A Research Of Mountainous Highway Damage Monitoring And Assessment Based On Remote Sensing

Road network traffic system is extremely easy destructed by the natural disaster. Road disaster not only brings a huge economic loss to disaster areas, but also some more serious loss result from rescue delay. After significant disaster occured, the multisource data is collected by different remote sensing platforms, it will be the foundation of disaster detection. Based on these multisource remote sensing data, several road damage information can be acquired, such as road damaged scope, road damaged degree and road damaged types information etc.. These kind of damage information will help for post-disaster road emergency repair, the emergency transportation organization and decision-making of rescue material schedule. The remote sensing survey is the product of information technology develops highly, it is very suitable in the fast and synthesis monitor of dynamic change of the wide range environment. Now, the technology has been succeeded used for some disaster monitor, such as earthquake, volcano, flood, landslide, snow and ice, forest-fire, hurricane and so on. But, it is difficult that remote sensing data is used to identify roadbed subsidence, roadbed collapse, the bridge break from the wide range road network and several kinds of block factors of road sections. In the world wide, the research has just been begin. This paper researched deeply three key technologies of significant road disaster monitor, including road scope detection, multisource data match and road damage information detection. The concrete research work may be divided into the following several parts:(1) Some remote sensing data sources can be used to evaluate road damage situation under disaster emergency condition. In the light of the different data combination, several feasible road damage detection schemes were deduced, and damage information content and precision of every kind of scheme were analyzed.(2) This paper analyzed road's ideal and non-ideal model characteristics in some high spatial resolution remote sensing images. Based on the model characteristics, a method that is combine arithmetic of MonteCalo particle filtering and arithmetic of improved Snake edge location is put forward. But when the road edge blurry, the detection result of the method is not good. In order to overcome the deficiency, a method of double edge tracking based on the road geometry characteristic restraint was proposed. Two restraint conditions were introduced into the method: The road can not present a extreme turn (constraint 1); Road's width can not have the sudden change (constraint 2). This method can solve the overwhelming majority road scope detection problems, for few road sections which violates the constraint conditions, a feasible detection plan was also proposed.(3) This paper is use for the least squares method and curve characteristics method match LiDAR data and optical image of mountainous area, and the later one is more reliable that is proved by experiment comparison and analysis. The curve match method is described as follows:firstly, it take use of w-d model describe a curve; And then it is use for the method of w-d model matching to find homonymous curves, Next it build price matrix by the similar relations of homonymous curve nodes. lastly, it confirm the corresponding relationships of homonymous curve nodes by the dynamic planning method.(4) This paper used the LiDAR data to analyze the road damage information. Firstly, a conclusion that the Li's morphology filtering algorithm is the most suitable for dealing with a mountainous area LiDAR data is obtained by analyze. And, a improved method based on Li's was proposed which can retain the bridge points during filter disturbance information. Next, a DSM is structured by point clouds. The LBP and VAR texture measures were used for analyzing DSM, whereupon the smooth and rough terrain can be distinct shown. And then the object-oriented multi-criterion division technology and the fuzzy C average value classify algorithm is taken use for differentiating the damage and the complete region. At last, according to volume change of all differential vexel in damage region, the road section damage degree (Landslide quantity or collapse quantity) can be precisely calculated, in addition according to several kind of damage models, the road section damage type (the burying damage, the bridge collapsing damage, the roadbed collapse damage) can be obtained.