The Research On The Precision Enhancement Of The Large Field Of View Visual Positioning System

In this paper, the aviation positioning accuracy in wide field of view visual positioning system has been analyzed. Due to the characteristics of large field of view,and high noise of the positioning camera system, the camera imaging model obtained by the traditional calibration methods and distortion correction technology cannot meet the measurement accuracy demand. Therefore, key techniques of improving the positioning technology has been examined by analyzing the image model of the positioning the camera model. Meanwhile, this paper also has significance in other position measurement problems with some common features.The main content and key technology of this paper are:1. Establish the constraint-based geometric four features feature point position and attitude measurement model and the pose solving algorithms of feature points through the connection between the conduct according to the design and the large field of vision demands precision positioning system.2. The effect of the camera distortion model on the pose measurement accuracy has been studied. The camera distortion model that is closest to the actual image is chosen according to the measurement error of spatial coordinates of feature points.The noise simulation experiments under different noise environments have illustrated the different anti-noise capability of different inner parameter distortion calibration model, and the positioning measurement experiment proves that a reasonable choice of the distortion model can effectively improve the positioning accuracy.3. On the basis of the selected camera distortion model, the effects of the external factors, that is, the choosing of the feature points on the inner parameter calibration has been studied, in the situation that the full-field imaging camera model cannot be described in a local scale calibration. Further, the impact of the calibration on the space positioning accuracy has been analyzed. The experiments prove that the space factor in calibration procedure has a significant impact on the positioning accuracy.4. A non-modeled imaging camera calibration method based ray tracing through experiments has been studied. Experiment shows that this method can effectively reduce the systematic errors in the calibration process. Thus, it not only has good noise immunity, but also can substantially improve the positioning accuracy,especially in measurement with expansion scope of the attitude angle. The imaging ray tracing model has very clear advantage of accuracy.