A Novel Approach for Recognition and Identification of Low-Level Flight Military Aircraft using Naive Bayes Classifier and Information Fusion

Abstract

A problem that has been faced by the Radar is if the aircraft flies at low level or near to the surface so its coming in the aerial-surveillance airspace cannot be detected and endangers the air sovereignty. The aircraft can be recognized and identified by carrying out a technique called Visual Aircraft Recognition (VACR) using a binocular. This technique requires military personnel that has capability carrying out the air surveillance from the ground. Surveillance is a time-consuming and tiring task so it can cause fatigue and impact to the results of the recognition and identification. To cope with this problem, we have designed and implemented a novel recognition and identification method using the combination of Naive Bayes Classifier (NBC) and information fusion. By using a dataset that consists of 45 military aircrafts, 35 civilian aircrafts, 40 military helicopters, and 35 civilian helicopters with 80:20 dataset distribution for the training scheme and the validation one, we obtained the recognition accuracy of 87.1%. We also found that the recognition and identification process can be speeded up 1.2 seconds when using information fusion.