Applied Mechanics and Materials · 2012 · 10 citations · 9 references
ComputingComputer ScienceComputer VisionMachine LearningRemote SensingFeature DetectionObject DetectionPattern RecognitionRemote Sensing Image AnalysisRemote Sensing ImagerySupport Vector MachinesFeature MatchingHigh-resolution Satellite Image AnalysisK-means ClusteringScale-invariant Feature DetectionScale-invariant Feature TransformScale-invariant Feature Transform (Sift)Spatial Pyramid MatchingSupport Vector Machine Classification
In this paper we propose an automatic ship detection method in High Resolution optical satellite images based on neighbor context information. First, a pre-detection of targets gives us candidates. For each candidate, we choose an extended region called candidate with neighborhood which comprises candidate and its neighbor area. Second, the patches of candidate with neighborhood are got by a regular grid, and their SIFT(Scale Invariant Feature Transform) features are extracted. Then the SIFT features of training images are clustered with the K-means algorithm to form a codebook of the patches. We quantize the patches of candidate with neighborhood according to this codebook and get the visual word representation. Finally by applying spatial pyramid matching, the candidates are classified with SVM (support vector machine). Experiment results are given for a set of images show that our method has got predominant performance.
9