Sea object detection using shape and hybrid color texture classification

Abstract

Sea target detection from remote sensing imagery is very important, with a wide array of applications in areas such as fishery management, vessel traffic services, and naval warfare. This paper focuses on the issue of ship detection from spaceborne optical images (SDSOI). Although advantages of synthetic aperture radar (SAR) result in that most of current ship detection approaches are based on SAR images. But disadvantages of SAR still exist. Such as the limited number of SAR sensors, the relatively long revisit cycle, and the relatively lower resolution. To overcome these disadvantages a new classification algorithm using color and texture is introduced for Ship detection. Color information is computationally cheap to learn and process. However in many cases, color alone does not provide enough information for classification. Texture information also can improve classification performance. This algorithm uses both color and texture features. In this approach for the construction of a hybrid color-texture space we are using mutual information and three aspects: 1) Classifies ship candidates 2) The relevant classes are automatically built by the samples' Appearances and 3) Shape and Texture features. Experimental results of SDSOI on a large image set captured by optical sensors from multiple satellites show that our approach is effective in distinguishing between ships and non ships, and obtains a satisfactory ship detection performance.. Feature extraction is done by the co-occurrence matrix with SVM (Support Vectors Machine) as a classifier. Therefore this algorithm may attain a very good classification rate. © 2011 Springer-Verlag Berlin Heidelberg.