A SYSTEM TO DETECT ANOMALY IN LIVE FEED OF AUTONOMOUS DRONE SURVEILLANCE USING COMPUTER VISION APPROACH

Abstract

Now a days, every city in the world has installed CCTV cameras for the monitoring the crowd and for public safety. But as these cameras are increasing, it has become difficult to manually monitor the feed which is received from these cameras. Also CCTV cameras are stationary, which is making them capturing the footage from particular angle while monitoring a place. These problems can be overcome by using the autonomous drone technology. By virtue of recent industrial developments, drones are increasingly employed in various domains. They can also be employed for surveillance as moving cameras. But again the problem of autonomous drone surveillance is that, it can only fly around without human interaction and transmit the live feed. It cannot interpret what is in that live feed. The solution to this problem is Computer Vision application through Deep Learning approach. Deep learning is a trend field of Computer Science which is helping to solve the problem like to autonomous car, detection of cancer cell and many other applications where it can possibly replace human eyes to see and interpret the data. In this paper, we are proposing a system design which will make use of the deep learning to detect the anomaly in the live feed received from an autonomous drone. The suggested anomaly detector system makes use of a deep neural network composed of a convolution neural network and a recurrent neural network, trained using supervised learning. Although we have not implemented a system but our main aim is to promote research efforts to resolve the impenetrability of anomaly detection in live videos.