LIVE CAMERA NETWORK
Camera Network Report — October 2026
WorldEye presents 74,229 live cameras. This monthly report counts the catalogue behind them — 76,575 records across 109 countries and 190 publishing sources — and publishes every figure with the dataset it came from, free to check or cite.
October 2026 edition · Published October 27, 2026 · First edition — the baseline for monthly editions that follow.
Key findings — October 2026
- One country holds the majority. The United States accounts for 41,355 catalogue records (54% of the world catalogue) — more than the other 108 countries combined (35,220).
- Japan is a strong second. 10,472 records (13.7% of the catalogue), 58.7% of all Asian records — and a single Japanese river network is the largest publisher in the catalogue (below).
- The top ten hold almost everything. The three largest countries hold 73.8% of records and the top ten hold 90.8%. Only 11 countries have 1,000+ cameras; the median country has just 8.
- The long tail is very long. 57 of 109 countries have fewer than 10 cameras — 168 records combined. Presence, not depth: the network touches those countries without yet covering them.
- Florida leads the United States. 4,935 records assigned to Florida by coordinates (11.9% of the US catalogue), ahead of Texas (4,390) and Georgia (4,288).
- Houston is the largest city cluster. 1,174 records carry a Houston label; London is second with 887. City labels are publisher feed labels — the method and its caveats are set out below.
Three numbers, three measures
WorldEye publishes three totals, because they measure three different things, and this report never mixes them. Live cameras — 74,229 — is the network's headline count: the cameras the globe presents. Catalogue records — 76,575 — is the size of the geo-attributed dataset behind the country pages, and it is the population every table in this report measures. The catalogue is the larger number because it retains records whose feeds were offline, moved, or between refreshes when the live count was taken.
The third measure is the served snapshot behind the Live Camera Index: Unique cameras served by the production catalogue API in the 23 October 2026 snapshot (84,455 records deduplicated by camera ID) — the Live Camera Index's served measure, wider than the globe headline. It runs wider than the directory — 113 countries against the directory's 109 — because the live API assembly includes cameras not yet attributed into the committed catalogue. No table in this report uses it; it is defined here so the three measures are never confused for one another.
| Measure | Count |
|---|---|
| Live cameras (network headline count) | 74,229 |
| Unique cameras served by the production API (Live Camera Index snapshot, 23 October 2026) | 83,795 |
| Catalogue records attributed to a country (this report's population) | 76,575 |
| Indexed catalogue records (before attribution) | 77,376 |
| Near-duplicate records dropped in the catalogue build | 798 |
| Records with no attributable country | 3 |
| Distinct publishing sources in the catalogue | 190 |
Country coverage — the top 15 of 109
The United States and Japan together hold 67.7% of the catalogue. The table shows the 15 largest of 109 countries; the full country table is in the downloadable dataset, and every country has its own camera page.
| Country | Records | Share of catalogue |
|---|---|---|
| United States | 41,355 | 54% |
| Japan | 10,472 | 13.7% |
| Canada | 4,685 | 6.1% |
| Taiwan | 3,243 | 4.2% |
| Spain | 2,120 | 2.8% |
| Austria | 1,725 | 2.3% |
| South Korea | 1,610 | 2.1% |
| Sweden | 1,569 | 2% |
| United Kingdom | 1,403 | 1.8% |
| Hong Kong | 1,350 | 1.8% |
| Italy | 1,030 | 1.3% |
| Finland | 825 | 1.1% |
| Norway | 746 | 1% |
| Iceland | 482 | 0.6% |
| Australia | 421 | 0.5% |
The long tail — 57 countries with fewer than 10 cameras (168 records combined): Singapore (9), Costa Rica (8), Ecuador (8), Georgia (7), Slovenia (7), Ukraine (7), Hungary (6), Moldova (6), