MelSafe
5 Minutes
MelSafe: A New Era of Safety and Convenience in the City of Melville
Automatic Number Plate Recognition (ANPR), or Automatic License Plate Recognition (ALPR) in North America, has quietly become one of the most transformative surveillance and compliance tools of the last half century. It’s the technology behind the cameras that spot speeding cars, enforce tolls, flag stolen vehicles, automate parking, and, increasingly, tie into wider smart city compliance systems.
What began as a niche police science experiment in 1970s Britain is today a global ecosystem worth billions, spanning law enforcement, transport, local government, airports, universities, and even homeowner associations. Cities like London, Sydney, San Francisco, and Auckland now run daily operations with ANPR embedded into their infrastructure. Councils and governments see it as a way to automate repetitive enforcement, free up human officers, and increase compliance rates. Vendors — from giants like Motorola Solutions to disruptors like Aero Ranger — are racing to shape what ANPR looks like in the AI era.
But ANPR isn’t just about efficiency and safety. It’s also a lightning rod for debate. Civil liberties groups warn that the technology is capable of building a real-time, nationwide surveillance dragnet if unchecked. Misreads, like confusing a “1” for an “I” or a “0” for an “O”, have occasionally led to wrongful fines or unnecessary stops. Community concerns around networks like Flock Safety highlight fears of overreach. And in jurisdictions like Canada, regulators have forced systems to throw away “non-hit” data entirely to preserve privacy.
This is what makes ANPR such a fascinating case study in 2025: it’s at once hero and controversy. A technology that has already proven its worth in catching criminals, automating tolls, and managing modern cities — but one that must constantly prove it deserves public trust.
This guide takes you on a deep dive into ANPR:
By the end, you’ll see ANPR not just as a set of cameras and algorithms, but as one of the most important — and contested — surveillance technologies of our time.
The story of ANPR begins in the UK in 1976, at the Police Scientific Development Branch (PSDB), a quiet research unit inside the Home Office. At a time when computers were still exotic in most workplaces, a small team led by Dr. Paul Wilkins and John Aston began experimenting with the idea of using optical character recognition (OCR) on video feeds to identify vehicles by their license plates.
It was an audacious idea. License plates were never designed for machines — they were designed for human eyes. They varied in size, colour, and clarity. They got dirty, dented, or obscured by bumpers or mudflaps. Lighting changed constantly — bright sun, heavy rain, fog. And computers of the 1970s were hardly fast enough to process video frames in real time.
Yet, through sheer persistence, the PSDB team built prototypes. Early tests involved black-and-white CCTV cameras linked to primitive OCR software. The system was slow and clunky, and cars often had to be stationary or moving at a crawl for a chance of a correct read.
By 1979, the prototypes had advanced enough to be field-tested. And in 1981, ANPR made history with its first successful police-assisted arrest: a stolen car was identified and stopped thanks to the new technology. For the officers involved, it felt like a leap into the future.
One of the first practical deployments came not on highways, but in infrastructure projects. In 1978, the technology was trialled at a UK river tunnel crossing as part of an experiment in automated tolling. Though crude, it demonstrated how ANPR could one day remove barriers, replacing toll booths with cameras.
At this stage, though, adoption was limited. The cost of computing and cameras was still high. Algorithms weren’t robust. The technology was an intriguing proof of concept, but it needed another decade of improvements before it could scale.
The 1980s were an era of steady iteration. The arrival of CCD sensors gave cameras sharper resolution. Personal computers became more powerful and affordable, enabling ANPR to run on standard PCs by the mid-1990s.
This period also saw experimentation with plate design to improve machine readability. Germany developed its FE-Schrift font in the 1990s — deliberately engineered so characters like “O” and “0” or “1” and “I” were visually distinct, reducing misreads. The UK standardised on the Charles Wright font, which provided uniform stroke widths for OCR clarity. These design choices mattered: they directly improved accuracy, reducing the “1 vs I” problem that plagued early systems.
The 1990s also marked the first attempts at wider rollouts. With cheaper computing and better cameras, ANPR units started appearing in UK traffic enforcement and local policing. Systems no longer required vehicles to be in predictable lanes or speeds; they could capture plates in more natural conditions.
While researchers built the tools, police leaders became ANPR’s early champions. In the UK, officers like Sir Bernard Hogan-Howe (later Commissioner of the Metropolitan Police) pushed for ANPR to be deployed against organized crime and terrorism. Hogan-Howe argued that plates were “the DNA of vehicles” — objective identifiers that could connect suspects to movements and crimes. His advocacy paved the way for nationwide adoption.
