Study And Realize On Geometric Calibration Of The Single-ccd With Four Bands Multi-spectral Camera

We must first know the inner orientation elements and lens optical distortion parameters of the imaging sensor when use aerospace remote-sensing image mapping topographic maps or to achieve the geographical coding of remote sensing images, the inner orientation element of sensor include the principal point position and the sensor principal distance, they have a significant impact on positioning accuracy of remote sensing images, Therefore, aerial remote sensing data must go through some processing to eliminate distortion in order to truly reflect the geometry features and accurate location information, the calibration of camera parameter is a key problem to guarantee the data quality could meet precision request.This paper first proposed a vision to construct three-dimensional calibration field for non-visible band remote sensing imager calibration, research and experiment on the quality of non-visible band image, and proposed an algorithm to extract the center of the control points at sub-pixel level based on gray value. Main contents of this article is the camera geometric calibration, firstly make use of direct linear transformation method to derive the initial value, and then use space resection with single-image and multi-image methods to iterative calibration the value of the internal and external orientation and distortion parameters of the CCD camera, and the last use a method through compute the object and pixel coordinate each other to verify and analyze the accuracy of the camera calibration result. Through analyze the calibration results and accuracy of the camera, make a conclusion that multi-image space resection is superior to single-image.Another main content of this paper is image geometric correction, described the general process of image geometric correction firstly, and correct calibration field images use the calibration of camera parameter, eliminate deformation non-linear distortion effectively, Through the way of test again and again to get the best camera distortion model, and then combined the exterior orientation elements of aviation images provided by POS to take a orthorectification process.