Landmark navigation in low illumination using omnidirectional camera

Landmark based vision navigation for mobile robot is critically dependent on successful recognition of landmarks. Landmarks, artificial or natural, are subjected to sufficient illumination in order for a successful recognition. Sufficient illumination is even more critical when the mobile robot is used for indoor. In this paper, experiments were conducted to recognize artificial landmarks using omnidirectional vision under low illumination. The objective of this paper is to demonstrate that landmark navigation in low illumination can be conducted without illumination invariance step and without images distortion correction. This landmark recognition performance thus demonstrates the robustness of landmarks especially under low light condition. The landmarks used were standard (ISO15417) Code-128 barcodes. The barcodes are placed besides turning machine and the illuminance is measured by a luxmeter on each barcode.