By the late 1990s, ANPR was no longer a lab curiosity. It was a tool police forces could depend on. And the stage was set for the technology’s global spread in the 2000s.
By the late 1990s, ANPR had moved from experimental prototypes to workable policing tools. But the technology’s global spread really began with bold projects in the early 2000s — each one showing how ANPR could solve big, visible problems.
Perhaps the most famous early project was London’s Congestion Charge, launched in 2003 under then-Mayor Ken Livingstone. The goal was simple: reduce traffic in central London by charging vehicles that entered the city center during peak hours.
ANPR was the backbone. Instead of toll booths or RFID tags, London installed a network of cameras at entry points. Each camera snapped the plates of vehicles entering, matched them against a paid database, and automatically issued fines for non-payers.
This was a watershed moment:
It also set off debates about surveillance creep. Civil liberties groups questioned whether a city-wide camera grid was proportionate. But politically, the project stuck. Congestion fell by 30%, revenue funded public transport, and ANPR became a permanent fixture of London’s transport network.
Meanwhile, on Europe’s highways, ANPR found another killer use-case: average-speed enforcement.
Instead of catching drivers only when they passed a radar gun, countries like Italy and the Netherlands deployed pairs of ANPR cameras kilometres apart. Each camera logged the vehicle’s plate and timestamp. If the average speed between the two points exceeded the limit, an automatic fine followed.
This system was:
Average-speed enforcement became one of the first continent-wide successes for ANPR, proving its worth beyond policing into broader traffic management.
In the U.S., adoption came later but grew rapidly in the wake of 9/11. National security concerns pushed agencies to invest in tools that could track criminal and terrorist movements.
Enter Steve Sevigny and Vigilant Solutions. Founded in 2009, Vigilant took a different tack: not just selling cameras, but building a giant commercial database of plate reads. By partnering with repossession companies and private operators, Vigilant amassed billions of plate records, which it sold back to police agencies as a subscription service.
For law enforcement, this was powerful:
But it also sparked public backlash. The American Civil Liberties Union (ACLU) published reports in 2013 warning of “mass routine location tracking” of innocent drivers. Jennifer Lynch of the Electronic Frontier Foundation (EFF) became one of the leading critics, warning that without oversight, ALPR could become a de facto nationwide surveillance grid.
Despite criticism, Vigilant’s model thrived. Motorola Solutions eventually acquired it, cementing ALPR as a core tool in U.S. policing.
In Canada, ANPR’s story unfolded differently. Pilot programs launched in the mid-2000s, but privacy commissioners intervened early.
Ann Cavoukian, Ontario’s commissioner, warned that retaining data on non-criminal vehicles was disproportionate. As a result, Canadian police forces often configured their systems to discard non-hit data immediately — meaning only matches to hotlists were stored.
This limited some of the “big data” investigative potential seen in the U.S., but it set a precedent: ANPR could only expand if it respected strict privacy boundaries. To this day, Canadian ANPR deployments are narrower and more carefully constrained than in most other Western countries.
In Australia, New South Wales Police trialled the first fixed ANPR camera in 2005. By 2009, mobile ANPR units were standard kit for highway patrol cars. These roof- or dash-mounted cameras could scan thousands of plates per shift, flagging unregistered vehicles, uninsured drivers, or stolen cars in real time.
Other states soon followed. By the 2010s, ANPR was entrenched in Australian policing, though still fragmented — each state running its own systems, with limited national integration.
In New Zealand, ANPR adoption was slower. But by the 2010s, community patrols began experimenting with it — using cameras to identify stolen vehicles and share data with police. More recently, platforms like Aero Ranger have formalised this, powering community safety groups in Upper Hutt and Wellington. This approach shows how ANPR can trickle down from national police forces to local volunteer groups, extending its reach in unexpected ways.
By the 2010s, some patterns were clear:
As ANPR matured from an experimental policing tool into a global industry, the shape of the market was determined not just by governments but by the companies that built the hardware, wrote the software, and developed the business models. Each vendor took a different path, influenced by the vision of its founders.
Origins and Vision
In the U.S., no company influenced the direction of ANPR more than Vigilant Solutions, founded in 2009 by Steve Sevigny. Sevigny’s insight was simple but disruptive: ANPR was not just about selling cameras — it was about collecting data.
