ThailandLogistics and transport2022 – ongoing

Expressway Authority of Thailand

Chalong Rat Expressway AI license-plate recognition pilot

EXAT researchers reported 97.5% plate-reading accuracy in a Chalong Rat Expressway pilot

WorkingCompany's own claim

Problem

Barrier-free toll collection required reliable vehicle identification. Staff also needed searchable vehicle records to investigate journeys and support retrospective toll collection.

Approach

EXAT developed an AI license-plate recognition system with a browser-based search application for the Chalong Rat Expressway pilot. Its researchers evaluated the system against recorded video at Lat Phrao toll plaza. The evaluation covered 987 vehicles across different times of day.

Results

97.5%license-plate reading accuracy in EXAT researchers' pilot evaluationCompany's own claim[1]

93%vehicle-color identification accuracy in the results table; the abstract and conclusion reported 96%Company's own claim[1]

93%vehicle-brand identification accuracy in the abstract and conclusion; the results table reported 96%Company's own claim[1]

Takeaway

Staff could correct vehicle records before issuing reports. Color and brand accuracy figures differed between the results tables and the abstract, so the lower figure is retained for each.

Sources

Automatic License Plate Recognition and Search Engine with AI Technology: A Case Study of Chalong Rat Expressway
27th National Convention on Civil EngineeringAug 2022Research paper or report

Source summaries

[1]27th National Convention on Civil EngineeringAug 2022

EXAT staff evaluated 987 vehicles against recorded video and reported 97.5% plate-reading accuracy. The abstract and conclusion reported 96% color accuracy and 93% brand accuracy, while the results tables reported 93% and 96%, respectively.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026