Developing visual functions for smart camera network
Cheng-Yuan Tang, Yi-Leh Wu, Kai-Hsuan Chan, Maw-Kae Hor, Whei-Wen Cheng · 2016
Developing visual functions in camera networks from multiple cameras is hot and challenging research topic. Based on the visual technologies including computer vision and image processing, in this paper, we developed some visual functions for smart camera networks. In our current progress, the visual functions developed for the camera network consist of four main functions. These four functions are listed in the following: (i) camera model and camera coordinate transformation; (ii) point correspondence and computing 3D; (iii) visual tracking and recognition; (iv) multiple view geometry and parameter estimation. In this paper, the problems of multiple views are divided into four sub-problems, such as (a) one-view, (b) two-view, (c) three-view and (d) N-view (N>3). Some contents of geometrical properties for these four sub-problems are described and implemented. In this paper, a prototype system for camera networks using visual functions has been developed. Some preliminary results on visual functions developed for the camera network are shown.