In this paper, in order to track UAV's and maritime targets an algorithm which is based on extraction of Local Binary Patterns (LBP) near salient points is proposed. The method extracts the LBP histograms from the regions near the detected salient points. Later these histograms are concatenated and used for object representation. The tracking is achieved by comparing the main histogram with the histograms extracted inside the search region. The proposed method is tested for both UAV and sea surface targets and experimental results are presented.