Multi Feature Descriptor Based Ship Wake Detection Using AdaBoost-Weighted Extreme Learning Machine

In this paper, a new ada boost-weighted extreme learning machine technique is proposed to detect ship wakes from synthetic aperture radar (SAR) images. To extract the features from SAR images a multi feature technique is proposed, which effectively extracts the pixel information. The extracted unique feature values are then imported to the adaboost weighted Extreme Learning Machine (ELM) for the classification of ship wakes. It is observed that the proposed technique is able to solve imbalanced data problems effectively and provides improved performance over other existing techniques.