Unmanned Aerial Vehicle (uav) Television Image Fast Positioning Method Research

Rapid positioning used UAV is a method based on TV image, which comprehensively used the resources from the UAV to get the coordinates of target quickly and accurately.This paper analyses the feature of UAV TV image, interior orientation, absolute orientation and nuclear line image generation. In order to extract coordinate of UAV TV image, we established the relationship between image coordinates, scanning coordinate System and the ground coordinate.Feature extraction is an important part of image analysis and image matching. In order to identify significant ground target point and obtain the important information of the target points, this paper used Moravec algorithm and Forstner algorithm to extract the characteristics of point about images because the gray surface on image is discontinuous changing. Taking the images of UAV TV was easily affected by all kinds of errors, the gray of image could change. In the process of image matching, the related functions and the system deformation parameters were introduced, which form the least squares image matching system.When constructing the digital elevation models, method of automation data collection was used and method of different DEM data interpolation was selected according to different terrain after comparing several DEM data sampling method. In order to improve the quality of data collection, the method of grossing grid data error detection and exclusion based on gradient information was selected.In order to correct the geometric distortion of the TV images caused by the changing of camera orientation, pitch angle,and focal length, the coarse correction and fine correction has been used acording to the image characteristics. When correcting the generation of image, considering the geometric accuracy of the resampling and the computing amount of the sampling and generating the corrected image by using a double-bilinear resampling method.Finally, according to the relationship between the television image established by the geometric correction and the corrected image, also based on knowing more than three control points, the digital target-location models based on the geometric correction have been established; moreover, the indirect target-location method was selected when the accuracy of the geometric correction can not meet with the requirements.