Javascript must be enabled for the correct page display

Developing an image recognition system

R. Deluster, (2015) Developing an image recognition system. Bachelor's Thesis, Industrial Engineering and Management.

[img] Text
Bachelor_Industrial_engineerin_1.pdf - Published Version
Restricted to RUG campus only

Download (2MB)
[img] Text
Toestemming1.pdf - Other
Restricted to Repository staff only

Download (41kB)

Abstract

The research to robotic fish at the Rijksuniversiteit Groningen has been put on the back burner since 2011. The group of Discrete Technology & Production Automation likes to breathe new life into the development of the robotic fish starting with this Integration Project. The people that will get involved exist mainly out of students Industrial Engineering & Management and PhD candidates. The robotic fish swim in a pool that is situated in the DTPA-lab of the Nijenborgh 4 at the Zernike campus. The fish communicate through software with the researchers. The current soft- ware is written in C++ which is a hard coding language to understand and to adjust for laymen. Therefore, the assignment is to write a new software in a MATLAB environment. This software will be the foundation for the future research. The scope of the new software is narrowed to developing a control, identification and track&trace function. The control function enables the researcher to steer the fish. The fish are to be recog- nised using a camera. Each fish gets two colored squares on its back for identification. The fish’ coordinates are extracted from the image using color recognition and BLOB analysis. When the identification of a fish fails, the track and trace function ensures that a location is estimated. The software is demonstrated in a demo and a manual is developed for future users. The software can be used as a foundation for future research and development. The identification function can be improved by implementing a benchmark that automatically detects the illumination circumstances. The track&trace function can be made more accurate when a Kalman filter is made.

Item Type: Thesis (Bachelor's Thesis)
Degree programme: Industrial Engineering and Management
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 15 Feb 2018 08:05
Last Modified: 15 Feb 2018 08:05
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/12863

Actions (login required)

View Item View Item