Help map the cameras

10 min readContributor guide

Useful camera records start with careful observations and clear sources. This guide explains how to document a device in OpenStreetMap so other people can check and use what you found.

How the data works

FlockDetour maintains a camera database assembled from OpenStreetMap, government publications and moderated community reports. OpenStreetMap is its largest camera-location source. An edit you publish there remains part of that shared dataset, available to other projects under its license. FlockDetour imports updates on its own schedule, so an OSM edit does not appear here immediately. The sources page explains the feeds, attribution and import dates.

OpenStreetMap asks that anything in it be verifiable — that another person standing where you stood would record the same thing. Everything below is downstream of that one rule.

Submitting a report to FlockDetour does not publish an edit to OpenStreetMap. Reports go through a separate moderation process; direct OpenStreetMap contributions from FlockDetour are not currently enabled. To make a public OSM edit, use the OSM editor or a supported mapping app described below.

Before you go anywhere

  • Stay on public ground

    Photographing something plainly visible from a public street is a constitutional right in the US, and the ACLU is unambiguous about that. It is equally unambiguous that the right to photograph is not a right to break other laws — trespass to get a better angle and you can still be charged with trespass.

  • Never touch the equipment

    Observe, note, photograph, leave. Interfering with, obstructing or damaging a device is a crime, it is not what this project is for, and it discredits everyone doing the careful version of this work.

  • Mind the traffic more than the camera

    Most of these sit at busy intersections and on-ramps. Pull over somewhere legal and safe first. Do not stand in a shoulder or read your phone in a live lane to get a bearing.

  • Skip people's homes

    This project documents government and commercial surveillance aimed at public roads. A neighbor's doorbell camera pointed at their own driveway is not that. The OSM community is genuinely split on residential cameras; staying out of the argument by only mapping public-facing infrastructure is both easier and kinder.

What to record while you're there

Six things. The first two are what separate a useful node from a dot on a map, and the rest are cheap to note now and impossible to recover once you have driven away.

  • The exact position

    Which pole, not which block. A node fifteen metres off can put a camera on the wrong side of an intersection and change every route calculated around it.

  • The direction it faces

    The single most valuable field, and the one most often left blank. A reader aimed at inbound traffic and one aimed at outbound are different facts about your street.

  • What kind of device it is

    A plate reader, a general-purpose camera, or something else. If you are not sure, record it as a camera rather than guessing at a plate reader.

  • The manufacturer, if you can read it

    A decal or a housing you recognize. If you cannot tell, leave it out — an unlabeled node is far better than a wrong attribution.

  • The mount and the power

    Own pole or existing structure; solar panel or cable to the ground. Cheap to note while you are there and impossible to recover later.

  • A photo, for your own verification

    Not required by the database, but it lets you check your own tags afterwards and settle a disagreement if someone questions the node.

Not sure what you are looking at? The identification guide covers the difference between a reader, an enforcement camera and a gunshot sensor.

The tags

OpenStreetMap describes things with key-value tags. These are the ones that matter for surveillance equipment. Worth knowing which is which: the base tag was approved by community vote in 2008, while several of the detail keys come from a draft proposal that was never formally ratified. They are in wide use and worth filling in — they just do not carry the same standing, and no guide should imply otherwise.

  • man_made=surveillanceRequired · Approved by community vote, 2008

    Declares the node a surveillance device. Nothing else works without it.

  • surveillance:typeRequired · De facto standard

    What is watching: camera, ALPR, guard, gunshot_detector. Use ALPR for a plate reader — the wiki lists camera:type=ALPR as an incorrect alternative to be corrected.

  • surveillanceOptional · De facto standard

    The kind of area watched: public for a street or plaza, outdoor for a lot or entrance, indoor for inside a building.

  • camera:directionOptional · In use

    Degrees clockwise from north, or a compass abbreviation. A hyphenated range describes an arc a moving camera sweeps.

  • camera:typeOptional · In use

    fixed for a device that cannot move, panning for one that rotates, dome for a PTZ under a dark cover.

  • camera:mountOptional · From an unratified draft

    pole, wall, street_lamp, traffic_signal, ceiling, building.

  • surveillance:zoneOptional · De facto standard

    The purpose of the area. traffic is the convention for roadside plate readers.

  • manufacturerOptional · Standard key

    The vendor name, spelled the way everyone else spells it — see the identification guide for the exact values.

  • operatorOptional · Standard key

    Who runs it, if you actually know — a police department, a homeowner association, a business. Do not guess.

A roadside plate reader, fully tagged
  • man_made=surveillance
  • surveillance:type=ALPR
  • surveillance=public
  • surveillance:zone=traffic
  • camera:type=fixed
  • camera:direction=<degrees>
  • manufacturer=<vendor>

How to submit

The OpenStreetMap web editor

Best if you are at a computer with notes and photos from a trip.

  1. Create a free OpenStreetMap account — everything below needs one, and it is the account that owns your edits.
  2. Find the exact spot on the map and zoom right in. Background imagery helps you place the point precisely; it is not a substitute for having seen the device.
  3. Choose the Point tool, click the location, then search the preset box for “surveillance”, “ALPR” or “CCTV” and pick Surveillance Camera.
  4. Fill in the direction using the compass control, then the remaining fields you noted in the field.
  5. Save, write a plain changeset comment saying what you added and that you surveyed it in person, and upload.

A mapping app on your phone

Best while you are actually standing there.

  1. Install a surveillance-mapping app that writes to OpenStreetMap — DeFlock publishes an open-source one for iOS and Android.
  2. Sign in with your OpenStreetMap account. The app uses OSM's own authorization flow, so it never handles your password.
  3. Drop a node, drag it onto the exact pole, and set the direction with the field-of-view control.
  4. Pick a device profile if the app offers one — it fills in the vendor tags correctly so you do not have to remember them.
  5. Submit. Uploads can take a while to appear; check your changesets on openstreetmap.org before submitting the same camera twice.

Rules and limits

Two things get contributions reverted more than anything else. The first is copying from a commercial map — Google Maps and Street View are copyrighted, and their terms forbid extracting content into another database. Your own eyes are the source; background imagery in an editor is only there to help you place a point you have already seen.

The second is mapping what you assumed rather than what you observed. A node you are unsure of is worse than no node, because somebody will route around a camera that is not there and trust a road that has one. If you are not certain, leave the field blank. Blank is honest; wrong is not.

Bad edits are handled by OpenStreetMap’s own community rather than by us — changesets are public, mistakes get discussed and reverted, and deliberate vandalism goes to a volunteer working group. Simple errors are not treated as vandalism, so a well-intentioned mistake is not something to be afraid of.

Other projects in this space

Other people map this too, and for some questions they are the better answer. Since everything above feeds a shared dataset, a camera you map improves all of these at once:

  • DeFlock

    A community ALPR map, an open-source mobile app for submitting cameras to OpenStreetMap, and information about local camera programs.

    deflock.org

  • Atlas of Surveillance

    EFF and the University of Nevada, Reno. Maps which agencies use which surveillance technology rather than where individual devices sit — better when your question is about a department instead of a street.

    atlasofsurveillance.org

  • Have I Been Flocked?

    Searches audit logs released through public records requests to show whether a plate was looked up. It answers a different question than a map does.

    haveibeenflocked.com/

  • OpenStreetMap

    A shared geographic dataset used by DeFlock and FlockDetour. Public edits can reach projects that import OSM data, with timing and filters that vary by project.

    www.openstreetmap.org