Efficient Video Stitching for an Eagle-Eye-Based Vision System

Computer vision has always been one of the research hotspots in the field of artificial intelligence. Inspired by the characteristics of eagle eye, this paper first designs a bionic visual imaging module that contains two wide-angle cameras and one telephoto camera. Then, video stitching and video fusion are accomplished by exploring the fact that relative poses of these three cameras keep constant. Results of extensive experiments show that the eagle-eye-based vision system can acquire large field of view and local high-resolution images simultaneously, which lays a good foundation for subsequent vision tasks.