Research Of The Target Detection Technology Based On Visible Light Polarization Imaging

Target detection is an important subject in the field of photoelectric tracking and scene recognition. With the increasing complexity of the natural scene and the continuous development of the camouflage technology, it is difficult for conventional detection methods to detect the camouflage targets effectively. As a new method of detection, imaging polarization detection distinguish between radiation and polarization characteristics of target and background, and obtain the intensity and polarization information at the same time, which can effectively enhance the significance of image details and improve the detection efficiency in the complex background.This paper mainly study the application of the polarization imaging technology in the field of target detection and scene recognition, mainly including the following contents:In the first part, the theoretical basis of polarization detection is introduced. Polarization imaging detection principle, transmission process and the general process of polarization images are introduced with the acquisition of the polarization image, the pretreatment of the image and the extraction of the polarization features.In the second part, polarization information fusion and polarization target enhancement in polarization imaging detection technology are studied. We propose a new method of visible light polarization image fusion based on nonsubsampled shearlet transform(NSST) on the basis of the Pyramid transform, wavelet transform, nonsubsampled contourlet transform(NSCT) and nonsubsampled shearlet transform(NSST) theory. We introduce guide filter theory in the process of target polarization enhancement, and put forward a dual scale guide filter enhancement algorithm. Finally, the subjective and objective evaluation methods of the polarization images fusion enhancement effect are described.In the third part, the target detection experiment of the visible and infrared light polarization imaging is carried out according to the different targets and scenarios on the built platform. The proposed polarization image fusion and enhancement method is used to fuse and enhance light intensity image and polarization characteristic image. The obtained experimental results are compared with the mean value method, Pyramid method, wavelet method and NSCT method in terms of subjective and objective effect and the experimental conclusion are drawn.