LPR technology to combat crime
In Switzerland, Spain, and Germany, several thousand license plates are stolen annually. The exact number varies depending on the region and police measures, but according to statistics, the number of stolen license plates ranges between 100,000 and 200,000 per year in Germany alone. These thefts have significant impacts and are often associated with other crimes.
Why are license plates stolen
The theft of license plates is a growing problem that results from various criminal motivations. One main reason for the theft of license plates is the avoidance of toll fees and traffic violations. Criminals use stolen license plates to pass through toll stations or commit traffic violations without the penalties falling back on them. This allows them to avoid significant costs and evade law enforcement
Another common reason for the theft of license plates is refueling without paying. The perpetrators attach stolen license plates to their vehicles and refuel at gas stations, then flee without settling the bill. This results in significant financial losses for gas station operators and makes it difficult to track down the perpetrators
Furthermore, stolen license plates are often used in other crimes. In burglaries or drug offenses, the perpetrators can conceal their identity and remain undetected by using stolen license plates. This significantly complicates police investigations and often allows the criminals to go unpunished
Another motive for the theft of license plates is the sale of stolen vehicles. Criminals use stolen license plates to temporarily “legalize” stolen vehicles. This facilitates the sale of the vehicles, as they appear to be legally registered vehicles at first glance. Buyers are deceived, and the stolen vehicles can be circulated unnoticed.
Accuracy and error rate of ANPR systems
The recognition rate of modern ANPR systems is typically stated by the main providers to be around 90% to 95%. However, studies have shown that commercially available solutions such as Sighthound achieve a recognition rate of 86%, PlateRecognizer 67%, and OpenALPR 77%. This discrepancy between the manufacturers’ claims and the results of independent studies indicates a real error rate that is higher than often claimed
An error rate of just 10% in ANPR systems means that one in ten license plates is incorrectly recognized or not captured. On a main road in Zurich with 2000 vehicles daily, 200 vehicles would thus not be correctly captured. This could lead to undetected traffic violations, increasing the risk of accidents, and impairing the efficiency of crime fighting, as stolen or wanted vehicles could go unnoticed. Additionally, police forces would need to allocate more resources to check false alarms, reducing the effectiveness of law enforcement. Measures to optimize traffic flow could also be affected, leading to more frequent and longer traffic jams. If you scale this up to a highway like the Swiss A1 with around 150,000 cars per day, 10% is a lot.
Vaidio: Advanced License Plate Recognition with YOLO and ResNet
Advantages of a high recognition rate in LPR technology for crime fighting
With a high recognition rate, LPR technology offers numerous advantages that can significantly improve the efficiency and effectiveness of security and law enforcement agencies
- Increased security: ANPR systems with high recognition accuracy identify stolen or wanted vehicles faster and more reliably. This allows law enforcement agencies to respond more effectively and enhance public safety
- Efficiency increase: With more precise ANPR systems, police authorities and security services use their resources more efficiently. Fewer false alarms and incorrect detections mean that less time and effort need to be spent on unnecessary investigations
- Better traffic monitoring: More accurate systems monitor and enforce traffic rules more effectively, leading to safer roads. This helps reduce traffic accidents and enforce speed limits and other traffic regulations.
- Reduction of crime: The quick identification and tracking of suspicious vehicles helps to reduce the overall crime rate. Law enforcement agencies can respond more quickly to criminal activities and pursue suspects
- Automated processes: Higher accuracy enables better automation of processes such as toll collection, parking management, and access control. This leads to more efficient management and reduces the need for manual interventions
Overall, it is evident that advanced ANPR systems like those from Vaidio, with their high recognition rate of 99.9%, can offer significant advantages for public safety and efficiency. By using modern deep learning techniques and specialized algorithms such as YOLOv5 and ResNet, the error rate is minimized, further enhancing the reliability and effectiveness of these systems.
Sources
- Astra; Federal Roads Office
- A Real-Time License Plate Detection and Recognition Model in Unconstrained Scenarios
- Automatic License Plate Recognition in Real-World Traffic Videos Captured in Unconstrained Environment by a Mobile Camera
- A Multi-Stage Deep-Learning-Based Vehicle and License Plate Recognition System with Real-Time Edge Inference
- NVIDIA
- Gutschild mobility blog
- authority.com
Anne-Katrin Michelmann
Date 01.09.2024