Research On Underwater Detection Device For Nuclear Fuel Assembly Defects Based On Line Structured Light

The fuel assemblies in nuclear power plants are deployed in harsh areas.With the influence of assembly stress and thermal stress,long-term operation will lead to component deformation or damage,which will increase the probability of nuclear accidents.Therefore,regular inspection of nuclear fuel assemblies is of great importance.A measurement model based on line structured light was proposed and a nuclear fuel component detection device was built in this paper to meet the requirements of related projects of State Nuclear Power Plant Service Company.Based on this device,the calibration,image processing and point cloud processing of the whole system are completed.Firstly,the system calibration methods were studied,including camera calibration,line structured light plane calibration,rotation center calibration and underwater calibration.For camera calibration,a suitable sub-pixel corner detection size was found to meet the minimum re-projection error.A stereoscopic target based on checkerboard is designed to realize the calibration of the line structured light plane.A method of calibrating the rotating center is proposed in this paper for rotational scanning.Finally,the equivalent focal length method is chosen as an underwater calibration method by comparing with experimental calibration and non-linear model calibration and equivalent focal length calibration.Then,an adaptive threshold method based on gradient is proposed for the uneven gray distribution of fuel rod scanning.The extraction of the stripe center was completed by combining the refinement and the gradient center of gravity,effectively reducing the noise from the surface of the fuel rod.The scanned point cloud is preprocessed,and the least square method is used to extract the plane features with less noise.The RANSAC algorithm is used to extract the cylinder axis with more noise.Finally,according to the measurement accuracy requirements,the structure of underwater device was designed,and the hardware selection and assembly were completed.The software development is based on win7 64-bit system by using crossplatform QT5.8.0 and VS2013.Open source OpenCV3.0 and PCL 1.8.0 libraries is used for image processing and point cloud processing respectively,realizing the algorithm and function module development of the whole system.By measuring the step gauge block,the device designed in this paper can meet the requirement of 0.5 mm accuracy at 600 mm.