BusEdge: Efficient Live Video Analytics for Transit Buses via Edge Computing

Many vehicles like transit buses are now routinely fitted with cameras. These live visual data are invaluable to achieve real-time traffic monitoring, but it is intractable to handle such a gigantic amount of data either locally or in the cloud due to computation or bandwidth limitations. In this work we propose a system that uses edge computing to achieve efficient live video analytics on transit buses. We call our system BusEdge. It uses an in-vehicle computer to preprocess the bus data locally and then transmits only the distilled data to the nearest cloudlet for further analysis. Our system provides an easily extensible and scalable platform for related applications to make use of the live bus data. A typical application, AutoDetectron, is developed upon the BusEdge platform to execute ad hoc search queries for a given object using the video stream from the bus. It integrates labeling, recursive learning and automatic model management to boost the development of a general object detection pipeline on BusEdge. To evaluate our proposed system and application, we deploy the system on a running bus and conduct extensive experiments. Experimental results demonstrate the efficiency and scalability of BusEdge and the object detection performance of Auto-Detectron. Code is available at https://github.com/CanboYe/BusEdge.