Novel camera calibration based on cooperative target in attitude measurement

In order to solve the problem of camera distortion in attitude measurement, a novel camera calibration method used for pose measurement is proposed. The predominant advantage of this technique is that it doesn't need to know either any movement information with regard to the model plane, or the correspondence between the model plane and images. The proposed calibration technique only needs to observe the calibration board at different orientations that unknown to observers, the function mapping between multiple images is established, thus seven of camera intrinsic parameters and extrinsic parameters can be subsequently computed in the first step. Then rotation and centroid of each feature are detected, and all camera parameters are refined by Maximum likelihood estimation criterion based on Levenberg-Marquardt algorithm. Finally, simulation and experiment results are experimented simultaneously. Results show that this model can effectively avoid the degradation of the image in the imaging process, and also ensure the accuracy of the calibration. The proposed calibration method is quite efficient to improve measurement accuracy and the accuracy of image positioning can be reach to 0.03 pixels, which represents the proposed method has comparable accuracy and reliability. Reconstruction error of 6 DOF doesn't exceed 1%. This technique is very useful for high accuracy solution of the cooperative target pose measurement compared with the existing method.