Tracking of Platforms above Sea-level in Visual Band Videos using Covariance based Tracker and Feature Selection

Abstract

In this study, a feature based tracker is developed in order to track platforms above sea-level in visual band videos. The covariance descriptor, known as an efficient method in object detection and classification problems, is used in the feature extraction phase of the tracker. The features or feature sets, extracted from the images in the database that is constructed for this work, are compared by using Support Vector Machine and by using the features or feature sets providing the best target/background separation, in the covariance based tracker successful tracking of platforms above sea-level is provided. Experimental results show that by adding or extracting appropriate features to covariance based descriptors, the proposed method can track different platforms successfully