Design of a vision enabled wireless sensor node

Abstract

This thesis describes a novel system architecture and implementation of a wireless visual sensor node. The proposed design of the node can be used to extract traffic information based on the visual description of road. In this research, the real-time performances and the capability to perform at low power consumption meanwhile obtaining accurate results were considered as the essential factors since a large amount of data need to be processed in an embedded level vision system. At first, a suitable vision algorithm is proposed to harvest the traffic condition on road. The functionality for each section of the algorithm has been performed by using carefully selected available vision techniques and image processing algorithms. The vehicle extraction from the current frame of view and the tracking process of the vehicle are identified as the most important functions in the algorithm. The vehicle extraction from the current frame is carried out by the ViBe algorithm with some modifications in order to acquire promising real time performances and the tracking process is carried out by a light weight but an accurate enough particle filtering technique. Moreover, the complete system is implemented in the FPSoC hardware system as a hardware and software co-design by considering advantages that can be obtained from different aspects. The performances of the system have been evaluated from many aspects for different standard data available from other research works. The conclusions and suggestions for further development have been presented at the end of this thesis.