Last updated:
🗺️ ALPR Cameras by US State — A Free Dataset
116,655 licence plate readers, mapped. We queried OpenStreetMap for every ALPR camera in all 50 states and DC, normalised by population, and published the whole thing — CSV, JSON, and the query, so you can check us.
- 16 Oct 2026 — corrections: the IJ count is refreshed to 277 (15 Oct). El Paso is now “more than 100 cameras”, not 150 (Spectrum News). The unsourced 75,000–100,000 deployment estimate is replaced by sourced growth figures, and the claim that the registry “drifts only slowly” is corrected. Maine is third-lowest per capita, not fourth. The per-state counts are unchanged.
- 277 local governments had cancelled or not renewed ALPR contracts since the start of 2025 as of 15 October 2026, per the Institute for Justice database (pauses and vendor switches excluded). IJ (read 16 Oct 2026)
- Named cancellations by state — VA: Henrico County (6 Oct, >120 cameras; VPM), Arlington County (Sep), Giles County sheriff (WDBJ7, 26 Aug); TX: El Paso (Sep, more than 100 cameras out of use after the state funding pause; Spectrum News, 14 Sep); OH: Hamilton County (Oct, early termination), Oberlin (end of Sep, over Flock's new user agreement and pricing rather than privacy); WI: Milwaukee County (Oct, early termination); MA: Nantucket (Sep, early termination); MI: South Lyon (5–2 vote, 29 Sep), Sturgis (let expire); NY: Newburgh (resolution 29 Sep); FL: Boynton Beach (contract expires 18 Nov 2026). IJ (6 Oct 2026); Jalopnik (4 Oct 2026)
- Why the trackers disagree: IJ counts only completed cancellations and non-renewals since January 2025 (277); DeFlock's “Recent Victories” counts cities that rejected plate readers, including proposals that never became contracts (153 per Jalopnik, 4 Oct). A Newsweek aggregation of DeFlock data counted 149 cities and counties that cancelled or rejected contracts or deactivated cameras from January to August 2026, more than half in August. Jalopnik (4 Oct 2026) Same wave, three definitions. Trackers: DeFlock The USA cancellations (an independent site), Who Approved This (contract approvals and renewal dates), and the IJ database above. Full context in the Flock guide.
Key findings
We counted 116,655 ALPR cameras tagged in OpenStreetMap across all 51 US jurisdictions on 17 July 2026 — a national rate of 34.3 per 100,000 residents. Raw counts mostly just track population, so the useful number is per-capita, and there the picture changes completely: Georgia leads at 70.4 per 100k — double the national rate, and Flock Safety is based in Atlanta TechCrunch (13 Mar 2025). Kansas is second at 62.8 despite holding 0.9% of the US population — and Kansas accounted for 3 of the 24 documented officer-stalking cases in our July 2026 catalogue. New York, despite 3,958 cameras, sits at just 19.9 per 100k. At the bottom: Alaska has zero tagged, and Maine, Vermont, Montana and New Hampshire all sit near 4 per 100k. The one caveat that governs everything here: OpenStreetMap is crowdsourced, so these counts measure deployment multiplied by volunteer mapping effort. Every figure should be read as one imperfect measurement, not a census.
DeFlock renders a map of ALPR cameras. Eyes Off Indiana publishes state-level statistics for one state. Nobody was publishing the national breakdown as a citable table — so we pulled it. Everything below is derived from a public OpenStreetMap tag and official Census population estimates. The query is on this page and the raw data is free. Don't trust the numbers; re-run them.
By Ned Walsch · Collected 17 July 2026 · OSM data © OpenStreetMap contributors (ODbL)
The method — reproduce this in 30 seconds
Want the cameras around one address instead of a state total? DeFlock’s ALPR map shows the same crowd-sourced OpenStreetMap camera data street by street. For 2025–26 state ALPR laws, see License-Plate Readers and the 2026 Midterms.
There is no proprietary data here and no scraping. OpenStreetMap has a public tag for these devices, surveillance:type=ALPR, and Overpass will count them inside any boundary it knows about. Paste this into Overpass Turbo, swap the state name, run it:
[out:json][timeout:90];
area["name"="Georgia"]["admin_level"="4"]["boundary"="administrative"]->.a;
node["surveillance:type"="ALPR"](area.a);
out count;
Change out count to out body to get every camera's coordinates instead of the total. Many nodes also carry operator, direction and manufacturer, so you can filter to one vendor or work out which way a camera faces. That's it — that's the whole dataset, and it's why we're comfortable publishing it: you don't have to believe us.
