mRMR-Tri-ConcaveHull detector for Floating Small Targets in Sea Clutter

Abstract For the feature-based detector of small targets in sea clutter, on one hand, three-dimensional convex hull-based deviates from distribution clutter vectors feature space and only combines information low-dimen-sional features. On other redundancy correla-tion between high-dimensional features are high. Consequently, we propose a for detecting floating (HDFS) this paper. First, mRMR algorithm to choose low-relatedness HDFS is proposed. For algorithm, target 8-dimensional (8-DFS), where have highest degree discrimination 3-dimen-sionalfeature (3-DFS). Second, distributionof 3-DFS concave or convex, which depends selection In most cases, concave. Using traditional hull match inevitably enlarges judgment area considerably de-creases detection probability. Due 3-DFS, new false alarm control-lable based (mRMR-Tri-ConcaveHull detector). mRMR-Tri-Concave-Hull detector, selected by mRMR, form that more suitable Con-caveHull detector. Through theexperimentalanalysisof meas-ured data, find proposed mRMR-Tri-ConcaveHull paper can significantly enhance performance compared with three-feature