An Approach for Spatial Optimization on Positioning Surveillance Cameras

Strategic crime monitoring, surveillance, and prevision in public security is a fundamental topic in public administration to efficiently control certain types of criminal behavior that affect citizens’ integrity and quality of life. Technological advances in computer networks and video surveillance cameras allow improving monitoring coverage by installing closed circuit television network systems comprising several high-resolution panoramic cameras in public roads and streets. This is specially useful in cities with high density of population where, naturally, crime density is high in specific zones. One of the main problems is to decide where to locate the surveillance video-cameras. To address this problem, we present an optimization-based methodology that suggests where the surveillance cameras are more likely to observe more crimes. In order to implement the optimization methodology, we propose the use of the Criminal Visibility Index that evaluates how optimal the position of a video-camera is, given the historical reported geo-referenced criminal incidence. The definition of this index is fundamental to pursue the cost function that can be solved by means of optimization algorithms for non-convex problems. Our proposal focuses on the spatial aspects of the optimization problem and relies on the implementation of a greedy algorithm that has the advantage to find a near-global optimal solution for any number of surveillance cameras limited by the available computing memory.