On Extracting Road Junctions From High-resolution SAR Images

Road junctions play a very important role in application of remote sensing,which is of great significance to studies on extraction.Firstly,road junctions,providing initial seed point and guide for semi-automatic extraction and establishment of road network,are an important part of road network.Secondly,road junction of being a basic element of road network can show position,linkage and orientation information of road.Finally,as ground control point,road junctions can effectively help build the model of the correct geometry in key steps.This study shows basic theoretical research and treatment scheme of extraction of road junction on high-solution SAR image,carrying out researches which focus on pretreatment SAR image(remove speckle noise),crude extraction and model confirmation of road junction on high-solution SAR image.In this paper,the main innovation points are:(1)Speckle noise brings the difficulty of extraction of road junctions.Given that speckles cause problems with the interpretation of SAR images,speckles must be reduced to obtain high-quality images in SAR follow-up processing.Edge preservation is also a crucial aspect to consider.However,these tasks are inconsistent.Therefore,an efficient algorithm is needed to solve this problem.This method is categorized as anisotropic diffusion and is called Euclidean distance anisotropic diffusion.Euclidean distance anisotropic diffusion is a modified SRAD method based on Euclidean distance.It combines the SRAD model with Euclidean distance to effectively reduce speckles and preserve edge.The experimental result showed that the method can reduce speckle noise in areas with high and low concentrations of noise speckles.(2)A method on directly detection of road junctions from SAR images is introduced to obtain road junction candidates.Synthetically uses morphological algorithm,binary operation and connected area calculation for removing false road junctions and detecting road junctions on high-solution SAR image,which provides candidate regions and reduce work load for next confirmation of road junctions.The experiment shows that this method can obtain road junction candidates without extracting road line characteristics for confirming the model of road junctions.(3)Targeting road junctions extracting from high resolution SAR images susceptible to various kinds of interference and contrast not obvious between roads and surroundings,a method for extracting road junctions by triangular model.This method can directly extract road junctions on high-solution SAR image without extraction of road before.The experiment shows this designed triangular model can better extract a variety of common road model including “L” model of two forks,“T” model of three forks,“Y” model of three forks,“+” model of four forks and “X” model of four forks.