Research On The Precision Of UAV Aerial Photography System

At present,with its high efficiency,short update period in all fields,fast access to spatial information and many other advantages,the UAV aerial photogrammetry technology has been widely used in large scale topographic mapping,environmental monitoring,disaster relief and military activities in a number of aspects.Compared with large scale digital mapping,the plane precision of UAV aerial photography system is low,and it is difficult to meet the precision requirement of large scale mapping due to the influence of different factors,such as hardware,weather and terrain conditions.This paper expounds on the UAV system and the aerial image quality factors and mapping of key technologies,and to the experimental area as an example,to explore the UAV system and methods to improve the mapping accuracy of digital elevation precision products.The main contents are as follows.1.The whole process of UAV aerial photography system is discussed.It includes the structure of UAV aerial photography system,the coordinate reference system,the factors affecting the aerial image quality of UAV and the corresponding countermeasures.This paper discusses the key technologies of UAV aerial photography system,including the external control of the image control point,the planning and design of the belt line,the pretreatment of aerial image and the aerial triangulation.2.Based on the experiment of the test area,from image overlap,image rotation angle,inclination angle and curvature and maintain high flight routes and test quality of the aerial image,and the empty three encryption processing industry image in the precision is analyzed.In addition,the accuracy level of the generated DOM is discussed,and the accuracy of the UAV aerial photography system is tested in many aspects.3.In order to solve the problem of low elevation precision of UAV aerial photography system,an optimal non-negative variable weight combination model based on three kinds of models,that is,planar fitting,quadric surface fitting and GA-BP neural network,is proposed.The model uses the difference between the actual measured value of the elevation point and the measured value in spatial model to fit the correction value closest to the actual difference at the checkpoint,so as to improve the elevation accuracy of the image.Experimental results show that the optimal non-negative variable weight combination model is better than the fitting effect of the single model and classical weight combination model,the elevation correction fitting residual maximum and minimum value and the error of mean squares is the smallest,the fitting residual is in a small range and the fitting results is the most stable.The accuracy of the image elevation is improved by 0.913 times after correction by the optimal non-negative variable weight combination model,which shows that the model is feasible and effective to improve the accuracy of image elevation.