Snapshot Polarization Imaging Based On Aperture Filter Of Light Field Camera

Optical imaging is an important method of information acquiring for people to understand nature and explore the unknown.With the continuous development of optoelectronic technology,optical imaging has developed from the ability of capturing light intensity to recording multi-dimensional optical information.Typical dimensions of optical information include amplitude,phase,polarization,and spectrum.Among them,polarization is closely related to the material,geometric shape,texture,and surface roughness of the object.Therefore,polarization imaging can be used as an important supplement to classical intensity imaging,and becomes a new technology for specific target detection in special environments.The traditional polarization imaging captures the polarized information of the scene in the form of time sequential imaging through the rotation of a single polarizer,which is simple and easy but has low time efficiency.In order to realize dynamic polarization imaging,a series of simultaneous polarization imaging techniques based on spatial multiplexing have been developed,including amplitude division.aperture division and focal plane division.The emergence of light field camera provides a potential new solution for dynamic polarization imaging.The light field camera can capture information from multiple viewing angles of the scene at the same time.By performing aperture filtering on the main lens,images of different viewing angles can be filtered separately to realize multi-functional imaging based on spatial multiplexing.This thesis proposes a snapshot polarization imaging system based on light field camera and aperture filtering,which can realize dynamic polarization imaging and has the advantages of small size and low cost.The thesis mainly completed the following works:1.The imaging principle of the focused light field camera is studied.Combined with the optical matrix transmission theory,the imaging process of the focused light field camera is simulated through MATLAB.The system structure and key parameters are analyzed and simulated,which provides a reference for the subsequent construction of the light field imaging system.The principle of aperture filtering is analyzed in the framework of light field imaging.The effect of the filter array is simulated,and the function of aperture filtering imaging with light field camera is verified through simulation experiments.2.The reconstruction algorithm of light field image under different filtering states is studied and realized.By decomposing and reconstructing the original image of the light field camera,the sub-images of the scene in different filtering states are obtained.Registration of the sub-images can compensate the spatial translation caused by the different viewing angles of the sub-images,which benefits the subsequent processing of the scene information.3.A focused light field camera with aperture filtering is established.The centers of micro lens are calibrated to improve the accuracy of sub-image reconstruction.An array of attenuation plates is placed at the aperture stop of the main lens,and images of the scene under different exposures are obtained.After sub-image registration and synthesis,the snapshot high dynamic range imaging is realized,which verifies the aperture filtering function of the light field camera.4.In the focused light field camera,a polarizer array is placed at the aperture stop of the main lens to realize snapshot polarization imaging.The accuracy of polarization imaging is tested with different polarization states.The results show that the root mean square error(RMSE)between the measured and the theoretical value of the light intensity changing with the polarization angle is 2.63%,which has a high polarization accuracy.Using the dynamic scene in low-light environment as the testing target,the Stokes vector,degree of polarization(DOP)and angle of polarization(AOP)of the scene are obtained through proposed snapshot polarization imaging.Compared with traditional intensity image,it is easier to identify and locate moving targets in low-light environments by using polarization image.The trajectory of the target detected by the proposed system has a high degree of coincidence with the planned trajectory,which verifies the feasibility of dynamic polarization imaging.