A Study On The Mathematic Model And Algorithm Of The Geometric Calibration Of Airborne Digital Sensor

The integrated remote sensing system of airborne digital camera and GNSS/IMU unit can acquire digital images, position and attitude data simultaneously, which resolves the primary issue of full digital photogrammetric mapping and becomes the mainstream equipment for airborne photography. A complete and systematic geometric calibration for its imaging parameters and GNSS/IMU mounting parameters is the key procedure to ensure final mapping accuracy.This thesis makes a comprehensive study on the geometric calibration of airborne digital camera. Its aim is to resolve the key algorithms of IMU misalignment angles, camera imaging parameters calibration and digital camera self-calibration, explore the technical methods for airborne digital sensor geometric calibration and draw a few useful conclusions from experimental tests. The main work and innovation of this paper is as follows:1. Different kinds of optical distortions of frame and multi-line array CCD camera are analyzed and summarized in detail, and corresponding additional parameters is set up, which provides theoretical support for the self-calibration adjustment of airborne digital camera.2. According to the equations which transform the GNSS/IMU navigation result to photogrammetric exterior orientation elements the rigorous mathematical model of IMU misalignment angles is derived. The Euler angles, unit quaternion and antisymmetric matrix three approaches have been introduced for two-step IMU misalignment calibration, experimental results show that the rigorous IMU misalignment mathematical model described by Euler angles proposed in this paper has better precision.3. GNSS/IMU supported single image resection and bundle adjustment two methods have been used for the interior orientation parameters calibration of frame CCD camera. Tests show that ideal and stable calibration results can be obtained with the introduction of GNSS/IMU data. If the exterior orientation parameters are measured by GNSS/IMU with high precision, interior orientation calibration with GNSS/IMU may not require ground control points.4. The self-calibration bundle adjustment of frame CCD images is investigated and GNSSS/IMU supported self-calibration bundle adjustment is set up. With the introduction of addition parameters the accuracy of 1 pixel on plane and 0.4 pixels on height can be obtained using DMC images, and corresponding precision promotion is reach to near 40% and 80% respectively.5. Based on low-order polynomial and orientation image interpolation two trajectory models and ETH and Brown these two kinds of additional parameters, the self-calibration bundle adjustment system of airborne three-line array images is established. Comparative tests have proved using multiple groups of ADS40 with different flights, different strips and different ground control point configurations that the precision promotion with self-calibration reaches to over 40%, 60% and 70% in three coordinate components. The conclusion that orientation image is the preferable adjustment model for three-line array images and both ETH and Brown are effective additional parameter models for the geometric calibration of three-line camera is drawn.6. The quaterion is introduced into the bundle adjustment with orientation image method. According to the characteristics of orientation images the simplified Spherical linear interpolation (Slerp) formulas are derived and corresponding bundle adjustment algorithm is proposed. Experimental results prove that this simplified approach can not only obtain equivalent position accuracy compared to traditional Euler angles but also improve the efficiency of error equations construction remarkably, at the same time the defects of Euler angles are avoided, so provides an alternative method for the adjustment of three-line images.