Target classification in perimeter protection with a micro-Doppler radar

In security surveillance at the perimeter of critical infrastructure, such as airports and power plants, approaching objects have to be detected and classified. Especially important is to distinguish between humans, animals and vehicles. In this paper, micro-Doppler data (from movement of internal parts of the target) have been collected with a small radar of a low-complexity and cost-effective type. From time-velocity diagrams of the data, some physical features have been extracted and used in a support vector machine classifier to distinguish between the classes "human", "animal" and "man-made object". Both the type of radar and the classes are suitable for perimeter protection. The classification result are rather good, 77% correct classification. Particularly interesting is the surprisingly good ability to distinguish between humans and animals. This also indicates that we can choose to have limitations in the radar and still solve the classification task.