Research On 3D Visualization Technology Of Structured Light In Radiation Environment

The Party’s 20th National Congress Report and the Fourteenth Five-Year Plan have pointed out the future development direction of nuclear energy in China,and clearly proposed that China should actively develop nuclear power in a safe and orderly manner.With the continuous development of nuclear power technology and the occurrence of nuclear power plant leaks,the issue of nuclear accident management has attracted extensive attention from scholars.After the nuclear leakage accident,industrial equipment at the accident site needs to be monitored,but humans cannot monitor the equipment at close range due to the lethal hazard of radiation energetic particles.In this thesis,we propose the study of structured light 3D visualization technology in radiation environment,which can monitor the equipment with high precision by using a simple system of projector and camera.In order to improve the reconstruction accuracy and reduce the exposure time of image acquisition equipment in radiation environment,this thesis focuses on radiation image filtering and unreliable phase rejection as follows:1.The introduction of the mathematical model of the monocular structured light system involves an analysis of the system’s calibration principle and method,and the optimization and enhancement of the calibration process,resulting in a 20.6%reduction in image acquisition time.Further analysis indicates that the primary source of error in the 3D reconstruction of the structured light system in a radiation environment stems from the impact of radiation noise,as well as the generation of unreliable phases due to shadow and background areas.2.To overcome the challenge of radiation noise that degrades image quality,we propose a radiation image composite filtering algorithm that combines median filtering and multi-frame averaging filtering,leveraging the radiation noise characteristics.In comparison with various common filtering algorithms,this algorithm achieves the highest peak signal-to-noise ratio and structural similarity ratio.Furthermore,the peak signal-to-noise ratios of filtered images at 23.5 Gy/h and 49.8 Gy/h radiation dose rates reach46.473 and 41.769,respectively,while the structural similarity ratios are 0.982 and 0.962.In contrast to the multi-frame averaging filtering method,the algorithm significantly reduces the number of frames required by 67.647%and 76.087%,respectively.Experimental results demonstrate that this algorithm effectively eliminates radiation noise and is more appropriate for image filtering in a radiation environment.3.To overcome the challenge of unreliable phases generated in shadow and background areas,we conducted an analysis of the various types of unreliable phases and the reasons for their generation.This thesis proposes the use of a modulation system to determine the reliability of phases,and puts forth an adaptive unreliable phase rejection method based on the modulation system.Through a comparison of absolute phase maps and 3D point clouds generated by common methods,we have verified the algorithm’s ability to efficiently and accurately reject unreliable phases in complex scenes.4.The radiation images were collected from the ⁶⁰Co irradiation chamber of the China Institute of Atomic Energy.The maximum error of the reconstructed model at 23.5Gy/h was 1.53 mm,the minimum error was 0.11 mm,and the average relative error was1.92%.At 49.8 Gy/h,the maximum error is 3.20 mm,the minimum error is 0.24 mm,and the average relative error is 2.49%.The reconstructed model accuracy has high precision,which verifies the feasibility of the work in this thesis.