STUDY OF THE VISUAL SUPPORT FOR NUCLEAR POWER PLANT DISMANTLEMENT WORK USING AUGMENTED REALITY TECHNOLOGY

It takes enormous labor and cost in the dismantlement works of nuclear power plants because their institutions/facilities are built complicatedly and firmly so as not to leak radioactive rays at all. Supporting the workers leads to solving the problem above since most of the dismantling is performed by hand. Therefore it is effective for the work support to display information about the work procedures or the pollution areas on the being dismantled object visually. We suggest a systemized method that performs the position posture estimate of the object using a characteristic around the natural feature point effectively and overlaps the 3D model. By evaluation experiments, we confirmed that the system could recognize the object from various courses and could track an object with a complicated shape robustly.