Research On Digitalization Of Airport Pavement Based On 3D Scanner

With the rapid development of aviation industry,the aviation safety has attached great importance to the governments and the populace.As the most important infrastructure of airport,the performance of airport pavement has great impact on the taking off and landing of aircraft.It is important for airport manager to evaluate pavement indexes scientifically and master the pavement properties correctly.This paper mainly studies on the digital system of airport pavement based on the 3D point clouds data which is collected by 3D scanner.The purpose is to provide a high-performance digital system to airport manager.This system can provide important technical support for airport manager during the pavement maintenance and decision-making.There are two important functions of the 3D digital system on airport pavement.On the one hand,it should reconstruct 3D scene of pavement.On the other hand,it can evaluate the commonly used indexes of pavement.Therefore,this paper is almost organized around 3D point clouds registration and the evaluation of flatness,gradient,damage and foreign body on airport pavement.The ICP algorithm costs a lot while searching for the corresponding points in the process of point clouds registration.And if there is no good initial value,the algorithm might not converge.So it is not applicable to the digital system of airport pavement.Considering that the point clouds of airport pavement have simple structure and texture,this paper firstly proposes a point clouds registration algorithm based on coded lights.During the experiment,we add additional time-coded lights which are controlled by hardware circuit to the scene.Then detect and identify the 2D coordinates of these lights in image sequences which are collected by scanner from different position.Finally,finding out the corresponding 3D coordinates of these lights in point clouds can achieve the point clouds registration.Then,this paper puts forward another point clouds registration algorithm,which based on the characteristic that the data of airport pavement are always surfaces with low curvatures.The low curvature surfaces in the overlapping area of point clouds are extracted by the regional growing.Then these surfaces are projected to 2D imageorthographically according to the mean normal.The SIFT is used to detect and find matching relationship between features of 2D images.Finally register the point clouds.In the digital software system,combined with differential GPS data,the algorithm completes the registration of actual airport pavement data.Finally,we built 3D digital software system of airport pavement based on the point clouds collected by 3D scanner.This part firstly analyzes the requirements of digital system.And we introduce the framework and the interface of software system.Then we descript the main modules of the software in detail.The key modules of the software system are: the preprocessing module,the evaluation module of flatness,gradient,damage and foreign body,and the results outputting module.We analyze the design ideas and related algorithms of each module.And then give out the final implementation of each module in the software system.