Study On Key Technology And Application Of Airborne Remote Sensing Auxiliary System

With the development of aerial remote sensing technique, Unmanned AerialVehicles(UAV) technology plays a tremendous role on national defence construction,basic surveying and mapping, earthquake disaster prevention and mitigation and lifeproduction. It has the advantages of low cost, hight precision, flexible, easy to control,and hight security. In the developed conturies as Europe and America, UAVtechnology has moved away from purely military use and made great development inthe field of bussiness. In the domestic, UAV technology is still in the initial stage.Data processing theory is improving stage step by step; software is still behind thedevelopment of hardware; real time is limited. The current lane planning and designmethod remains in the artifical calculation state with low efficiency, difficult editingand taking no account of other geographic information data. While image qualitychecking depends on manul extraction of corresponding points. This manul methodhas low degree of automation, heavy workload and it cannot timely make evaluationwhen the mission ended, especially cannot meet the rapid response to disasters.Airborne remote sensing auxilary system is put forward to meet the fast lane planningand design and fast image quality checking.According to the actual requirements, airborne remote sensing auxilary systemcontains two modules: lane planning and design module and image quality checkingmodule. Referencing to the geographic information data, the first module has twomainly functions to plan and design lanes: based on areas to arbitrary azimuth designand based on linear features to arbitrary azimuth design. According to the relatedstandard of aerial photography, the second module has the functions checkinglongitudinal overlap, side overlap and swing angle and generating reports.Airborne remote sensing auxilary system deponds on the followiing aspects:design and organization of data and file structure, computer vision for display andexpression method, data analysis for basic geographic information, spatial entity dataabstract expression, algorithm for lane design, algorithm for feature matching andmethod of image quality checking. In view of all the above, referencing to theresearch status at home and board, using existing conditions, works are done as follows:1. In the lane planning and design module, point, line, plane and other baiscspatial data structures are designed. Based on these structures, area, lanes, section linestructures with attributes are put forward. At the same time, in order to meet thedesing need, sensor, design parameter and unspatial structures are constructed. Inimage quality checking module, lane and image structures are designed.2. Using GDAL and OpenGL open-source library, this system gives a goodsupport to multi-source geographic information data. When the terrain data is beenprocessing, storage and expression are separated to speed up its display. Optimizingthe terrian data search algorithm, terrian data statistics and calculation are morequickly and accurately.3. According to the design requirements, based on aerial photogrammetricspecification, systems gives two methods to design lanes: based on area and basedlinear features.4. Through studying the algorithms of image feature matching and using OpenCVopen-source library, SURF algorithm is used in this system. The same pair points willbe faster to extract by SURF. Then it will be easier to calculate the relative positionsof each image, and then aerial matching and lanes connecting can be going onactually.5. Through discussing the image quality checking methods, the algorithm uses thesame pair points to calculate the overlapping and rorating is put forward. By detectingaerial longitudinal overlap, side overlap and swing angle to determine whether theflight quality is all right.The innovation of this paper lies in: making full use of the advantage ofopen-source libruary; optimizing the algorithm of searching and statistias of terrainpoints; using textures display geographic information data; giving out the algorithm ofdesigning lanes; putting out a new tree-node sorting method; realizing SURFalgorithm whitch greatly improves the matching speed and precision.The study shows that: using texture, terrain date display efficiency is moresmoothly than rendering point by point, especially in the situation of large amount ofelevation points, texture expression will be more fluency. Compared with thetraditional lane designing method, based on area and based on linear featuresdesigning method will be more flexible, easer to modify and more efficient. SURFalgorithm of image feature matching can quickly extract the same pair points. Takethe flight data of An county in Sichuang as a sample, the whole image quality check processing takes less than half an hour, which mainly achieve the image quality quickcheck.