| Due to the urbanization process accelerated, in recent years, road update very quickly.The traditional surveying and mapping methods can’t effectively and timely get roadinformation, and more and more can’t satisfy people’s needs, especially in road navigation.People urgently need a quick precise method to obtain road information. And Airborne laserradar (Airborne Light Detection and Ranging, LiDAR) system can quickly provide a greatdeal of high precision surface mapping data, which makes fast accurate surveying andmapping become possible. With the related technology continues to mature and socialdemand continues to expand, airborne LiDAR system is becoming a research hotspot in thefield of surveying and mapping.Airborne LiDAR system is a new type of sensor and spatial measurement system,which collect laser, global positioning system (GPS) and inertial navigation system (INS)three technology, be able to quickly and accurately obtain the information of the ground3dcoordinates, the reflection of the intensity, the reflection times. The system is graduallybeing used widely in recent years. Airborne LiDAR system can provide dense lattice dataless than1m in horizontal distance, accuracy15~20cm in vertical distance. At the sametime, the airborne LiDAR system has a strong adaptability, whose dependence to weather issmall, influence to the the shadow and the height of the sun is less susceptible. UsingLiDAR data can acquire very good aerial photo and3d road information other means canhardly access, and restore realistic3d road model very likely.This paper makes a study to use the height and intensity information provided byLiDAR data extracting three-dimensional road. Hierarchical algorithm is used to get groundpoint. Then a local layered extraction method based on road geometry characteristics isused to road extraction. For the road, in addition to the characteristics of its strength, moreimportant is the geometric feature. Roads’ geometric features include continuity, ribbon,average strength, mesh structure. However, geometric characteristics are difficult to beprecisely quantitative, so this paper mainly selected a few easier quantitative roadgeometric feature to discuss, including continuation, topological characteristics, intensitydistribution uniformity, etc. At first this paper has studied the layered algorithm’s application in LiDAR dataprocessing. Layered algorithm, makes the full consideration of high level information ofLiDAR data, and makes information in different height present independency, not onlymake the computer can deal with more huge airborne LiDAR data, but also reduce thealgorithm complexity and shorten the program running time.Then extract DTM based on height information of LiDAR data combined withhierarchical algorithm. According to height information, the original LiDAR data arelayered, and its connectivity is used to judge surface and non-surface. As a result of theground’s slow slope changing trend and mutually connected, so large connected area wouldbe produced. Analysis these large areas connected region can get DTM.This paper also studied the Fourier descriptor’s application on LiDAR data processing.Fourier descriptor is a Fourier transform coefficient in graphics boundary line, transformgraphics boundary curve signal into frequency domain to analysis. DFT (namely discreteFourier transform) linear transformation is reversible, so information won’t gain or loss inDFT transformation process.Finally, extracted road based on the strength of the LiDAR data information combinedwith hierarchical algorithm and Fourier descriptor. Local geometric feature filtering methodis used for road extraction in this paper combined with the geometrical characteristics of theroad. Due to large area in the road, the judgment of road shape will be very difficult, andthe sensitivity is low, very easily introduce non-road area into road area. Therefore, in thisarticle road extraction will be first process in local area, and finally each area be merged. Inthe local road extraction, we will use the layered extraction strategy strength, the thicknessof the layer decided by the local road smooth degree. Layer will gradually move up alongthe strength, in the moving process, the structure with the road characteristics is extracted.The experimental results show that the layered algorithm in LiDAR point cloud dataprocessing application is completely feasible, which can effectively reduce the timecomplexity and achieve higher extraction accuracy, and is very suitable for road extractionbuilt in wide range of LiDAR point cloud. |
PDF Full Text Request