Implementation and Evaluation of a Wide-Range Human-Sensing System Based on Cooperating Multiple Range Image Sensors

A museum is an important place for science education for children. The learning method in the museum is reading exhibits and explanations. Museums are investing efforts to quantify interests using questionnaires and sensors to improve their exhibitions and explanations. Therefore, even in places where many people gather, such as in museums, it is necessary to quantify people's interest by sensing behavior of multiple people. However, this has not yet been realized. We aim to quantify the interest by sensing a wide range of human behavior for multiple people by coordinating multiple noncontact sensors. When coordinating multiple sensors, the coordinates and the time of each sensor differ. To solve these problems, coordinates were transformed using a simultaneous transformation matrix and time synchronization was performed using unified time. The effectiveness of this proposal was verified through experimental evaluation. Furthermore, we evaluated the actual museum content. In this paper, we describe the proposed method and the results of the evaluation experiment.