Unmanned Aerial Vehicle Remote Sensing Images And Digital Elevation Model Of Three Dimension Visualization Research

Elevation Model of Three Dimension Visualization Research Abstract:In recent years, with the development of remote sensing, GPS navigation, as well as the computer technology, more conveniently and technology integration of UAV remote sensing technology comes naturally. Its application field also expands continually, in the dynamic monitoring; digital construction and disaster emergency domain has a more mature use. For the large UAV image scale and rotation angle, the stitching matching technology is more mature. But, the research for large soale UAV image extraction of high resolution DEM is less relatively, how to build digital visualization model rapidly according to images terrain features is the final application embodiment of the UAV remote sensing image, and is also the premise of its further development in various fields. In this article, we first summarize the current application research status of UAV image in three dimension production. As the characteristics of high resolution and a certain degree of overlap of UAV images, we propose a method of building three dimension digital models.This paper’s research work mainly includes the following several aspects:(1) The research for large scale UAV image extraction of high resolution DEM is less relatively, with the problems such as low resolution and lack of effective quality evaluation. This article selects root mean square error and surface roughness as factors in the evaluation of using the terrain feature points extraction of high resolution DEM data quality to avoid the blindness of DEM data in digital terrain application aspect.(2) Considering the influence of the UAV image terrain elevation difference, from the theory of image distortion caused by error, analyzes the process of UAV image geometric distortion and we main make the correct processing for the error caused by the exterior orientation elements. Using characteristic points of control in combination with polynomial and collinear equation processing methods for different terrain, makes image data in three dimension applications is more ideal.(3) In view of the large amount of data, especially building digital model for large-scale urban region, the modeling is slow. In this article, combining with the traditional block algorithm and triangular mesh generation algorithm, we propose an adaptive threshold block triangulation algorithm, improving the speed of modeling, under the condition of guarantee of the quality also improves the automation level.(4) Combining with the traditional binary tree and the quad tree, we put forward a multi-resolution model based on a viewpoint. According to the distance between the terrains feature points and the observation point, analysis the relationship of resolution of dynamic transformation and use elaboration, simplification in processing suitably according to the terrain complexity then solve the low efficiency of three dimension visual model expression effectively. Application of the model, implement image superposition of two dimension and three dimension visualization simulation of track in baidu maps and Google earth, respectively.