Automatic License Plate Redaction

Law enforcement agencies and the public everywhere are becoming familiar with automatic license plate readers (ALPRs). These types of systems while they have become commonplace have actually changed, improved and become far more sophisticated since their beginnings. Like many software and computing systems new ideas and better features are integrated all the time creating a situation for constant change, learning and renewal.

Automatic license plate detection today can be used in a variety of situations and for criminal investigations. The cameras can be mounted virtually anywhere. While often installed on many police cruisers, these systems can be mounted on street or light poles, overpasses, and even mobile stations with speed sensors. The camera takes a snapshot or video that also uses reading software to process the information that comes along with it such as the license plate numbers, time, date, location and even in some instances the number of passengers in the vehicle.

Types of Data Collected

The data collected from automatic license plate detection systems can be used in a variety of ways once the information is loaded into the main server or database. Using a variety of queries, officers can use the information in the system to determine if the vehicle has been near or suspected of being involved in a crime scene. Long-term data can show the travel patterns of the vehicle, which can help solve some investigations. The data can even be more broadly defined to output results of other vehicles that may be typically found nearby.

Linked up, law enforcement agencies can combine their data and share amongst themselves to be more effective in policing on a larger scale. Locally collected data can be shared with many agencies with a network to build on evidence that may cross many jurisdictions. The data collected can be stored for up to 5 years and be referenced at any time. The amount of data that can be shared in regional systems can be mind-boggling, as there are some automatic license plate detection systems that brag of the ability to capture and store the data on thousands of vehicles per minute.

Public Transparency and Privacy

ALPR data can be troubling to the general public when concerns about civilian privacy and how it can be used to negatively impact private citizens who are engaged in protected First Amendment activities. The databases that store this sensitive information while at times can be part of a larger police department system or server, more often than not departments use private companies to maintain these databases. These companies make the information available to all law enforcement agencies, even those who may not have their own ALPR system.

Police most often use the data for immediate police action, which can be current, real-time investigations where the information is needed to solve a crime or take urgent precautions for public safety. The information can also be used to solve ongoing cases or historical investigations. Crimes that have already been committed in which law enforcement is working surrounding details, such as checking for other vehicles found in the vicinity when the crime occurred.

Freedom of Information Act (FOIA) requests can come from courts, the public, or even local journalists and news channels when a crime of local interest or public safety is involved. Before any video footage can be released for public consumption, the video must go through a redaction process. Video redaction in evidentiary videos often entails going frame by frame to ensure that private citizens, minors or other protected information is blurred or blocked out from view. Legal requirements, which can vary by jurisdiction, mandate that privacy concerns are taken into account, are treated seriously, and action taken to protect citizen’s privacy. In some instances, lives can even be at stake.

Transparency is also a vital importance to law enforcement agencies, news organizations and the general public because it is the basis for building trust between citizens and officers. When local communities have a good understanding of their own law enforcement agency’s practices and protocols the level of trust grows. Trusting law enforcement versus fearing them, can make the difference when a private citizen decides to pick up the phone and report a crime. They have to know they will be protected and can trust that their local departments will always come through for them.

Automatic License Plate Redaction

Like many types of video redaction processes in which data sets are entered, searched for, then blurred out for privacy, license plate video redaction is no different. Video redaction for all types of police video evidence goes through the same approach as put in place by the department’s video evidence protocols and policies. These policies generally define how video evidence is transmitted, who handles the information, the department or officers in charge of video redaction, storage and requirements for disbursement to courts, legal counsel, or journalists.

As with other privacy concerns, such as personal identity with an individual’s face, license plates can give out personal information that can also put people at risk when solving crimes. License plate video redaction can be done by entering in the data points to the AI interface which can then search through frame by frame to locate and blur or block out any license plates in the video footage. Using quality video redaction software lowers the risk of any missed video frames or license plates getting through the redaction process.

Video redaction software is a must have for many police departments and other law enforcement agencies due to the time and cost savings that it provides. The added benefit of a machine learning AI system being more accurate than human officers gives a greater safety measure for all. Individual officers are still needed to enter the data sets for redaction and to review the results before any information is released. The dual system allows the redaction to be done much faster, more data to be stored and the review and release to the courts done with ease.

Automatic Video Redaction Saves

License plate detection and redaction can be fairly simple for the complex machine learning AI-based video redaction software. The data to be searched for or tagged is entered by the information specialists or trained law enforcement officer and the system then searches for all objects that match the data set within the video. Due to the uniformity of license plates, the rectangular shape, the pattern of numbers and letters, powerful video redaction software can easily locate and remove license plates footage from any video.

Automating license plate detection and redaction can be easier than faces but still offers some complex situations that may make all plates difficult to detect, such as partially obstructed plates or side views. Algorithms, or steps of instructions, in locating all license plates in an image became more proficient when the first instruction was not the shape but to locate areas with white or yellow concentrations in the photograph. The second step is to read for characters or numbers and then finally the shape of the area that the colors encompass.

While there is still a need for manual checking of footage before it is released the automation of license plate video redaction cut hours of manual labor from the department overhead. To individually check and redact video without automated systems, a single 5-minute video could take as long as 50 minutes to edit. The value of added accuracy is a plus. Automatic license plate redaction systems have several benefits for every law enforcement agency; four of the main benefits are listed here:

• Time: A single case with an estimated 7,700MB of data to be analyzed can take an individual officer up to 40 hours of forensic processing time. Any time that this amount of time can be reduced to mere minutes or even an hour or two, then minimally reviewed for accuracy, the time saved also equals money saved.
• Money: Hours of redaction time cost departments overhead. There are not any federal funds available to cover these costs currently and many states give limited if any funding to their law enforcement agencies for video services. Any way that a department can get a more accurate job accomplished with less man hours increases productivity in other areas and overall morale in the department.
• Public Safety: Automatic license plate detection and redaction improves public safety. Using the detection cameras, police are able to find more stolen vehicles, missing persons, or apprehend wanted suspects. The systems automatically alert officers to uninsured drivers, suspended licenses, and even previously arrested violent offenders. This type of automation to alert officers in real-time creates safer neighborhoods and communities everywhere.
• Public Trust: By allowing access to information in a speedy fashion, it gives a greater transparency between policing agencies and the public it serves. With automatic license plate detection and redaction, the video processing time is shortened and if the crime is in the local interest or concerning public safety it can be readily released quickly, increasing community trust, and possibly saving lives.

Setting the Hot-List

One of the great advantages to automatic license plate detection and redaction software systems is the ability to use the system in real-time crime solving solutions. Local law enforcement agencies often have their own daily list of wanted felons and stolen vehicles. The information available on the vehicles is called the hot-list and is pre-loaded into the system. While on patrols, officers’ vehicle cameras can pick up these plates and automatically notify law enforcement. The information is given as to the nature of the offense, the type of vehicle, the name and license information on the offender, and other real-time data designed to make the arrest go smoother and keep both the officer and community safe. Overall, the advantages of automatic license plate detection and redaction systems are far reaching and will continue to benefit all departments.