Read this before you cite anything below
surveillance:type=ALPR tag captures any plate reader a volunteer identifies — police, private, tolling, parking enforcement. OSM may also over-count in places, through duplicate nodes or devices wrongly tagged as plate readers. We can't resolve this from the data, so we're flagging it rather than picking whichever reading flatters the dataset. Treat these figures as one imperfect measurement to be cross-checked, not an authority.Still anchored to 17 July 2026 as of October 2026. This snapshot is dated 17 July 2026, and the registry has grown fast since. RoadSigns.com, analysing DeFlock and OpenStreetMap data, counted 135,844 mapped ALPR locations of all brands on 26 August 2026. That is a net gain of 19,760 since its own 17 July snapshot. RoadSigns.com Jalopnik reported on 4 October that DeFlock maps 143,356 US plate readers. Jalopnik (4 Oct 2026) We keep this table anchored to its pull date rather than chasing a moving total; re-run the query yourself for a live number. Note that a cancelled contract (see the October 2026 update) only lowers a state's count once the cameras come down and a volunteer deletes the node, so the table lags the contract map by months.
The data: all 51 jurisdictions, ranked by saturation
Sorted by cameras per 100,000 residents, because raw counts mostly just rediscover which states have the most people. Bar shading is relative to the highest rate.
| # | State | Cameras (OSM) | Population | Per 100k |
|---|---|---|---|---|
| 1 | Georgia | 7,872 | 11,180,878 | 70.4 |
| 2 | Kansas | 1,866 | 2,970,606 | 62.8 |
| 3 | Ohio | 5,961 | 11,883,304 | 50.2 |
| 4 | Illinois | 6,197 | 12,710,158 | 48.8 |
| 5 | Missouri | 2,925 | 6,245,466 | 46.8 |
| 6 | Indiana | 3,203 | 6,924,275 | 46.3 |
| 7 | Texas | 14,257 | 31,290,831 | 45.6 |
| 8 | California | 17,870 | 39,431,263 | 45.3 |
| 9 | Alabama | 2,243 | 5,157,699 | 43.5 |
| 10 | Colorado | 2,394 | 5,957,493 | 40.2 |
| 11 | Tennessee | 2,894 | 7,227,750 | 40.0 |
| 12 | New Mexico | 848 | 2,130,256 | 39.8 |
| 13 | Louisiana | 1,654 | 4,597,740 | 36.0 |
| 14 | Delaware | 375 | 1,051,917 | 35.6 |
| 15 | Kentucky | 1,582 | 4,588,372 | 34.5 |
| 16 | Florida | 7,972 | 23,372,215 | 34.1 |
| 17 | Wisconsin | 2,011 | 5,960,975 | 33.7 |
| 18 | Michigan | 3,377 | 10,140,459 | 33.3 |
| 19 | Virginia | 2,883 | 8,811,195 | 32.7 |
| 20 | Oklahoma | 1,249 | 4,095,393 | 30.5 |
| 21 | Arizona | 2,311 | 7,582,384 | 30.5 |
| 22 | Arkansas | 933 | 3,088,354 | 30.2 |
| 23 | South Carolina | 1,594 | 5,478,831 | 29.1 |
| 24 | North Carolina | 3,124 | 11,046,024 | 28.3 |
| 25 | Utah | 956 | 3,503,613 | 27.3 |
| 26 | Mississippi | 781 | 2,943,045 | 26.5 |
| 27 | Iowa | 820 | 3,241,488 | 25.3 |
| 28 | Rhode Island | 267 | 1,112,308 | 24.0 |
| 29 | Washington | 1,906 | 7,958,180 | 23.9 |
| 30 | Nebraska | 468 | 2,005,465 | 23.3 |
| 31 | Connecticut | 850 | 3,675,069 | 23.1 |
| 32 | Minnesota | 1,339 | 5,793,151 | 23.1 |
| 33 | Nevada | 717 | 3,267,467 | 21.9 |
| 34 | New York | 3,958 | 19,867,248 | 19.9 |
| 35 | Idaho | 331 | 2,001,619 | 16.5 |
| 36 | North Dakota | 126 | 796,568 | 15.8 |
| 37 | Pennsylvania | 2,017 | 13,078,751 | 15.4 |
| 38 | New Jersey | 1,464 | 9,500,851 | 15.4 |
| 39 | Massachusetts | 1,026 | 7,136,171 | 14.4 |
| 40 | South Dakota | 133 | 924,669 | 14.4 |
| 41 | Maryland | 833 | 6,263,220 | 13.3 |
| 42 | District of Columbia | 86 | 702,250 | 12.2 |
| 43 | West Virginia | 202 | 1,769,979 | 11.4 |
| 44 | Oregon | 460 | 4,272,371 | 10.8 |
| 45 | Wyoming | 61 | 587,618 | 10.4 |
| 46 | New Hampshire | 65 | 1,409,032 | 4.6 |
| 47 | Vermont | 28 | 648,493 | 4.3 |
| 48 | Montana | 49 | 1,137,233 | 4.3 |
| 49 | Maine | 58 | 1,405,012 | 4.1 |
| 50 | Hawaii | 59 | 1,446,146 | 4.1 |
| 51 | Alaska | 0 | 740,133 | 0.0 |
| UNITED STATES | 116,655 | 340,110,988 | 34.3 |
What we checked before publishing
A dataset is only as good as its failure modes, so here is what we tested and what it showed. All of it is re-runnable.
