Railway Extraction Based On LiDAR Data

As a comprehensive transportation means,railway plays an important role in the national economic development.The railway is also the significant rail infrastructure element for surveying.But there are many shortcomings in current railway surveying,such as long measurement period and low efficiency,which are difficult to meet the requirements for the fast accurate extraction of the thematic railway elements.Therefore,it is urgent to research a new method of railway track extraction,and it is also a problem to be solved in mapping geographic information industry.The LiDAR(Light Detection and Ranging)system has such merits: quick data acquisition,comprehensive access to information,high measurement accuracy and high automation degree,which provides a good foundation to extract the railway.By using the comprehensive high-resolution remote sensing images,this paper realizes the automatic extraction of railway track based on the different characteristics of Airborne and Vehicle-borne data.These main contents are as follows:(1)On the basic of data fusion of airborne Li DAR point cloud and images,the spectrum and gray and three-dimensional feature information are used to extract the railway.And according to the smooth embankment and the local topography of the railway,the paper realizes the automatic recognition and extraction of the railway area.(2)The paper uses Gaussian smoothing differential geometry and linear feature extraction algorithm to extract the railway line based on high-resolution remote image.Then,establish rail parallel lines algorithm model based on the noise cancellation and the connected line segments and detect the track accurately.(3)For high-density airborne LiDAR data,on the basis of obtaining the railway mask region,we well-define the rail constraints and classify railway point clouds.Finally,we fit the rail line by the least squares method.(4)For the character of vehicle-borne Li DAR,the paper builds the railway track buffer by using the corrected GPS centerline.According to the shape characteristic of the track cross-section,the paper constructs template matching operator which is suitable for the extraction of the track and automatically extract the track based on the filtered vehicle-borne data.