By Joseph Kania
A plate hit is not evidence. It is a lead. What you do with it over the next few hours and days is what decides whether the case you hand the prosecutor holds up when the defense goes after it.
I worked organized auto theft for the City of Pittsburgh Bureau of Police for close to 10 years, and ALPR touched nearly every case I built. I later sat as lead expert witness in the prosecutions that came out of them. Testifying teaches you things working the street does not. You find out in a hurry which investigators understood what an ALPR hit proved and which ones assumed it proved more.
This is for the investigator who wants more out of ALPR than a real-time alert. The technology helps us work. It does not do the work for us.
Where the investigation begins
Almost every ALPR case starts in one of three places, each asking something different of you:
- Known plate: This is the easiest scenario. A stolen vehicle, a wanted subject with a car registered to him, a BOLO out of a neighboring jurisdiction. You have the number, so ALPR either locates the vehicle in real time or reconstructs where it has been. Strongest lead you are going to get.
- A description without a plate: This is harder. A witness at a robbery remembers a dark-colored SUV. Responding officers give you a partial and little else. Cases like that used to die on the desk. Now, if you have coverage on the corridors that matter, you can pull every detection within a time frame and a defined radius, then filter by make, model and color where your system supports it. That turns somewhere in the city into a set you can review. Expect a lot of wrong ones first.
- A location and a clock: That’s it, that’s all you have. A shooting on a corner between 10:15 and 10:45 p.m. Every vehicle that passed a nearby camera in that window is potentially relevant. Most are not. A few belong to witnesses who never came forward. One or two might be the car used in the crime. You have a list of vehicles that were present, and present is a long way from involved.
Building timelines and identifying movement patterns
One thing comes before the rest. Run timeline queries only against plates tied to an active investigation, with a purpose you have written down. Querying outside that is the practice that has generated public criticism of ALPR, and it costs credibility for every agency using the technology the right way.
A single detection tells you one vehicle passed one camera at one time. Pull 30 days of detections on that same plate, plot them against their timestamps, and the vehicle starts describing its owner’s life to you.
Home and work surface first. A car hitting the same intersections every weekday between 7:00 and 7:30 in the morning, then again between 5:30 and 6:00 in the evening, tells you roughly where the driver sleeps and works.
Destinations and associations come next. When that same vehicle keeps turning up at a location you are already looking at for unrelated reasons, at consistent times, that is a relationship worth pulling on.
Anomalies are often the most useful part of all. Once you know what normal looks like, a detection that breaks the pattern becomes a lead in its own right.
Two cautions. No detection does not mean no movement, because cameras have coverage gaps. And patterns need history, so give yourself a month before drawing conclusions about how somebody moves. That said, if the same gap keeps recurring in the same shape, look at it twice.
When the gap in the data is the lead
I just said a gap does not mean a vehicle sat still. But a gap that repeats itself in the same shape is telling you something, and a case I worked in 2024 turned on nothing else.
We had vehicles fleeing traffic stops. An officer would attempt a stop, the driver would run, and the car would pass an ALPR camera on the way. What caught my attention was that we had daytime reads on these plates and nothing at all after dark, even though the stops were happening in the evening. The plates were in the system. They just stopped existing once the sun went down.
Finding the signature
I ran the detections through PlateSmart ALPR and charted the reads against time of day, and there was the answer. License plates are manufactured with retroreflective sheeting. That sheeting bounces a camera’s infrared illumination back to the lens and lets ALPR read a plate in low light. These plates were printed. No reflective surface at all. In daylight, the cameras picked them up on ambient light like any other plate. At night, there was nothing for the infrared to reflect off, so the plate read as blank.
Turning the signature into a target list
Once I knew what to look for, I queried the ALPR database for the same daytime-only pattern across other plates in the region. Turned out, it wasn’t just one vehicle. Close to a dozen came back, each running the same pattern of daytime reads, nighttime blanks and evening stops that ended in a pursuit. That gave me a working target list I did not have earlier.
Building the map
With the target list in hand, I pulled the full detection history on each plate and mapped it against time and location. The same cluster of intersections showed up repeatedly, at similar hours, forming a corridor that pointed at a shared destination. That destination turned out to be a chop shop. ALPR did not put me at the door on its own, but it narrowed a region-wide search to a corridor small enough to work with conventional methods.
What ALPR contributed at conviction
The chop shop investigation ran to a much larger narcotics case in the Pittsburgh region, and multiple convictions came out of it. The ALPR record carried real evidentiary weight. It placed specific vehicles at specific times and locations in a way witness testimony could not, corroborated informant statements about movement patterns, and gave the jury a timeline they could follow without having to trust any single source on its own. The vehicles came back stolen, most out of Ohio, taken off dealership lots and out of private driveways. Not random grabs. High-end MOPAR products, fast enough to lose a marked unit and sought after in the street racing scene nationwide. Somebody was stealing to order.
