Mobile automated video surveillance systems deployed in mobile environments are used to monitor and analyze numerous situations and take necessary actions in real time. These systems include various mobile nodes equipped with numerous sensors and involve application of sophisticated image and video processing algorithms. The execution of these algorithms requires a vast amount of computing and storage resources. To address the issue a traditional approach is to send collected data to an application on a cloud accessible through an infrastructure-based system such as cellular network. This approach has several issues such as high transmission energy consumption and communication latency. In addition, this approach cannot be used in situations where pre-existing communication infrastructure is not available. This paper focuses on a recent approach in which multiple mobile devices interconnected through a mobile ad hoc network are combined to create a virtual supercomputing node called a mobile ad hoc cloud which is then used to support execution of automated video surveillance application. In order to fulfil real-time requirements associated with automated video surveillance application, a task allocation scheme has been proposed. Compared to existing schemes, proposed scheme focuses on deadline-oriented tasks and energy efficiency.