基于最小互信息的自适应偏振差分成像方法

Adaptive polarization difference imaging approach based on minimum mutual information

偏振信息解析是偏振成像探测涉及的一个重要课题。针对传统的偏振差分成像方法存在的不足,从信息论的角度出发,提出了一种基于最小互信息的自适应偏振差分成像方法。首先,利用偏振信息解析得到的斯托克斯参量图像,计算得到偏振片透光轴与所选参考坐标轴的成任意夹角的出射光强度图像;然后,以归一化互信息为评价指标,寻找具有最大不相关的两幅出射光强度图像;最后,根据找到的两幅图像的差分结果进行成像。实验结果表明,提出的方法不仅具有更好的抑制背景、突出目标细节特征的能力,而且得到的图像质量更高,从而验证了该方法的可行性和有效性。

Polarization information analysis is an important topic for polarization imaging detection.To overcome the drawback of the conventional polarization difference imaging method,from the viewpoint of information theory,a kind of adaptive polarization difference imaging approach based on the principle of minimum mutual information was introduced.Firstly,according to the Stokes parameters images obtained by polarization information analysis,the outgoing light intensity image oriented by any angle between the transmission axis of the filter and the selected reference coordinate axis was calculated.And then,taking the normalized mutual information as the evaluation index,two outgoing light intensity images with maximum irrelevance were found out.Finally,these images were used to imaging by difference operation.Experimental results show that the proposed method not only has better ability to restrain the background and pop out the details of targets,but also can obtain better image quality which testifies its feasibility and validity.