How a four-member government team turned red-light cameras into speed enforcers in 3 months

Roadbuster was a government-built software tool that made existing red-light cameras useful for speeding enforcement. It solved a major systems problem: the traffic police’s old system could not process speeding images from those cameras in bulk or create official cases from them. A permanent vendor solution was being developed, but its delivery date was uncertain.

The four-person Open Government Products team built Roadbuster at an annual hackathon. It collected camera images, used optical character recognition to read vehicle registration numbers, and consolidated speed, location, and plate data. That information was formatted for upload into the case management system, allowing motorists to receive offence letters.

Roadbuster processed 120,000 speeding cases over two years. Speeding violations fell by 66 per cent at locations where it operated. After traffic police upgraded their systems, Roadbuster was retired, but its trial showed how quickly a focused tool could improve enforcement.

Red-light cameras were mainly associated with catching vehicles that crossed intersections against a red signal. However, according to the article, they also inherently had the ability to capture speeding images. The limitation was not the cameras themselves. It was the traffic police’s legacy system, which could not operationalise or process that capability.

A speeding image contained more than a single photograph. It needed associated information, including the vehicle’s registration number, speed, and location. Roadbuster gathered this material centrally and used optical character recognition to extract usable information from images. It then consolidated the data into a format compatible with the case management system.

That connection let traffic police use existing cameras for another enforcement purpose. The approach avoided waiting for the vendor’s permanent integrated solution. It also enabled officers to notify motorists with letters describing where and what speeding offence had occurred.

Roadbuster went from an urgent idea to an operating enforcement tool in three months. Darren Lau, a senior product manager at Open Government Products, and three colleagues took on the problem during OGP’s annual hackathon. Their speed mattered because traffic accidents and fatalities were rising, while the traffic police needed more cameras to catch speeding.

The team did not wait for the vendor’s permanent solution. It built a practical bridge between the red-light cameras and the police case management system. The tool gathered images, extracted information through optical character recognition, consolidated the data, and produced records that could be uploaded as speeding cases.

The team rolled out Roadbuster and improved it along the way. That rapid, experimental approach allowed the police to begin using the cameras sooner. Two years later, upgraded police systems could handle the capability directly, so Roadbuster was retired after completing its interim role.

Roadbuster handled 120,000 speeding cases during its two-year operating period. That is the clearest measure of its scale: the tool transformed a large stream of camera information into records that traffic police could act on. Before it existed, the legacy system could not process speeding cases from red-light cameras in bulk.

Its reported enforcement effect was also substantial. Speeding violations fell by 66 per cent at locations where Roadbuster was activated. OGP director Lee Hongyi said the pattern was not a coincidence because the same trend appeared at cameras turned on later. The article does not provide the underlying violation totals or a separate accident count.

Roadbuster was eventually retired after traffic police upgraded their systems. Its 120,000 cases and the reduction at activated locations show that a temporary digital tool had a measurable operational effect. It also helped officers process cases more efficiently and spend more time on other operational work.

After Roadbuster was introduced, red-light cameras could support active speeding enforcement rather than merely producing unused speeding images. The tool made it possible to create cases, notify motorists by letter, and process large volumes of evidence. Speeding violations fell by 66 per cent at locations where the system was activated.

Roadbuster automated laborious steps in the enforcement workflow. It gathered images centrally, used optical character recognition to extract vehicle registration numbers, and consolidated speed and location data. The resulting table could be uploaded into the case management system. This turned camera output into cases officers could review and process.

The Singapore Police Force said the automation improved efficiency and freed officers for other operational work. After two years, traffic police upgraded their own systems to take on the capability. Roadbuster was then retired, but the underlying enforcement function continued through the upgraded systems rather than disappearing.

Roadbuster worked because government teams and technology vendors coordinated across several systems. The project involved Open Government Products, the traffic police, the Home Team Science and Technology Agency, its vendor NCS, and Jenoptik, the red-light camera vendor. Each participant controlled part of the technical or enforcement process.

OGP built the tool that gathered images and extracted key information. The traffic police needed the resulting data for their case management system. HTX and NCS were involved in that system environment, while Jenoptik’s cameras supplied the original images and data. Jenoptik and HTX adapted their systems to ingest Roadbuster’s consolidated table format.

The teams handled friction through frequent communication, clearly scoped roles, and joint testing. Their systems had to work together reliably before cases could be processed. The collaboration also showed why the project was challenging: Roadbuster had to connect an old enforcement workflow, camera technology, vendor systems, and police operations in a short time.

A red-light violation can often be documented with a static image showing a vehicle at an intersection. A speeding case needs more evidence and context. Traffic police must know the vehicle’s speed, location, and registration number, then connect those details to a case that can be reviewed and sent to the motorist.

Roadbuster handled this extra work by collecting images in a central location and applying optical character recognition. It extracted vehicle registration numbers and other usable information, then consolidated plate numbers, speed, and location into a table. That table could be uploaded into the case management system. The system also had to handle potentially higher image volumes.

This processing made speeding enforcement possible at scale. The Singapore Police Force said Roadbuster automated time-consuming steps and helped officers work more efficiently. The later upgrade of police systems suggests the capability became part of the regular enforcement infrastructure after the temporary tool was retired.