An autonomous drones (or Unmanned Aerial Vehicles - UAVs) could potentially increase the chances of finding missing persons while reducing rescue time, thus increasing the overall survival chance.

We describe the development of a system that is able to detect humans in images obtained from a drone’s downwards facing camera. With such a system, the drone could automatically send the GPS coordinates to the rescue teams where it detected people.

We use the Histogram of Oriented Gradients (HOG) as main feature extractor and a Support Vector Machine (SVM) as a classifier. In this project we deal only with the computer vision aspect of the system.