Key Technologies Of Extraction And Importance Classification Of Broken End Road Based On Remote Sensing

The problem of broken end road affects people’s daily life and hinders traffic development.But as a special road form,broken end road in specific scene plays an important role in clue guidance for the identification of important targets.Road extraction and target recognition using remote sensing images has always been a research hotspot and difficulty in the field of remote sensing applications.However,there is little research on the detection of broken end road based on remote sensing images and the application of broken end road in target recognition.Based on this,this paper proposes an algorithm for broken end road detection based on remote sensing images.Considering the special location characteristics of the broken end road,this paper puts forward the importance classification principle of broken end road based on the surface attribute of the related area of broken end road,and verified the clue guidance role of broken end road in target recognition application.The main research results of this paper are as follows:(1)Taking GF-2 multispectral image as experimental data and road buffer as limiting condition,the road edge is extracted by using double angle threshold phase grouping algorithm and local quadratic edge detection method..Firstly,the edge line groups in the image are extracted by double-angle threshold phase grouping algorithm,and the geometric characteristics,continuity and parallelism of road edge line groups are used to screen the edge line groups.Then,the location of road interruption is determined according to the position relationship between the line groups,and the local secondary edge detection of the road interruption area is carried out by using the remote sensing images of different time phases registered,so as to realize the pseudo-interruption connection of the road.Experimental results show that the algorithm proposed in this paper can accurately and completely extract the road edge while screening the road edge to a large extent..(2)An algorithm of broken road detection based on the distance between breakpoints and the identification of true and false breaks is proposed after the edge line group is extracted.Firstly,the minimum distance between the end points and the end points of the same line group and the minimum distance between the end points of the opposite line group were calculated,and the distance threshold was set to preliminarily determine the location of the broken point.Then,the road buffer mask is used to screen out the false detection broken end road points in the road to realize the broken end road detection.Experiments show that the true broken ends in the remote sensing images of simple road network and complex road network are correctly detected,and there is no missing detection.(3)The multi dimensional feature expression of the surface of the associated area of the broken end road and the importance classification of the broken end roads are studied.Firstly,the extension direction of the guillotine is calculated according to the terminal coordinates of the edge line group of the road where the guillotine is located,and then the associated region of the guillotine is determined.Then,the remote sensing index in the associated region is calculated,and multidimensional surface feature information is obtained by combining DEM data and vector geological data.A classification principle of importance of broken roads is proposed based on surface feature of the associated region of broken end roads.(4)Taking three special target areas such as port,tunnel and satellite launch base as examples,the feasibility of this technical route is verified through demonstration and application,which is based on remote sensing image to find dead road,importance classification based on surface attributes of the area associated with broken end road,and finally finding important target area with broken end road as a clue.In addition to image preprocessing,other experimental steps are realized by C++ and Python programming language.Except for some threshold values which need to be manually set and adjusted according to the actual needs,the automatic detection of broken roads based on remote sensing image,and the application function of land surface feature statistics and importance classification in the area associated with broken end road.