Research On The Algorithm Of 3D Road Alignment Identification Based On Remote Sensing Image

As the most important element information of the map,the road alignment parameters are of great significance for the update of road network information and the safety of automobile driving.Therefore,it is necessary to identify the road alignment,and obtain linear parameters that meet the accuracy.Applying the road information in actual production and life.However,at present,there are problems such as difficulty in obtaining road alignment data and excessively rough line recognition.In this paper,the following research work is done on the problems of road alignment recognition:Firstly,considering the time-consuming and labor-intensive acquisition of the linear road data by traditional measurement and the difficulty of obtaining data due to the high cost of other acquisition methods.This paper proposes the experimental data of using remote sensing image extraction of linear road vector data as planar linear shape recognition;Combined with the oblique photogrammetry 3D model to collect 3D road data for profile linear recognition,so as to achieve 3D road alignment recognition.Among them,the planar linear data based on remote sensing imagery is obtained by using a single spectral information and obvious shape characteristics of the road on the high image,and this is used as a classification index.The random forest is used as the classifier to extract the road alignment,and the road centerline is extracted in combination with the multivariate adaptive regression algorithm to obtain the linear vector point data.The road coordinate data identified by the profile line is obtained by using the 3D model processing software for stereoscopic acquisition of the 3D modelSecondly,in view of the problem that the current line recognition is too rough,the obtained line parameters are not comprehensive,and the accuracy needs to be improved,this paper combines the road design specifications to improve the Hough transformation algorithm and combine the IGGIII resistance estimation to calculate the road alignment parameters,and then realize the three-dimensional road alignment recognition of the flat and longitudinal linear shape.By comparing the design road data parameters with the experimental parameters,the feasibility of the road identification algorithm in this paper is verified.In this paper,the research content of linear pattern recognition is divided into two parts,the first part is the detection of linear data,the Hough algorithm is used to detect linear shapes,and the unbiased linear data is obtained by combining with IGIII for differential resistance.The second part is the solution of road alignment parameters and the overall adjustment of road alignment,which uses the linear condition of the road to constrain the adjustment and obtain the linear parameters that meet the accuracy requirements.In the line shape recognition,the flat and longitudinal line recognition methods are the same,the difference is that different detection algorithms are used to detect the line shape,the former uses the straight line Hough algorithm and the round curve Hough algorithm for detection,and the latter uses the straight line Hough algorithm and the parabolic Hough algorithm.Therefore,the research content of this paper is summarized as extracting road coordinate data from remote sensing images and oblique photogrammetry 3D models,and then using Hough algorithm and combined with IGIII resistance to detect linear shapes,and finally performing the overall adjustment of road alignment parameters,and verifying the feasibility of road alignment recognition algorithm through comparative experiments,and then realizing three-dimensional road alignment recognition.