Image processing techniques are successfully applied to the automated recognition of motor vehicle number-plates under a large variety of conditions. Recent progress in this field allows one to record information even when human observers fail. New video equipment is on the market and there is an increasing demand on testing them to compare their performances under experimental conditions as similar as possible to the operational ones. In the absence of international standards, this report describes a new method for assessing the performance of video equipment for automated number-plate recognition. The experimental set-up is presented for evaluating both the capability in recognizing alphanumeric characters of number-plates in standard condition and the robustness against noise sources. The method consists of laboratory tests only and avoids field trials, so that different equipment can be characterized with an improved accuracy and compared.
Experimental set-up for the characterization of automated number-plate recognizers
Measurement ; 26 , 2 ; 103-114
1999
12 Seiten, 8 Quellen
Article (Journal)
English
Generation of Handwritten Characters with Bayesian Network Based On-line Handwriting Recognizers
British Library Conference Proceedings | 2003
|