation and has high efficiency in measurement.Corresponding experimental studies of Mueller matrix measurement are performed,and the average relative error between the experimental results and the theoretical results is less than 0.033.(2)A kind of novel Stokes polarimetric descattering algorithm based on low-pass filtering in the frequency domain is proposed.The deep analysis on existing polarimetric descattering algorithm based on Stokes vector shows that there are several drawbacks due to the quantum noise.They usually need bias parameters to help to obtain high-quality descattered images.These bias parameters always need to be adjusted in real-time under different scattering environment condition or in different RGB channels,which make the polarimetric descattering method less robust.The innovations of this technique include: fourier transformation is applied to the image;and a 5×5 low-pass filter is used for filtering in the frequency domain.Due to the fact that both the directly transmitted light and the noise signals are with high frequencies,the image remains scattering light signal after low-pass filtering.Then,the polarized characteristics and intensity of the scattering light can be precisely estimated without the bias parameters,so the parameter fixation in the image degradation model is realized.The corresponding underwater polarimetric desscattering experiment is carried out,and the experimental results show that the image contrast could be promoted to over 100%.This method is also suitable for hazy images dehazing process in the atmosphere.This algorithm also has a short running time,the computational efficiency has enhanced over 70% compared to the previously proposed method.(3)A three-dimensional polarimetric imaging technique based on the fringe projection of structured light is proposed.Three-dimensional imaging can obtain the profile of the object,while polarimetric imaging can obtain the structure,roughness,material and other characteristics of the object.In order to get more information of the object,firstly,the four phase-shifted sinusoidal fringes are projected on the target;secondly,a flood-fill method is introduced to unwrap the phase;finally,the degree of polarization and angle of polarization information are obtained to reconstruct the three-dimensional polarimetric images of the target.Experiments based on monochromatic and color polarimetric camera are carried out,respectively.The experimental results demonstrate that three-dimensional polarimetric imaging can obtain not only the three-dimensional information of the target,but also the polarization characteristics of it.The polarization information in the RGB channels are also different.This technique simultaneously combines the advantages of three-dimensional imaging and polarimetric imaging,so that more information of an object can be obtained,thus the features of object are more obvious. |
PDF Full Text Request