Research On Multi-channel Polarization Imaging Detection Technology Of Non-coaxial Type

The ordinary imaging detection can only obtain the intensity information of the target,and its detection ability is limited in the complex environment.Polarization imaging can enhance the contrast and improve the detection and recognition ability of the target in the complex background by using the polarization characteristic difference between the target and the background.Nowadays,it has been widely used in military,medical,astronomy and other fields.The ordinary polarization imaging system can be divided into time-division polarization imaging system and simultaneous polarization imaging system.The time-division polarization imaging system has poor real-time performance,large imaging error,and it can't detect dynamic targets.The simultaneous polarization imaging system has low energy utilization,low resolution and its detection ability is limited in low light environment.According to these problems,we design a multi-channel polarization imaging system of non-coaxial type.The system uses a four-channel array structure,each channel access to 0 °,45 °,90 ° and 135 ° polarization intensity image.And then,using the SIFT algorithm and Stokes vector to complete registration between channel and calculating the degree of linear polarization,angle of polarization and characteristic factor of polarization of the target.In this way,polarization imaging is completed.We set up a mathematical model for this kind of system structure.We simulate and analyze the errors caused by the distance between channels,the parallelism of optical axis and the polarizer polarization angle on the polarization imaging results.And we summarize the range of system structure parameters with the polarization imaging accuracy greater than 1%.According to the disadvantage that the polarization imaging of the system will lose the background intensity,a polarization image fusion algorithm based on Nonsubsampled Contourlet Transform(NSCT)and Adaptive Pulse Coupled Neural Network(PCNN)is proposed to fuse the light intensity and polarization characteristic factor represented by parameter I in Stokes vector.We use the polarization imaging system designed in this paper to experiment in low light,hidden and remote conditions.The experimental results show that the imaging system and fusion algorithm can significantly improve the detection ability of targets in multiple environments.