Development of a general-purpose vision system for robots and vehicles

Abstract

As technology advances over time and with the rise of artificial intelligence, there is a growing demand for robots and autonomous vehicles to improve the lives of humans. A vision system is crucial for these machines to be able to interact with the human world, allowing them to ‘see’ and perform meaningful tasks such as avoiding obstacles. This report seeks to address that issue and in the process develop a general-purpose vision system for robots and vehicles. The report begins by giving an introduction to the project through describing the background and needs of using vision systems. Next, a review of knowledge on modern vision systems is presented. Following that, blueprints of hardware and software components making up a general-purpose vision system are proposed. Using the various hardware and software as stated in the blueprints, a prototype vision system is designed. The system has four main functions: colour detection, shape detection, object detection and facial recognition. These functions can be customised to different user requirements, such that the vision system becomes general-purpose. Finally, the performance of the developed prototype is tested, and its advantages and limitations are evaluated in detail.