An Embedded System Design for Wildlife Detection and Identification

When there is a need to monitor wildlife animals in the vast wilderness, power consumption is a challenging aspect in systems designed for such purposes. In this work, a low-power embedded-camera system for wildlife detection and identification is implemented. The design objectives of the system are to reduce the power consumption and system load of operation in an embedded environment through a prior movement detection strategy. An asynchronous multi-threaded concurrency mechanism is employed to increase the Frame Per Second (FPS). This system is beneficial for many applications in various wilderness areas, such as traffic alarms when large wild animals are crossing a highway, public warning systems of potential threats in urban areas indicating the presence of dangerous animals, and wildlife migrations monitoring by biologists.