Research On Road Extraction Based On Airborne LiDAR Point Cloud

As the lifeblood of the national economy, road plays a pivotal role. With swift development of economy, the road traffic conditions change with each passing day. So how to get the information of road timely, accurately and efficiently become the most urgent desire of people. Airborne LiDAR system shows the tremendous development potential with its ability to obtain3D information. At present, the research of airborne LiDAR technology focuses on point clouds filtering, building extracting and rebuilding, besides improving the function of system and adjusting the precision of point clouds. But we can see, most of the filtrations are aiming at smooth city territory, no matter what kind of algorithm, it will express a nice effect; the research of road extracting much less than filtering and building extracting, and most of them processing with remote sensing image, but if the images are not being got at the same time with point clouds, the deviation will be hard to be avoid in the process of rectification, and then it will affect the final precision, with the development of airborne LiDAR technology, the point clouds with spectral information may be obtained. In consideration of the factors above, this study presented a research of uneven terrain road extracting based on airborne LiDAR point clouds. The major contents and conclusions are as follows:(1) Summarized the status of airborne LiDAR system and technology development at home and abroad, introduced the composition, working theory, characteristic of point clouds and application fields of the system.(2) Labored some kinds of classical filtering algorithm at the moment, combined the uneven terrain characteristic of the study area, and used the algorithm of Advance Gradually TIN Encipherment to filter point clouds first. It can reduce the effects of buildings and vegetations in the process of road extracting when we get digital products such as DTM, DEM etc. It can also save time, ensure accuracy in the postprocessing. Filtered error points before building the TIN net, presented a algorithm of Cylindrical Error Filtering, introduced the detail of error filtering process, solved the problem of the former studies simply considered the lowest point belongs to ground made the result of extracting wrong. And then improved the algorithm, reduced iteration times, saved operation time. (3) Based on the ground points that had been extracted, extracted road by points' echo intensity information. Demarcated the extent of road intensity, and then used the Density Control to refine the first road extracting result.(4) Select the study area which is not flat, use the algorithms presented in this paper to get DTM and extract road, compare the filtering result with commercial software TerraScan, and compare the road extract result with the digital map that measured by scale1:500. According to the experiment in this paper, we can see these algorithms' validity, insufficient, and some shortages of airborne LiDAR system that independent research and development by our country.