2009 · 12 citations · 3 references
ComputingGeoinformaticsComputer VisionRemote SensingUrban MorphologyEdge DetectionImage SegmentationRemote Sensing Image AnalysisWatershed DelineationHigh-resolution ImagingMorphological Image SegmentationRoad ExtractionRobustnessWatershed Segmentation
Nowadays, Google Earth is very popular in the world. Lots of people are interested in this software. The purpose of this paper is to explore the method of road extraction in high resolution urban images from Google Earth. There are several basic characteristics of roads, if more characteristics are used for the road extraction, good results can be achieved. In this paper, watershed dual-thresholds algorithm is used to obtain the initial outline of roads, multi-weighted method is used to get the exact edge of roads, morphology is used to eliminate noise and thin lines, and shape index is used to delete non-road areas at last. At the end of the paper, two different types of road of experiments are to be done. And in accordance with the outcome of experiments, the algorithm proposed in this paper has a high precision and a good robustness in the various types of urban road extraction.
3