Image quality improvement for Fukushima Daiichi remote operations using denoising prior to super resolution

Tanifuji, Yuta; Imabuchi, Takashi; Kawabata, Kuniaki

Proceedings of the 31st International Symposium on Artificial Life and Robotics (AROB 31st 2026), p.1011 - 1016, 2026/01

In decommissioning work at the Fukushima Daiichi Nuclear Power Station (1F), operators must rely on robot camera images degraded by turbidity, floating matter, lens contamination, and radiation induced noise, with no clean reference images available. This study investigates a lightweight restoration pipeline combining Noise2Noise (N2N) denoising and FSRCNN super resolution, and shows that the N2NFSRCNN configuration most consistently improves visibility while suppressing noise and artificial textures, according to NIQE and PIQE scores and subjective evaluation.

DX platform for evaluating radiation and environment perturbation; An Application to Fukushima Daiichi decommissioning

Nakashima, Shinsuke*; Wu, J.*; Hanari, Toshihide; Imabuchi, Takashi; Matsuhira, Nobuto*; Kawabata, Kuniaki; An, Q.*; Yamashita, Atsuhi*

Proceedings of IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR2025), p.127 - 132, 2025/10

https://doi.org/10.1109/SSRR68451.2025.11391292

Discrimination of structures in plant using deep learning models trained by 3D CAD semantics

Imabuchi, Takashi; Kawabata, Kuniaki

Artificial Life and Robotics, 30(1), p.184 - 195, 2025/02

https://doi.org/10.1007/s10015-024-00989-w

Image selection method from image sequence to improve computational efficiency of 3D reconstruction; Application of fixed threshold to remove redundant images

Hanari, Toshihide; Nakamura, Keita*; Imabuchi, Takashi; Kawabata, Kuniaki

Journal of Robotics and Mechatronics, 36(6), p.1537 - 1549, 2024/12

https://doi.org/10.20965/jrm.2024.p1537

This paper describes three-dimensional (3D) reconstruction processes introducing the image selection method for efficiently generating a 3D model from an image sequence. To obtain suitable images for efficient 3D reconstruction, we tried to apply the image selection method to remove the redundant images in the image sequence. By the proposed method, the suitable images were selected from the image sequence based on optical flow measures and a fixed threshold. As a result, the proposed method can reduce the computational cost for the 3D reconstruction processes based on the image sequence acquired by the camera. We confirmed that the computational cost of the 3D reconstruction processes can reduce while keeping the 3D reconstruction accuracy at a constant level.

Measuring unit for synchronously collecting air dose rate and measurement position

Kawabata, Kuniaki; Imabuchi, Takashi; Shirasaki, Norihito*; Suzuki, Soichiro; Ito, Rintaro; Aoki, Yuto; Omori, Takazumi

ROBOMECH Journal (Internet), 11, p.11_1 - 11_11, 2024/09

https://doi.org/10.1186/s40648-024-00279-x

Discrimination of Plant Structures in 3D Point Cloud Through Back-Projection of Labels Derived from 2D Semantic Segmentation

Imabuchi, Takashi; Kawabata, Kuniaki

Journal of Robotics and Mechatronics, 36(1), p.63 - 70, 2024/02

https://doi.org/10.20965/jrm.2024.p0063

Validation of the Be ground-state molecular structure using Be()He triple differential reaction cross-section measurements

Li, P. J.*; Beaumel, D.*; Lee, J.*; Assi, M.*; Chen, S.*; Franchoo, S.*; Gibelin, J.*; Hammache, F.*; Harada, T.*; Kanada-En'yo, Yoshiko*; et al.

Physical Review Letters, 131(21), p.212501_1 - 212501_7, 2023/11

https://doi.org/10.1103/PhysRevLett.131.212501

The cluster structure of the neutron-rich isotope Be has been probed via the () reaction. The triple differential cross-section was extracted and compared to distorted-wave impulse approximation reaction calculations performed in a microscopic framework using the Tohsaki-Horiuchi-Schuck-Rpke wave function and the wave function deduced from Antisymmetrized Molecular Dynamics calculations. The remarkable agreement between calculated and measured cross-sections in both shape and magnitude validates the description of the Be ground-state as a rather compact nuclear molecule.

A Study on generalization capability of trained structure discrimination network based on 3D point cloud

Imabuchi, Takashi; Kawabata, Kuniaki

Proceedings of 20th International Conference on Ubiquitous Robots (UR 2023), p.632 - 633, 2023/06

Discrimination of structures in a plant facility based on projected image created from colored 3D point cloud data

Imabuchi, Takashi; Kawabata, Kuniaki

Proceedings of 2023 IEEE/SICE International Symposium on System Integration (SII 2023) (Internet), p.396 - 400, 2023/01

https://doi.org/10.1109/SII55687.2023.10039472

Discrimination of the structures in nuclear facility by deep learning based on 3D point cloud data

Imabuchi, Takashi; Tanifuji, Yuta; Kawabata, Kuniaki

Proceedings of 2022 IEEE/SICE International Symposium on System Integration (SII 2022) (Internet), p.1036 - 1040, 2022/01

https://doi.org/10.1109/SII52469.2022.9708845

This paper describes a method for discrimination of the structures in nuclear power station by deep learning based on 3D point cloud data. In order to promote safe and steady decommissioning work, it is important to estimate and assume the condition in nuclear power station based on the measured sensor data. Especially, the data of the dose rate in the workspace is useful to plan the decommissioning task and, the shape and the material property of the structures in the workspace are required for the dose rate simulation. Shape data can be obtained by such as 3D Scan, however, it is difficult to acquire the material property data of the objects. Therefore, we consider that it is possible that the major material property can be estimated from the category of the structures in nuclear power station. In this paper, we proposed a structure discrimination method by 3D semantic segmentation with 3D point cloud data that consists of labeled points by referring category labels of CAD data of existing nuclear facility. We reported discrimination performance of the proposed method by hold-out validation.