光滑物体表面反射光偏振特征分析及反射光分离技术

2021,43(6):135-144
张焱
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073,atrthreefire@sina.com
张景华
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
石志广
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
张宇
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
凌峰
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
刘荻
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
索玉昌
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
师晓冉
国防科技大学 电子科学学院 ATR 国防科技重点实验室, 湖南 长沙 410073
摘要:
物体表面反射光由镜面反射光和漫反射光组成,针对光滑物体表面镜面反射光和漫反射光的分离问题,建立了一个结合镜面反射和漫反射共同作用的偏振态表征模型,并使用该模型对混合反射区域偏振特征进行分析。基于镜面反射光和漫反射光的偏振特征,根据镜面反射光和漫反射光在光滑物体表面垂直方向及平行方向上的分布关系,采用偏振正交分解的方法实现了镜面反射成分和漫反射成分的求解。利用镜面反射光图像和漫反射光图像之间的相关性,通过梯度下降算法得到镜面反射成分和漫反射成分归一化互相关最小值及其对应的镜面反射偏振度和漫反射偏振度,从而实现反射光的分离。通过对不同场景中获取的反射光图像进行测试,实验结果表明,该算法具有较好的反射光分离效果。
基金项目:
国家自然科学基金资助项目(62075239,61302145);国家部委基金资助项目(9140C800302KG01)

Analysis of polarization characteristics of reflected lights on smooth object surface and separation of reflected lights

ZHANG Yan
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China,atrthreefire@sina.com
ZHANG Jinghua
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
SHI Zhiguang
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
ZHANG Yu
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
LING Feng
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
LIU Di
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
SUO Yuchang
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
SHI Xiaoran
National Key Laboratory of Science and Technology on Automatic Target Recognition, College of Electronic Science and Technology, National University of Defense Technology, Changsha 410073, China
Abstract:
Reflected lights on object surfaces are composed of specular and diffuse reflection lights. To separate the specular and diffuse reflection lights on a smooth object surface, a polarization state characterization model based on the collaborative effect of specular and diffuse reflections was proposed and was used to analyze the polarization characteristics in a mixed reflection region. Based on their polarization characteristics and their distribution relations perpendicular and parallel to the smooth object surface, the specular and diffuse reflection lights were separated by polarization orthogonal decomposition. According to the correlation features between the specular and diffuse reflection components, the normalized correlation minimum between the specular and diffuse reflection lights and its corresponding to the degrees of specular and diffuse reflection polarization were calculated by the gradient descent algorithm, thus enabling the successful separation of the reflected lights. By testing reflected light images obtained from different scenes, the results indicate that the proposed algorithm could effectively separate reflection lights.
收稿日期:
2021-09-10
     下载PDF全文