Abstract

The usage of UAV-based monitoring and surveillance has been rapidly increasing in the past few years. The UAV surveillance is quite efficient compared to CCTV cameras or regular patrol by security personnel as the flight reduces the time of the patrol and at the same time covers more area thus increasing not only the number of patrols but also increasing the radius of the patrol. Over the years UAV surveillance has been adopted by many nations but it these mainly used only for military purposes or in disaster spots if we use it in urban cities, we can cover the back alleys that are not under the CCTV coverage and the areas that are hard use CCTV cameras and patrol or even in private properties like mansions or companies that has huge gardens, etc., which greatly reduce many criminal activities also prevents crime such as burglaries, trespassing private properties, etc., it does not only help for police but also can be used in private properties. In this model, we propose a model system for urban UAV surveillance which will be able to patrol the designated area in the designated path and also detects person and weapons such as knife or gun.

Keywords:Deep learning; UAV; Automated surveillance; MAVLink; Open source