Study of inland river maritime radar image Gabor wavelet edge recognition based on direction adaptive

Abstract

To realize automatic extracting edge of bank and bridge from inland marine time radar image, the first step is to get the detection region and realize superposition with the electronic chart display system automatically. Image processing of the radar image should be done firstly, to get radar coastline and the edge of the bridge and location information. The basic way of image processing in the target extraction based on two extraction methods, morphology and color characteristic. Color feature is mainly suitable for color image target. Especially for clear goals, according to color component achieve target extraction. Morphological features for feature extraction. The main method is to realize the target of morphology through the preset or template matching target to get extraction. Maritime radar image without templates, the conventional extraction method is mainly been used in morphology to extract the target. Generally implemented to extract the linear through Hough transform, for the non linear characteristics of target used fractal way. Fractal needs through a large amount of data access to lines of latitude interval. Interval target extraction based on latitude, with large amount of calculation. In this paper, through combination of Hough transform to get the tendency of shoreline bridge information, we use the Gabor wavelet certain tendency feature of shoreline and Bridges to extract the target, in a certain extent, reduce the amount of calculation, and achieve the target of extraction.