| Check | Why it matters | Result |
|---|---|---|
| Is Alaska's zero real? | A zero usually means a broken query, not an empty state. | Real. Alaska's boundary resolves and returns 27 surveillance nodes of other types — there are simply no ALPR tags. The zero is an absence of mapping, not an error. |
| Does tag case cost us cameras? | If some mappers write alpr and we query ALPR, we'd undercount silently. | No. A case-insensitive regex against Georgia returned an identical count to the exact-match query. The tag is used consistently in uppercase. |
| Does the count drift? | If OSM changes hourly, a static number is misleading. | Yes, and materially. Georgia read 7,872 when we pulled it and 7,885 four hours later — thirteen cameras added in an afternoon. This is a living dataset. Cite the date or re-run the query. |
| Do other tags overlap? | If ALPRs are also tagged another way, we'd miss them. | Partially unresolved. Georgia carries 6,820 nodes tagged surveillance:zone=traffic, which covers speed and traffic cameras and may overlap ALPRs in ways we haven't disentangled. We did not merge them — this count is the ALPR tag only. |
What stands out — and what it does and doesn't prove
Georgia, at double the national rate
Georgia tops the per-capita table at 70.4 per 100k, against a national rate of 34.3. It's also where Flock Safety is headquartered. That is a striking coincidence, and it is worth saying plainly that a coincidence is all this dataset can establish. Two ordinary explanations fit equally well: a company sells hardest in its home market, which is unremarkable; or a hometown company generates local publicity, which motivates local volunteers to go mapping. Both produce exactly this number. We can't distinguish them from here — and the second one would mean Georgia is merely the best-mapped state, not the most-surveilled.
Kansas: 0.9% of the population, 12.5% of the stalking cases (July 2026)
Kansas ranks second at 62.8 per 100k. Separately, our catalogue of documented misuse includes three Kansas officers — Victor Heiar (Kechi), Lee Nygaard (Sedgwick) and Kyle Rector (Bonner Springs). Three of twenty-four cases, from a state with under one percent of the country's people.
The bottom of the table
Alaska: zero. Then Hawaii (4.1), Maine (4.1), Montana (4.3), Vermont (4.3), New Hampshire (4.6). Maine is, as far as we can establish, the only state that bans private ALPR use outright: its law says “a person may not use an automated license plate recognition system,” with exceptions only for the state transportation department and law-enforcement agencies 29-A M.R.S. §2117-A. That is tempting to point at, until you notice that every other state down here is rural and none of them has that law. Rurality alone predicts this cluster. The law may contribute; this dataset cannot show it does.
More interesting is New York: 3,958 cameras, seventh by raw count, but only 19.9 per 100k — well below the national rate and barely a quarter of Georgia's. Whatever is driving saturation, it isn't simply urban density.
The number grew
Our pull counted 116,655 on 17 July 2026. Six weeks later RoadSigns.com counted 135,844 mapped locations of all brands (26 August), and by 4 October Jalopnik reported DeFlock mapping 143,356. RoadSigns.com Jalopnik (4 Oct 2026) The likeliest explanation is more mapping as well as more cameras: ALPR became a mainstream story across 2025 and 2026, and stories bring volunteers. Different counters may also use different boundaries. Either way it's the argument for publishing a method instead of a number — re-run the query and see what today says.
What this is actually for
Frequently asked questions
What exactly does this dataset count?
Nodes in OpenStreetMap tagged surveillance:type=ALPR, within each US state's administrative boundary, counted on 17 July 2026 via the Overpass API. Population is the US Census Bureau's Vintage 2024 state estimates. The rate column is simply cameras divided by population, times 100,000. Every number is reproducible — the query is on this page, and the raw CSV and JSON are free to download.