Two things came out of that case
ALPR data has physical properties and not just informational ones, and knowing what a plate has to be made of for a camera to see it turned missing reads into a fraud indicator. I also never would have found the pattern one plate at a time. It only showed up in aggregate.
Connecting ALPR to the rest of the record
ALPR is one source among many, and it is worth the least when it stands alone.
On my cases, I pulled CAD history and incident reports for the times and locations in question. Security video from businesses along a vehicle’s known route, obtained by asking or by subpoena. Witness statements. Cell tower data and, in complicated cases, financial records, both obtained through proper legal process. Each tells you a piece of something, and ALPR is frequently what ties them together.
A plate detection at 10:32 p.m. at a given intersection is a data point and not much more. Set it next to a CAD call at that location, a business owner two doors down with video covering the window, and a witness describing a matching vehicle, and now you have something a prosecutor can use.
That is the real function of this technology. It tells you which other records are worth your time. Investigators who treat it as the answer instead of the roadmap walk past the evidence that makes the case, and a case resting on ALPR alone has one point of failure that defense counsel will find.
When is it a lead, and when is it evidence?
An ALPR hit tells you a plate matching a specific number passed a specific camera at a specific time. That is the whole claim. Any competent defense attorney knows the three things it does not tell you, and you should know them better:
- Who was driving. The registered owner is frequently not behind the wheel. Family members, friends and thieves all drive cars that are not theirs.
- Whether the plate belonged on that vehicle. Cloning, swapping and counterfeits are standard tradecraft in organized vehicle crime.
- Whether the vehicle was actually there. Calibration drift, misreads and, in rare cases, deliberate tampering produce detections that are simply wrong.
None of that makes ALPR unreliable. It makes an ALPR hit a starting point that must be corroborated before it is fit for a courtroom. Cases come apart on the stand when investigators treat a lead as though it has already been proven.
Documentation and query discipline
The prosecutions that held up came down to paperwork, not the strength of the hit.
Write down why before you run the plate. A case number, an incident type, supervisor authorization if your policy calls for it. When a defense attorney asks eight months later why you queried that plate, the answer must already be in the record.
Preserve full metadata on every hit that makes it into the file. Image, timestamp, GPS coordinates, camera information. The defense will look for gaps in precisely that record.
Document the analysis while you are doing it. Which detections did you look at, which did you rule out, and why? Anything you did not write down, they will argue you never did.
Handle exculpatory material with the same care. A single query can produce evidence for both sides at once. Both halves belong in the record. Ordinary Brady practice.
| Common mistakes to avoid |
|
1. Treating a plate hit as arrest justification. A hit on a stolen vehicle gives you grounds to stop the car and investigate. Standing alone, it is not probable cause to arrest whoever happens to be driving. |
|
2. Assuming the registered owner is the driver. The habit comes out of routine traffic enforcement. Carry it into investigative work and it will hurt you. |
|
3. Skipping corroboration because the clock is running. With a suspect at large, every hit feels like the break the case needed. Corroborate before the arrest and before the warrant application. |
Closing thoughts
ALPR has reshaped how organized auto theft gets investigated, and it can do the same in nearly any other case category. But a hit is worth only as much as the officer at the terminal makes it worth.
The detectives I watched close the strongest cases were never the ones running the most queries. They were the ones who treated every hit as a lead, corroborated all of it, documented each step as they took it, and could walk a jury from a plate on an ALPR capture to a defendant in the box without losing anybody.
That is the craft of the work. No amount of sophistication in the technology is going to replace it.
The cases of misuse that have made recent headlines are real, and they deserve scrutiny. But they should not cast a shadow over what this technology does in the hands of investigators who work with it with discipline. Used ethically and responsibly, ALPR helps close cases, save lives, recover stolen vehicles, bring offenders to justice and make our communities safer every day.
About the author
Joseph Kania is an account executive with PlateSmart Technologies and a former detective with the City of Pittsburgh Bureau of Police, where he served for nearly a decade. During his law enforcement career, he specialized in organized auto theft investigations, used ALPR and vehicle tracking systems to identify, locate and recover stolen vehicles, and served as the lead expert witness in criminal proceedings involving vehicle crimes and technology-driven investigations. Today, Kania works with law enforcement agencies throughout Western Pennsylvania and across the country, advising on the implementation and strategic use of ALPR and vehicle intelligence solutions.
