2017 · 17 citations · 10 references
ComputingEarth SciencesGeoinformaticsComputer VisionDigital ImagingRemote SensingDigital Image ProcessingEdge DetectionImage AnalysisImage SegmentationRemote Sensing Image AnalysisTexture AnalysisCanny Edge DetectorDigital Image EnhancementEdge FeaturesImage Feature ExtractionNearest-neighbor ClassificationRemote Sensing Image ClassificationShoreline MappingTexture Feature Extraction
This paper investigates the potential of using a Gamma Correction-based image enhancement and Canny edge detector to extract the shoreline position from coastal video images. Within the framework of the shoreline extraction system, Gamma correction has updated the contrast of coastal imagery to segmented and classified efficiently in land, sea and sky areas. In this study, the value of the Gamma correction factor was determined to be 1.3. At this value, the validation of the cluster value in the Self Organising Map (SOM) segmentation has shown an increasing perform of cluster, from the reasonable cluster to the strong cluster. Furthermore, the mechanism of classification of land, sea and sky objects in coastal imagery is done by extracting the texture features on each cluster. Then, feature information is stored for the classification training process using K-Nearest Neighbor (K-NN). The shoreline extraction procedure is based on the identification of the land class in the K-NN testing process. This study has implemented a binary image transformation method combined with a Canny edge detector to define pixel boundaries of land and non-land areas. The experimental results have shown that the proposed method capable of producing a continuous shoreline in accordance with the texture of the coastal path in coastal imagery.
10