Research On Image Enhancement Technique Of Visible Light Polarization Imaging

Visible light image is easily affected by external environment factors, the artificial object hided in natural "background is undetectable because of the low contrast of the image. Target detection and recognition has become increasingly difficult with the development of military camouflage and new stealth materials. Polarization imaging techniques can detect target surface polarization information, target and the background can be distinguished by using polarization difference between artificial and natural objects. In order to improve the detection and recognition rate of artificial target, this paper mainly studies the image enhancement technology of optical polarization imaging.This paper focuses on two aspects, the image enhancement with polarization based on the Stokes parameters and on Mueller matrix respectively, and a set of Stokes automatic measurement system is set up for the polarization image acquisition. As the polarization image can highlight the artificial object in a good way, providing more detail and texture information, a polarization image enhancement method based on Stokes parameter is proposed in this paper. First, three polarization images are clustered in the Poincare sphere, which are calculated by Stokes parameters, in this step, we get the polarization characteristic image; second, the wavelet transform is used to fuse the intensity image and the polarization characteristic image, in this step, we obtain the final enhancement image, the results show that the method in this paper can effectively improve the detection and recognition rate of the artificial object, realizing the rapid goal detection. Through the analysis of the typical targets’Mueller matrixs, we find that a certain difference exists between the polarization characteristic of different materials’ Mueller matrixs, which can be useto realize target’s detection and recognition, thus a polarization image enhancement method based on Mueller matrix is proposed. Firstly, the wavelet transform is used to fuse the parametric images of Mueller matrix, then the RGB pseudocolor enhancement method based on threshold segmentation is employed on the fusion image, finally we obtain the final enhancement image. The result shows that, the enhancement method can effectively distinguish target and the background, improving the detection and recognition rate of target.