Abstract
It is difficult to implement camera calibration directly on the embedded platform which has limited capacity where there are a lot of image feature data to process and complex computation to execute during the camera calibration. A simple and efficient camera calibration method is proposed based on the camera imaging model, in which the relationships between the Euclidean three-dimensional points and the planar points of 2D images are established by means of the coordinate translation and rotation, and the method decreases the complexity of the calibration algorithm because it is based on 2D feature points. Finally, a camera calibration system is constructed to get camera parameters on the hardware platform of the Tiny6410, a typical embedded system platform, and the accuracy of calibration results is verified by re-projection. Test results showed that the method is convenient and has high accuracy.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have