Vigilant built one of the largest commercial plate databases in the world. By partnering with repossession companies, private security fleets, and contractors, it aggregated billions of license plate scans — far beyond what police agencies could capture alone.
Technology
Vigilant specialised in proprietary dual-lens cameras (infrared + colour) mounted on patrol cars and fixed locations. The cameras streamed plate reads into the LEARN (Law Enforcement Archival and Reporting Network) platform, a central cloud where agencies could:
Business Model
Unlike traditional hardware vendors, Vigilant leaned on subscription access to its plate database. Agencies paid for queries, alerts, and integration. This made the company profitable and sticky — once agencies had access to billions of records, they didn’t want to give it up.
Adoption and Controversy
Law enforcement loved the capability. Investigators used Vigilant’s tools to track suspects, locate stolen cars, and build patterns of movement. But civil liberties groups like the ACLU and EFF were alarmed. Jennifer Lynch of EFF called Vigilant’s system “a tool for mass surveillance by proxy.”
Despite controversy, Vigilant thrived and was eventually acquired by Motorola Solutions, embedding ALPR into Motorola’s broader public safety portfolio. Today, Motorola sells it as part of a total intelligence suite for police, linking ANPR with body cameras, radios, and command centre software.
Origins and Vision
In 2017, Garrett Langley, a former solar entrepreneur, founded Flock Safety with a bold pitch: give neighborhoods and small police departments the power of ANPR without the complexity.
Langley’s framing was different from Vigilant’s: where Vigilant sold “big data” to police, Flock sold community protection. His promise: with one solar-powered camera and an annual subscription, any neighbourhood or HOA could log every car entering or leaving.
Technology
Flock’s cameras are designed to be easy to deploy:
Business Model
Flock is a pure subscription service, costing around $2,500 per camera per year, which includes hardware, installation, maintenance, data storage, and software. The simplicity made it appealing to non-technical buyers.
Adoption
Flock exploded in growth, with over 1,200 police departments and 2,000 neighbourhoods in the U.S. adopting it within five years.
Controversy
Flock’s strength — its nationwide sharing network (Talon) — also became its biggest flashpoint. Critics worried it could be used to track sensitive journeys (e.g., visiting clinics). Civil liberties groups argued that “private neighbourhoods shouldn’t become nationwide surveillance nodes.”
Langley countered that Flock only sells to communities and police, and that data belongs to the local agency. Still, the debates continue.
Origins and Vision
Founded by Eric Yoo, Plate Recognizer represents a different philosophy: ANPR as a software service rather than a hardware product.
Technology
Plate Recognizer’s system is API-driven:
Business Model
Plate Recognizer’s pricing is transparent and low-cost:
This democratized ANPR, making it accessible to small councils, parking operators, and startups.
Adoption
Plate Recognizer is used globally, often embedded inside other products. Parking operators, smart access systems, and even small municipalities use it to automate enforcement or entry control.
Positioning
Yoo positions Plate Recogniser not as a “big brother” surveillance company, but as a toolkit provider. Its flexibility and affordability make it popular where budgets are tight.
Origins and Vision
From Australia comes Aero Ranger, led by John Colebrook. Unlike its competitors, Aero Ranger doesn’t position itself as just an ANPR vendor. Its vision is broader: AI-powered compliance ecosystems.
Technology
Business Model
Aero Ranger pioneered Hardware-as-a-Service for ANPR. Councils and agencies pay a monthly subscription covering equipment, software, and support. This makes budgeting predictable and lowers barriers to entry.
Adoption
In just 15 months, Aero Ranger has been adopted by 40+ cities, including London, Perth, Sydney, Melbourne, San Francisco, and Auckland. Community patrols in New Zealand also use it to spot stolen vehicles.
Positioning
Aero Ranger differentiates itself by embedding ANPR into local government workflows. Instead of selling “just cameras,” it sells a way to compress compliance work into digital automation — freeing officers from paperwork and boosting efficiency.
One of the paradoxes of ANPR is that its success depends not only on cameras and algorithms, but also on the humble license plate itself. Plates weren’t originally designed for machines, and that mismatch explains many of the technology’s growing pains — and the evolution of plate standards worldwide.
Globally, most countries follow the ISO 7591 standard, which specifies retro-reflective sheeting for plates. This ensures that at night, when illuminated by headlights or infrared lamps, plates bounce light back toward the source. For ANPR, this is a gift: the reflective background makes alphanumeric characters stand out against the plate.
