Classification of marine targets using two HRR radars

In this article, a novel method based on measuring the number of range cells in detected range profiles has been introduced according to two different aspects by two same high range resolution radars. At first, using rectangular approximation and measurement of the length and width of marine targets, calculation of the number of range cells is used in detected range profiles. On the following, according to the measurement of the angle between the two radars and target, Feature spaces are formed. Finally, classification of marine targets is formed using neural network. After completing the above steps and for testing the accuracy of the proposed algorithm, three real floatings are simulated in feko software. Then, range profiles obtained are mixed with noise and are imported to the neural network designed. Thus, the algorithm accuracy is measured according to different levels of signal to noise ratio. Accuracy of the algorithm, in certain confines, is above the 99%.