Research On Dual Camera Polarization Imaging Technology Based On Stokes Vector

The polarization characteristics of light can reflect the essential characteristics of matter.When matter in nature is irradiated by light,the incident light interacts with the matter,and the polarization state of the reflected light changes immediately.Different types and states of matter will show specific characteristics.Polarization characteristics provide a new way of thinking for target detection and identification.Aiming at the problems of traditional polarization imaging technology such as large mechanical vibration and incomplete information acquisition,this thesis proposes a dual-camera polarization imaging technology based on Stokes vector and liquid crystal polarization modulation as the core,and develops a new type of time-sharing polarization imaging detection system.This thesis focuses on the polarization imaging detection technology,and first carry out in-depth research on the mechanism of imaging detection.The focus includes the Stokes vector method and the measurement of the Stokes vector,the polarization modulation technology based on the liquid crystal adjustable phase retarder(LCVR),and the typical time-sharing polarization imaging system.Secondly,starting from the polarization mechanism,build a dual-camera polarization imaging system,which is mainly divided into imaging and detection unit,liquid crystal drive control unit,image extraction and processing unit,and then design each unit,involving component selection,schematic drawing,and software development etc.Finally,the polarization image processing algorithm is researched,including polarization image preprocessing,registration,and fusion.In terms of polarization image preprocessing,this thesis introduces the concept of variance in the median filter algorithm to improve the noise filtering ability.In terms of registration,by analyzing the factors that cause the polarization image mismatch,a feature point-based image registration algorithm is adopted for the transmission image and the reflection image.In terms of image fusion,the first component image and polarization characteristic image of the Stokes vector are processed by an improved two-dimensional discrete wavelet transform,and compared with the weighted average method and the pixel absolute value selection method.The dual-camera polarization imaging system designed in this thesis uses liquid crystal flexible full-polarization modulation technology.Under the premise of ensuring the detection accuracy of the system,it can also solve the problem of incomplete information acquisition,realize complete Stokes vector measurement,and obtain linear and circular polarization images at the same time.Through indoor and outdoor polarization imaging detection experiments,it is shown that the polarization degree image can highlight more detailed features of the target object than the linear polarization degree image.In addition,through carrying out indoor and outdoor polarization image fusion and comparison experiments,the results show that the algorithm in this thesis can effectively improve the clarity and contrast of the image.