Landmark Tracking Using Unrectified Omniirectional Image for an Automated Guided Vehicle

• Main Authors: Zahari, Taha, Jessnor Arif, Mat Jizat
• Other Authors: Kim, Jong-Hwan
• Format: Book Section
• Language: English
• Published: Springer 2015
• Subjects: TS Manufactures
• Online Access: http://umpir.ump.edu.my/id/eprint/9118/1/fkp-2015-zahari-Landmark%20Tracking%20Using%20Unrectified.pdf; http://umpir.ump.edu.my/id/eprint/9118/; http://dx.doi.org/10.1007/978-3-319-16841-8_60

In this paper, a study on landmark tracking using unrectified omnidirectional image for automated guided vehicle is presented. Omnidirectional image from a catadioptric camera may appear distorted against the height of an object. However, for a flat object on the floor, the distortion is negligible thus can be advantageous for on-the-ground landmark; Landmark used in this study was Code-128 standard barcode. The barcode is modified to suit the detection process where the barcode adopted cyan instead of white background and bears a red strip on top for orientation. The image processing can directly begin tracking landmarks when no distortion rectification in the image was required. We adopted a topological map approach where the automated guided vehicle moves from landmark to landmark. Experiments were conducted on a small four wheel drive, four wheel steering automated guided vehicle. The results were measured through number of successful consequent tracking of the landmark.