融合Kinect与γ相机图像的放射性区域重建与定位  

Reconstruction and localization of radioactive area fusing images from Kinect andγcamera

作　　者：王刚 肖宇峰[1] 郑又能 田星皓 WANG Gang;XIAO Yufeng;ZHENG Youneng;TIAN Xinghao(School of Information Engineering,Southwest University of Science and Technology,Mianyang 621010,China;Institute of Applied Nuclear Technology,China Institute of Atomic Energy,Beijing 102413,China)

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010 [2]中国原子能科学研究院核技术应用研究所,北京102413

出　　处：《应用光学》2020年第5期965-972,共8页Journal of Applied Optics

基　　金：国家重点研发计划(2019YFB1310800);国家核能开发项目([2016]1295);四川省重点研发计划(19GJHZ0197)。

摘　　要：针对核设施退役、核应急处置过程中放射性分布信息的可视化需求,提出了一种基于Kinect与γ相机图像信息融合的放射性区域重建与定位方法。首先,基于γ相机的特殊成像方式,构建了Kinect与γ相机组合成像模型,并完成相机组合联合标定;其次,基于视觉地图构建方法,建立了核辐射环境稠密点云地图并得到Kinect位姿;然后,提取γ相机图像中的放射性分布信息,根据相机组合模型计算地图中的放射性区域点云;最后,基于最小包围盒对γ相机成像的中心区域进行三维定位。在实验中,通过将Kinect和γ相机数据同步与空间对齐,在少量γ相机图像的情况下,实现了单个点源的三维分布重建模型与辐射场景地图的融合。在8×12 m^2的实验室环境中点源定位的均方根误差为0.11 m,证明了本文方法的有效性。Aiming at the visualization requirements of radioactive distribution information during the decommissioning of nuclear facilities and nuclear emergency disposal,a reconstruction and localization method of radioactive area fusing the image information from Kinect andγcamera was proposed.First,a combined imaging model of Kinect andγcamera was constructed and its joint calibration was implemented based on the special imaging method ofγcamera.Second,based on the visual map construction method,a dense point cloud map of the radioactive environment was established and the pose of Kinect was obtained.Then,the radioactive distribution information was extracted from theγcamera images,and the point cloud of the radioactive area in the map was calculated according to the camera combination model.Finally,the 3D localization of the central area forγcamera imaging was performed based on the minimum bounding box.In the experiment,by synchronizing the Kinect andγcamera data with spatial alignment,the fusion of 3D distribution reconstruction model and radioactive scene map of the single point source was realized in the case of a fewγcamera images.The root-mean-square error of the point source localization is 0.11 m in a laboratory environment of 8×12 m2,which verifies the effectiveness of the proposed method.

关 键 词：三维重建 定位 信息融合 联合标定 Γ相机 放射性区域 

分 类 号：TN209[电子电信—物理电子学] TP391.41[自动化与计算机技术—计算机应用技术]