Automatic Configuration of Video-Surveillance Applications: a Model-Driven Experience

Abstract

The number of video-surveillance applications has increased in the last years due to the increasing security requirements, especially in public buildings. In that sense, new types of applications have appeared with new services. As an example, digital applications allow the possibility of remotely controlling the video through internet. However, most of these applications are implemented ad-hoc. They are very tied to the programming level reducing thus their reusability, configurability and maintainability. In this setting, we propose a Model-Driven approach to develop video-surveillance applications. This approach is based on the utilization of a Domain Specific Language (DSL) that allows the design of these applications at the modeling level. The utilization of models enhance the reusability of the services provided by the systems but also aims at managing their variability making them highly configurable. The models built based on this DSL are then used by model transformations that are able to automatically generate the configuration of the applications so that they may be dynamically reconfigured. Moreover, the approach provides different versions of the system, namely a standalone application, a web application or an application for mobile devices. All these applications allow the remote control of the video installation and the integration of external services such as SMS alerts, vehicles' plate recognition or smart home control.