Depth Reasoning System from a 2-D Image and a Proximity Sensor.
Kazuo Yamaba, Hiroshi Tominaga, Tatsuya Nakamura, Yoichi Miyake · TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C · 1993
In robotics research there is an increasing number of papers on stereo or binocular vision using two or more cameras combined with artificial intelligence methods. This paper deals with a newly developed depth reasoning system from a two-dimensional (2-D) image with a single view and from a proximity sensor. The system consists of only one camera, an image processor and the proximity sensor. The proximity sensor is constructed of a pair of sending/receiving ultrasonic sensors. Several basic experiments using checkered three dimensional objects and a new type of depth calculation were carried out using sensor fusion technology. A new shape of the membership function was discussed on the basis of two equations using a modified implicit function. As a result, the relationship between the measured distance value and reasoned depth value was found to be linear within a wide range.