Vision Software for Humanoid Robot Soccer

Exhibitor: Anthony Peters

Supervisor: Gordon Wyeth

Research Group: Complex and Intelligent Systems

Industry Sector: Scientific / Research Services

The previous work on GuRoo’s vision system has seen the independent development of the necessary hardware and software subsystems. The hardware element for this vision system comprises of an image capture system and an image processing system. The software element consists of refined algorithms to perform the image processing functions, which achieve the goal of object detection and analysis.

Independently, both the hardware and software subsystems have been implemented and their functionality verified experimentally. However, no attempt has been made to integrate these systems with GuRoo, and demonstrate their functionality in conjunction with other GuRoo behaviours.

The scope of this project involved the collaboration of the existing subsystems and verifying their ability to function cooperatively, along with the integration of the vision system with GuRoo.

Before this integration was attempted, the various debugging applications used in conjunction with the vision system, which were initially written for a Linux execution environment, needed to be ported to the Microsoft WIN32 environment.

In addition to the redevelopment of the debugging applications, various add-ons were developed for these applications which allow the visual analysis and verification of the image processing functions. These add-ons consisted of the definition of a communication protocol for use between the image processing system and the PC.

Thesis Document (PDF)