Recognition of colorful objects in variant backgrounds andillumination conditions

摘要

This paper presents a fuzzy based method for recognizing colorobjects in a complex background under varying illumination. Fuzzy rulesare generated using a fuzzy associative memory training method to copewith chromatic distortion. The color model used is thehue-saturation-value color model. In this paper, the proposed method isapplied to stop sign recognition in real-world scenes that may haveincorrect video camera focus and/or other color distortions. The resultsof this research can help the GPSVanTM system generate mapsautomatically