Other ISO standards cover plate size, durability, and even how reflective materials should perform after years of wear. For agencies deploying ANPR, these standards mean predictability: as long as plates meet them, recognition software can be tuned accordingly.
United Kingdom & Australia:
Both countries use the Charles Wright font, standardised in 2001 for the UK and adopted by several Australian states. Its uniform stroke width and simple sans-serif design make it easy for OCR systems to distinguish characters, though early systems still struggled with “1” vs “I”.
Germany:
In the 1990s, Germany developed FE-Schrift (Fälschungserschwerende Schrift, or “forgery-impeding script”). Its characters are subtly distorted: the number “0” is oval, the letter “O” is round; the “1” has a serif, the “I” is wider. These quirks are invisible to most drivers but crucial for machines. They also make it harder to alter plates — turning a “3” into an “8” with tape, for example.
United States:
Here lies the biggest challenge. Each U.S. state issues its own plates, often with multiple designs, vanity options, and special backgrounds. Some use stylised fonts or multi-colour gradients. This means an ANPR system in the U.S. must be trained on hundreds of variations, increasing error rates.
Few issues illustrate the plate design challenge better than character confusion:
In the 1990s and early 2000s, these errors caused false positives — innocent drivers flagged because their plate was misread as a hotlisted one. Some even received wrongful fines. In the UK, media stories about drivers being ticketed despite being miles away from incidents dented early trust in ANPR.
Solutions:
Europe:
The EU harmonized plates with the blue band and country code, easing cross-border enforcement. Still, fonts and spacing differ per country.
Middle East:
Plates often mix scripts (Arabic + Western digits). ANPR must handle multiple OCR alphabets simultaneously.
Asia:
In countries like Japan and Korea, multi-line plates and non-Latin characters add complexity.
United States:
For vendors, plate diversity defines their competitive advantage.
The bottom line: better plate design = better ANPR performance. Where governments mandate machine-readable fonts and reflective sheeting, read rates can hit 95–98%. Where plates are messy, multi-style, or paper-based, accuracy can drop into the 70s.
For councils and agencies, plate design and standards aren’t academic — they directly affect enforcement efficiency. A 10% drop in accuracy could mean millions in lost tolls, unpaid parking, or missed stolen cars.
That’s why some governments have redesigned plates specifically for ANPR. The Netherlands, for instance, added subtle character spacing tweaks in 2002 to improve machine accuracy. Australia standardized reflective materials across states in the 2010s.
As ANPR becomes embedded in compliance and safety, expect more countries to harmonize plate designs for AI readability.
ANPR may have proven its worth in law enforcement and traffic management, but its expansion has always been shadowed by questions of privacy, proportionality, and oversight. For every project hailed as a success, there has been a debate about what it means for ordinary people to have their movements tracked by machines.
The UK, home of ANPR’s invention, is also home to some of its toughest debates.
By the 2000s, Britain had the densest ANPR network in Europe. Cameras tracked millions of vehicles daily, with data fed into the National ANPR Data Centre (NADC). Police could search this database for patterns of movement, suspects, or vehicles linked to crime.
But with scale came scrutiny. Civil liberties groups and individuals like Ed Bridges challenged mass camera use in court, arguing that blanket surveillance without cause was disproportionate.
In response, the Home Office introduced NASPLE (National ANPR Standards for Policing and Law Enforcement):
This framework allowed ANPR to continue while reassuring the public that data wouldn’t be kept indefinitely or misused.
In the U.S., ANPR (ALPR) grew rapidly in the 2000s, particularly through vendors like Vigilant Solutions and later Flock Safety. But unlike the UK, the U.S. lacked a central framework. Instead, rules were set state by state — or sometimes not at all.
That vacuum of oversight drew fire.
Some states reacted:
But others allowed wide-open use. This patchwork continues to this day: one police department may delete plate data in 30 days, while another may keep it for 5 years.
Canada took a different path. From the start, privacy commissioners intervened to limit the scope.
In Ontario, Ann Cavoukian made it clear: non-hit data (vehicles not linked to crime) should not be retained. Police forces adjusted, configuring their systems to delete anything that didn’t match a hotlist in real time.
The Royal Canadian Mounted Police (RCMP) initially resisted, but pressure from multiple provinces forced them to comply. Today, Canadian ANPR deployments are some of the most privacy-conscious in the world — limiting their investigative power, but protecting citizens from dragnet tracking.
