基于BRDF可见光偏振成像技术的海面溢油油种识别研究  

Oil spill–type identification based on BRDF visible polarization imaging technology on sea surface

作　　者：刘双 马毅[1,4] 李英超 姜宗辰 杜凯 王斌 LIU Shuang;MA Yi;LI Yingchao;JIANG Zongchen;DU Kai;WANG Bin(First Institute of Oceanology,Ministry of Natural Resources,Qingdao 266061,China;College of Oceanography and Space Informatics,China University of Petroleum(East China),Qingdao 266580,China;School of Opto-Electronic Engineering,Changchun University of Science and Technology,Changchun 130022,China;Technology Innovation Center for Ocean Telemetry,Ministry of Natural Resources,Qingdao 266061,China;Jilin Provincial Key Laboratory of Space Optoelectronics Technology,Changchun 130022,China;Harbin Institute of Technology,Harbin 150001,China)

机构地区：[1]自然资源部第一海洋研究所,山东青岛266061 [2]中国石油大学(华东)海洋与空间信息学院,山东青岛266580 [3]长春理工大学光电工程学院,吉林长春130022 [4]自然资源部海洋遥测技术创新中心,山东青岛266061 [5]吉林省空间光电技术重点实验室,吉林长春130022 [6]哈尔滨工业大学,黑龙江哈尔滨150001

出　　处：《海洋科学》2024年第1期11-24,共14页Marine Sciences

基　　金：国家自然科学基金重大项目(61890964);山东省联合基金项目(U1906217);山东省自然科学基金青年基金项目(ZR2022QD075)。

摘　　要：海上溢油事件发生后需要及时判断溢油种类,为溢油事故处置提供及时有效的应对措施。本文利用BRDF多角度测量装置和可见光偏振相机对不同油种和清洁海水进行观测,基于可见光偏振特征筛选出最佳观测Stokes参量和对应观测几何,构建了清洁海水和不同油种数据集,并建立了基于BRDF偏振特征的DBN溢油油种识别模型。研究发现,在实验室可控条件下,相对方位角–60°、0°、–30°组合时油种识别效果较好,光源天顶角和相机观测角二者在接近±50°时识别效果较好;多次重复实验显示,海水、原油、重油、汽油、柴油、棕榈油的最优识别率分别为90%、86.27%、84%、80.44%、82.08%和82%;基于可见光偏振特征所建立的油种识别模型,对于海面溢油种类区分具有较好的效果。The oil spill type needs to be promptly identified after the occurrence of a marine oil spill to provide timely and effective response measures for its disposal.The BRDF multi-angle measurement device and visible polarization camera were used to observe clean seawater and different oil types.The best observation Stokes parameters and corresponding observation geometries were selected based on the visible polarization characteristics.The datasets of clean seawater and different oil types were constructed,and the DBN oil spill–type identification model was established based on the BRDF visible polarization characteristics.Under controlled conditions in a laboratory,the combination of relative azimuth angles of−60°,0°,and−30°produces better oil identification results,while the light source zenith angle and camera observation angle have better recognition effects when approaching±50°.Multiple repetitions of experiments showed that the optimal recognition rates for seawater,crude oil,heavy oil,gasoline,diesel oil,and palm oil were 90%,86.27%,84%,80.44%,82.08%,and 82%,respectively.The experimental results show that the oil-type recognition model based on the visible polarization characteristics has a good effect on the oil spill–type identification on the sea surface.

关 键 词：油种识别 双向反射分布函数(BRDF) 偏振 观测几何 偏振度 

分 类 号：P76[天文地球—海洋科学]