Improvement of robot arm control using computer vision approach

Uroš Tomšič (2010) Improvement of robot arm control using computer vision approach. EngD thesis.

PDF Download (1196Kb)

Abstract

The diploma work presents a control system with added computer vision system for more accurate robot hand positioning. In process automation computer equipment adapted for industrial environment is usually used. Such adaptability brings about very high costs. Since these systems are usually not compatible with consumer goods systems, it is necessary to use sensor systems adapted for operation in industry also in friendlier environments. The diploma work demonstrates that moderate efforts enable significant improvement of control systems with elements of consumer goods and the use of freely accessible software libraries. In the aspect of the approach mentioned, we tried to improve the control of the robot hand. We built our own control system into which the elements of computer vision are included. The main part of the diploma work describing the development of the entire system is divided into two parts. In the first part the robot hand control system, the programme structure, the automatic control cycle with or without the use of an image analysis and the setting of the system into the initial state are presented. The second part is focused on the development of the SCADA system in the C# programme language in the Visual Studio 2008 environment. The system includes the basic visualisation, the robot hand control in the two automatic and manual operation regimes, the alarm/notification interface, the setting of relevant data, saving of data in the data base, processing of certain data and their graphic/table review. The system also includes a web camera interface that enables a user to supervise the robot hand operation. It is used in an analysis and processing during the fine positioning of the robot hand. It also includes the incorporation of a live image into the system including the image analysis, setting of the control points and description of data archiving and processing.

Actions (login required)