High-speed 3D modeling for nuclear reactor environment based on feature extraction results from video images (Contract research); FY2023 Nuclear Energy Science & Technology and Human Resource Development Project

Collaborative Laboratories for Advanced Decommissioning Science; Sapporo University*

JAEA-Review 2025-033, 71 Pages, 2025/11

JAEA-Review-2025-033.pdf:4.48MB

The Collaborative Laboratories for Advanced Decommissioning Science (CLADS), Japan Atomic Energy Agency (JAEA), had been conducting the Nuclear Energy Science & Technology and Human Resource Development Project (hereafter referred to "the Project") in FY2023. The Project aims to contribute to solving problems in the nuclear energy field represented by the decommissioning of the Fukushima Daiichi Nuclear Power Station (1F), Tokyo Electric Power Company Holdings, Inc. (TEPCO). For this purpose, intelligence was collected from all over the world, and basic research and human resource development were promoted by closely integrating/collaborating knowledge and experiences in various fields beyond the barrier of conventional organizations and research fields. The sponsor of the Project was moved from the Ministry of Education, Culture, Sports, Science and Technology to JAEA since the newly adopted proposals in FY2018. On this occasion, JAEA constructed a new research system where JAEA-academia collaboration is reinforced and medium-to-long term research/development and human resource development contributing to the decommissioning are stably and consecutively implemented. Among the adopted proposals in FY2023, this report summarizes the research results of the "High-speed 3D modeling for nuclear reactor environment based on feature extraction results from video images" conducted in FY2023. The present study aims to develop a 3D model for a workspace that maximizes the amount of information based on the features extracted from video, which is taken when surveying the primary containment vessel and inside the reactor building as part of the decommissioning of 1F, considering within a specified time. In FY2023, we verified extracting effective shooting conditions for obtaining 3D reconstruction based on photogrammetry and the method extracting feature values that can generate 3D restoration results from a small amount of data within a specified time based on deep learning. In addition, we applied point cloud data extracted from video to segmentation and classified it into parts with instance labels.

A Study on the effects of photogrammetry by the camera angle of view using computer simulation

Nakamura, Keita; Hanari, Toshihide; Matsumoto, Taku; Kawabata, Kuniaki; Yashiro, Hiroshi*

Journal of Robotics and Mechatronics, 36(1), p.115 - 124, 2024/02

https://doi.org/10.20965/jrm.2024.p0115

Visualization of radioactive substances using a freely moving gamma-ray imager based on Structure from Motion

Sato, Yuki; Minemoto, Kojiro*; Nemoto, Makoto*; Torii, Tatsuo

Journal of Nuclear Engineering and Radiation Science, 7(4), p.042003_1 - 042003_12, 2021/10

https://doi.org/10.1115/1.4048843

3D position and radioactivity estimation of radiation source by a simple directional radiation detector combined with structure from motion

Sato, Yuki; Minemoto, Kojiro*; Nemoto, Makoto*

Radiation Measurements, 142, p.106557_1 - 106557_6, 2021/03

https://doi.org/10.1016/j.radmeas.2021.106557

3D modeling for grasping the internal state of the nuclear reactor in decommissioning; Effects and evaluation of image quality to 3D reconstruction accuracy

Hanari, Toshihide; Imabuchi, Takashi; Kawabata, Kuniaki

no journal, , 

This paper proposes to introduce a quantitative image quality assessment of 3D modeling to grasp the internal state of the nuclear reactor for the decommissioning of the Fukushima Daiichi Nuclear Power Station. We attempted to perform a feasibility investigation of a quantitative evaluation of image sequences as an index of 3D reconstruction accuracy. As a result, we confirmed that brightness and contrast parts in quantitative evaluation scores of image sequences were correlated with the reconstruction accuracy of 3D models.

Robust image selection method based on multi-modal detection for efficient 3D reconstruction from images

Hanari, Toshihide; Kawabata, Kuniaki; Imabuchi, Takashi

no journal, , 

This presentation describes the results of 3D reconstruction based on images, introducing a robust image selection method by multi-modal detection for a remote operation on decommissioning tasks at the Fukushima Daiichi Nuclear Power Station. We have been developing a 3D reconstruction method based on images to improve the spatial awareness of the robot operators during the decommissioning tasks. For the display of 3D reconstruction models on demand of the operators, a reduction of the computational cost of 3D reconstruction is required. Thus, we tried to introduce an image selection process in the 3D reconstruction. We performed a verification of the 3D reconstruction introducing the proposed method to the image sequence. The results suggest that adequate images can be selected from the image sequence to reduce the computational cost of the 3D reconstruction, and the speed-up of the 3D reconstruction processes can be achieved for displaying the 3D model on demand of the operators.

Integration of multiple local point clouds based on estimated parameters in photogrammetry with QR code

Nakamura, Keita*; Baba, Keita*; Watanobe, Yutaka*; Matsumoto, Taku; Hanari, Toshihide; Kawabata, Kuniaki

no journal, , 

This study proposes a method for integrating reconstructed models by partial-to-partial registration using photogrammetry reconstructed models and QR codes. It has been considered difficult to integrate photogrammetry reconstructed models because the scale of each reconstructed model is different each time. In this study, we solve this problem by placing QR codes of known size in the environment for reconstruction and scaling each reconstructed model based on the size of the QR code. To verify this method, we compared the accuracy of the integrated model with that of the reconstructed model from all images. The comparison results show that a tolerance of 20 mm is highly accurate. We consider that this approach will be effective in reducing the time required for mapping using robotic and photogrammetric methods.