Abstract


AR systems use see-through HMDs that superimpose virtual objects upon the wearer¡Çs view of real world. It offers great possibilities to support NPP maintenance work. Using AR technique, 3D position can be indicated more intuitively, so field work can be more accurate and efficient and human errors can be reduced. Moreover, it can avoid workers entering dangerous region. There are still some problems for practical application of AR in real NPP environment For example, recognize and measure features in long distance and ensure the high tracking accuracy. For the application of augmented reality in plant maintenance work, real-time tracking and measuring technology with higher accuracy must be developed. Because of the complexity of the NPP environment, this study focused on the tracking method using cameras and paper-based markers. In practical application, a great number of markers maybe need to be pasted in the plant environment. And we have to measure the three-dimensional positions of these markers. When number of the markers is large, it takes much time to measure the positions of the markers. To solve this problem, we develop an automatic marker registration system. this system is composed of a camera, a laser rangefinder, two motion bases which fixed on tripod, (One is for camera, the other is for laser) and a PC connected to them. Circular fiducial marker was adopted here [1]. The relative position between camera and laser is fixed in advanced. Markers are placed at fiducial points (to build the world ordinate system) and other appropriate positions in the environment. After system starting, markers are recognized by recognition process through camera at first, and their positions are estimated. Then the system drives the laser to measure the 3D-position of each marker automatically. Based on the estimated positions, the measure process is high efficient. Finally the data of position is stored in the computer. Experimental evaluation was taken in Fugen nuclear plant following above steps by special people who were familiar with the marker measure process but never used this system before. Although some other problems are to be solved, this system is expected to be useful for supporting fieldwork.

Powered by JDBS Ver.2.1.0
(C) 2020 Hirotake Ishii