Recognizing, Distinguishing and Tracking Enemy Army by Missile’s RGB-D Sensors, to Decrease Civilian’s Casualty, in Missile War

As technology grows, it change everything and war is not a difference. Using human resource in direct battle has cost, so for decreasing human death, far distance battles takes advantage, and in this field missiles have priority. All the controllable missiles have GPS and a visual camera in the rocket tip (warhead), which uses image processing techniques to track and hit. By increasing such a war in the world, especially in Middle East, and with the aim of reducing civilian’s mass destruction, this paper proposed a new method. This paper proposes a method to use depth or range cameras beside of color ones to increase the accuracy and better functionality on the day and night time, for preventing civilian’s death. Depth sensor which is attached to the tip of missile, must distinguish civilians from enemy clothing and texture patterns and change the way if target was civilians or just terminate the functionality of the missile with help of color sensor. For testing the system, Kinect V.2 is employed as color and range sensor and it will work with any other infrared sensors based on proposed theory. As it is clear infrared cameras could calculate distance in day or night time which is so crucial in missile war. Features like Speeded-Up Robust Features (SURF) and Local Phase Quantization (LPQ) is employed along with K Nearest Neighbor (KNN) classifier for fast detection (because missiles are so fast, times matter). Also Kalman tracking is using for tracking moving enemies. For validating, a database is collected from internet plus self-made data recorded using Kinect V.2 sensor. Achieved results shows fast recognition and action in our test. This method could be used in real environment and with stronger devices, but the base is same.

Keywords:

Missile , Civilian , Depth sensor , Kinect V.2 , Kalman filter tracking