In Australia, ANPR expanded across states largely without national debate. Mobile patrol units became standard from 2009 onward. Only in recent years have questions been raised about state-to-state data sharing and the potential for a centralised system.
In New Zealand, adoption has been slower and more community-focused — for example, Aero Ranger powering community patrols in Upper Hutt. Public concern there has been less about mass surveillance and more about ensuring the technology is used proportionately and transparently.
The most recent wave of criticism has focused on Flock Safety in the U.S.
Flock’s Talon network allows police in one jurisdiction to search data collected by cameras nationwide for up to 30 days. For law enforcement, this is invaluable — a stolen car crossing state lines can be tracked instantly.
But critics argue it creates a de facto nationwide surveillance system without legislation. Fears have been raised that such networks could be used to track individuals seeking healthcare, attending protests, or simply moving between states.
Civil liberties groups call it “outsourcing mass surveillance to the suburbs.” Flock insists the data belongs to local agencies and that their system is solving real crimes. Both are true — which is why the debate remains heated.
Beyond policy, there have been real cases of misuse:
These incidents highlight why auditing and transparency are as important as the technology itself.
The public pushback has shaped ANPR’s evolution in important ways:
The outcome is not uniform. In some places (like London), ANPR is accepted as part of city life. In others (like San Francisco), debates rage over whether it tips into surveillance overreach.
These individuals and groups are as much part of ANPR’s story as the engineers and police leaders who built it. They’ve ensured the technology matured under scrutiny, not in secrecy.
If there’s one thing every engineer, officer, or council ranger will tell you: ANPR lives and dies by deployment details. The algorithm might be state-of-the-art, the back-end cloud may be rock solid, but if the camera is placed poorly or the light floods out the plate, the system fails.
The golden rule: get the angle right.
In practice, police forces in the UK discovered that mounting on motorway gantries produced near-perfect reads. Meanwhile, suburban deployments in the U.S. (mailbox-height poles from Flock Safety, for example) trade some accuracy for affordability and community acceptance.
Plates look simple. They aren’t.
Tips learned the hard way:
Specialist vendors like Motorola’s Vigilant and Plate Recognizer supply software that auto-adjusts exposure. But even the best code can’t save a badly lit deployment.
One fascinating lesson: climate dictates ANPR design.
Flock Safety built its reputation on low-cost fixed cameras communities could afford. Aero Ranger flipped the model — making mobile subscription-based ANPR accessible to councils in Australia and NZ. That subscription approach eliminates capital costs but requires trust in recurring budgets.
Every enforcement technology creates countermeasures. ANPR is no exception.
Systems respond with:
Vendors all boast “>95% accuracy.” Real-world operators laugh at this.
The key lesson: ANPR is not infallible. That’s why retention, auditing, and officer discretion matter. In the UK, wrongful fines early on damaged trust. In Australia, false positives on rural roads nearly derailed community trust projects until councils improved signage and audit processes.
Deployment doesn’t stop with the camera.
This is where vendors differentiate:
ANPR isn’t just about reading plates anymore. The 2020s marked a tipping point: the technology began evolving from a single-purpose enforcement tool into a data-driven ecosystem for mobility, compliance, and safety.
Historically, plate recognition relied on big servers in data centres. That’s changing fast.
Australia and NZ councils lean toward edge-first models (driven by Aero Ranger) because it reduces network dependencies in rural deployments. U.S. police remain comfortable with cloud-first approaches due to FBI/DOJ data-sharing frameworks.
Plate recognition is just the start. The new frontier is Vehicle Fingerprinting:
Motorola’s Vigilant has pioneered integration into police RMS. Aero Ranger is weaving this into council compliance — imagine a ranger car that not only spots expired registrations but also notes cracked windscreens or illegal modifications.
ANPR started as a single tool, faced public suspicion, and is now returning as part of a bigger ecosystem that creates more value for communities.
Aero Ranger is arguably the only major vendor betting entirely on subscription-only ANPR. No capital sales, no outright ownership — just recurring SaaS tied to vehicles and rangers.
So far, it’s worked. Adoption across Australia, NZ, and even the UK shows councils prefer predictable OPEX over lumpy CAPEX — especially when solutions integrate into broader compliance.
With great adoption comes great suspicion.
Unlike Wi-Fi or Bluetooth, ANPR has no universal open standard.
Industry groups are pushing for a National Open Data Standard (especially in Australia), but adoption remains uncertain.
