Mobile surveillance has been recently proposed as a replacement for traditional road-side surveillance system. One important aspect of a mobile surveillance system is to decide what the events of interest are. In this work, we adopt the concept of participatory sensing for mobile surveillance by involving a human into the loop of data collection. We introduce a virtual credit based mechanism to motivate the participants to collect data and share their bandwidth. Finally, we use a large-scale real- world vehicle trace to evaluate the feasibility of our proposed framework.