Comprehensive Approach to the Evaluation of Off-Line License Plate Recognition Data

The aim of the article is to present a comprehensive procedure for processing and evaluating directional data from vehicle license plates, focusing on the specific challenges of areas that are sparsely or not at all equipped with permanently located standard license plate recognition systems. This remains a current issue, especially in smaller towns; it leads to the implementation of short-term directional traffic surveys, often using inexpensive measurement devices, in order to obtain directional traffic data. In this research, a procedure for evaluating license plate recognition data is proposed with a primary focus on its simple adaptability and automation for any subsequent use. The data sources considered are primarily the above-mentioned traffic surveys; however, the proposed evaluation procedure is theoretically transferable to any off-line data obtained from license plate recognition systems. Identifying potential inaccuracies in the data is also an integral part of the evaluation process. The design of the proposed procedure follows the Checkland soft systems methodology and the functionality of the resulting procedure was validated through a case study of a directional survey in Prague. The proposed procedure contributes to greater accuracy of the conclusions drawn from evaluated traffic engineering parameters under non-ideal, but common conditions of smaller cities, not only in the Czech Republic.