A Novel Algorithm for Moving Target Detection

Abstract

The approach for this project is simple. For the detection of moving, we had employed the method of image subtraction and absolute frame differencing of consecutive frames coming from the static camera for making a robust approach for signature locking of the detected target. Through the image subtraction technique, we extract the coordinates of the moving target from the image and to a designed mechanical system that can operate the gun. A red signature is accurately a red object in the real-time video. The reason behind a red object is that a thermal camera detects the object on the base of the temperature gradient and the thermal camera detects an aircraft in the form of a red object. The thermal camera is so expensive that why we chose the visual camera and selected red object as a signature so that this algorithm is also capable to detect and track a thermal target from the thermal camera. And after detecting the coordinate of the required target sent to our gun barrel to shoot that target for future enhancement. There exists a lot of applications for our project, especially from a security point of view. This thesis explained how to detect and track red targets in a live video. There are distinctive ways to deal with recognizing the red/warm shading in each edge. One of these prevalent methodologies incorporates the change of the entire RGB outline into comparing the HSV (Hue-Saturation-Value) plane pursued by extraction of pixel esteems just for red. One can pick a range that mirrors every single existing shade of red to distinguish and separate all hues in the frame.