| Summary: | 碩士 === 亞洲大學 === 生物資訊學系碩士班 === 98 === In computer vision, most studies of three-dimensional reconstruction focus on using multiple cameras. However, there are some advantages in using single camera, e.g. lower system complexity, simpler calibration procedure, low cost, and shorter development time. In this study, we propose a new and simple approach for 3D trajectory reconstruction using single camera. In this approach, after calibration, the world coordinates of objects can be retrieved from given information of the object's image coordinates (X,Y) and its geometric area. The depth of objects can be found by a transformation of a quadratic function of object's geometric area and coordinates on image, and thus the trajectory of the moving objects can be reconstructed. This study is useful in practice for low-precision and low-cost 3D reconstruction applications in the sense of performance-cost ratio.
|
|---|
