Research On Optical Sensor Geometric Calibration Technology

Optical sensor geometric calibration is one of the basic problems of research inphotogrammetry, with the diffusion using of civil UAV in aerial photogrammetry, nonphotometry sensor geometry calibration problem become more prominent, it has a veryimportant role to study. Based on this background, this paper study in the geometric calibrationof optical sensor in the control field factor influence, the geometric calibration model and remotesensing image distortion correction technique. Specific research work mainly includes:1. Completed the research of simulated data generation technology. Designed and achievedthe algorithm and software to generate a set of simulated data, which used in optical sensorcalibration software commissioning and validation. On this basis, further analog sensoraccidental error, provide standard data for subsequent verification algorithm study.2. Completed the research of the factors of optical sensor in the geometric calibrationcontrol field, finally obtained control points distribution program. Because control points have agreat effect on the precision of camera calibration in control field, this paper introducedDLT(Direct Linear Transformation) calibration method, and the correctness of the this method isvalidated with simulated data, and respectively research on the factors which the number ofcontrol points, distribution and relative high in the control field effecting on calibration accuracyof the experiment, through the experiment analysis and comparison, we can obtain control pointsdistribution program.3. Proposed an improved space resection optical sensor geometry calibration method withmulti-images, this method can effectively overcome the problem of correlation in the process ofparameters solution. As parameters correlation have a great effect on its final solution in thetraditional space resection with multi-images, this paper puts forward a solution based ongrouping parameter for multi-images space resection, i.e. grouping the correlation parametersand respectively for a solution. We using simulated data to verify the correctness of the method,and then using measurement data to calculate. Through the experiments, this method is effectiveto overcome the problem of correlation in the process of parameters solution and cansuccessfully work out each calibration parameter.4. Achieved remote sensing image distortion correction technology. On the basis ofobtaining optical sensor calibration parameters, which used in the remote sensing imagecorrection, and applied in actual engineering projects, achieved satisfactory results.