Uav Remote Sensing Image Application Research On Aerial Triangulation

As UAV technology matures, a growing number of UAV aerial applied.Compared with traditional satellite remote sensing to obtain data, unmanned aerialvehicles due to their own advantage can be quickly and efficiently respond to naturaldisasters and other emergencies, the small cost of a regional high-resolution images tofacilitate decision-making. UAV image capture characteristics of the image obtainedhigher degree of overlap, a large tilt angle, ordinary digital camera its own distortion, sothe UAV to get the data has its own new problems and difficulties in the post-processing.In order to solve these new problems that exist in the UAV image, this paper studied theUAV remote sensing image processing in the following:1、 A comprehensive summary of the history of development of photogrammetryand the POS system application on the basis of UAV aerial photography, the use of POSlocation information for aerial data quality checks, the flight strip to identify, to thedegree of overlap check for heading next judgment Airlines flying data quality andfollow-up orientation, aerial triangulation processing and so on.2、 Fast and accurate matching algorithms, optimized and improved by adding twoconstraints on the basis of the SIFT algorithm, coarse matching using SIFT algorithm;and then delete the "many" and "many" false match point, leaving only the "one-to-onematch point; match point connection between the slope roughly equal to that condition,then by virtue of the same name and slope constraints to remove the error, and improvethe matching accuracy; block limit the shape of the matching point distributed control,makes the same name as the matching points roughly evenly distributed, this canincrease the accuracy of the follow-up relative orientation of subsequent processing.3、 Researchers used continuous like the relative orientation of the basicprinciples, matching earlier on the basis of the degree of overlap with the same namethree or more matching points for the global computing model connected with theconditions, using the least squares method Equation of France, and completed therelative orientation. 4、 On the basis of the traditional mainframe image digital aerial triangulationattempt to explore the UAV remote sensing image bundle aerotriangulation. The initialvalue of the bundle block adjustment according to the absolute orientation elements sumcalculated in the relative orientation of directional elements and strips level Solutions,using the least squares normal equation in the linearized loop solution. Generatethree-dimensional simulation of the pixel map, the last positive and negative countorthorectification, and finally get the image mosaic. And well-known softwareEARDAS LPS (Digital Photogrammetry) module for the accuracy comparison of theexperimental results, the accuracy of this article bundle adjustment algorithm to meetthe mapping requiremen.