Is this a count of every ALPR camera in America?
No, and this is the single most important caveat on the page. OpenStreetMap is crowdsourced: a camera appears in this dataset only if a volunteer found it, identified it, and tagged it. So each number reflects two things multiplied together — how many cameras exist, and how much mapping effort has happened there. A state with an active DeFlock community will look denser than an equally-surveilled state where nobody has gone looking. Treat these as a floor, never a census. The real deployment count is higher everywhere, and unknowably so.
Why does Georgia top the per-capita table?
Georgia has 70.4 ALPR cameras tagged per 100,000 residents — roughly double the national rate of 34.3, and the highest of any state. Flock Safety is headquartered in Atlanta. That is a genuinely striking coincidence and we are flagging it as exactly that: a coincidence worth investigating, not a demonstrated causal link. There are at least two innocent explanations. Flock may simply have sold hardest in its home market, which would be unremarkable commercial behaviour. Or Georgia may have more people mapping cameras — the same publicity that surrounds a hometown company can motivate volunteers. Both would produce this result. We can't separate them from this dataset alone, and anyone who tells you they can from a single correlation is overreaching.
What's going on with Kansas?
Kansas is second per-capita at 62.8 per 100,000 — nearly double the national rate, from a state with 0.9% of the US population. Separately, our catalogue of documented ALPR stalking cases includes three Kansas officers: Victor Heiar in Kechi, Lee Nygaard in Sedgwick and Kyle Rector in Bonner Springs. That was 3 of the 24 cases in our catalogue in July 2026 — 12.5% of the total then — from 0.9% of the population. The catalogue now follows the Institute for Justice’s list, which had 82 police stalking cases on 16 October 2026; we have not yet recounted Kansas against it. We want to be careful here, because the obvious reading is probably wrong. Kansas may not have more misuse; it may have more detection. Cases surface when someone audits, when a victim is told they can check, or when a local reporter files a records request. A state that catches three officers may be doing oversight better than a state that has caught none. The honest conclusion is that both numbers are unusual and neither explains the other.
Maine is near the bottom — does that mean anything?
Maine has 4.1 cameras per 100,000 people, third-lowest in the country. Maine is also, as far as we can establish, the only state that bans private ALPR use outright. That's suggestive and we'd love to claim it, but the confound is obvious: Maine is rural, sparsely populated and cold, and the low-density states clustered at the bottom of this table — Vermont, Montana, New Hampshire, Wyoming — mostly don't have that law. Rurality alone predicts the result. The law may well contribute; this dataset can't show that it does.
How does 116,655 compare with Flock’s own camera count?
They count different things. Flock does not publish an official count of its own cameras; in July 2026 it was reported as saying more than 120,000. This page counts every plate reader of any brand tagged in OpenStreetMap: 116,655 on 17 July 2026. That OpenStreetMap count has kept growing. RoadSigns.com counted 135,844 mapped locations of all brands on 26 August 2026, and Jalopnik reported DeFlock mapping 143,356 on 4 October. Some of that growth is likely more mapping, not only more cameras. Either way it demonstrates the point of publishing the method rather than the number — you can re-run the query today and see what it says now.
Why publish this when DeFlock already has a map?
Because a map answers 'where' and a table answers 'how much', and nobody was publishing the second one nationally. DeFlock does excellent work and renders the same underlying tag — this is not a competitor to it, it's an aggregation of the same public data into a form you can cite in a council meeting, drop into a spreadsheet, or check a claim against. Eyes Off Indiana does something similar for one state and does it well. This is the fifty-state version, free, with the query attached so you never have to trust us.
Can I use this data?
Yes. The underlying OpenStreetMap data is licensed ODbL, which requires attribution to OpenStreetMap contributors. This aggregation is free to reuse — cite Max Intel and the collection date if it's useful to you, and re-run the query if the numbers matter, because they will have changed. If you're a journalist or a researcher and want the per-county breakdown or a different cut, the method here scales to any boundary OSM knows about, which is most of them.
Do the 2026 cancellations change the per-state camera counts?
Not yet, and not in this table. The counts are a dated 17 July 2026 OpenStreetMap snapshot and we have not re-pulled them. A cancellation such as Henrico County VA (over 120 cameras, October 2026) or El Paso TX (more than 100 cameras, September 2026) lowers a state's count only after the hardware is physically removed and a volunteer deletes the OSM node, which can take months; the Institute for Justice database counted 277 cancelled or non-renewed contracts as of 15 October 2026, so expect Virginia, Texas, Ohio and Wisconsin to drift down on a future re-run. Re-run the query on this page for today's figure.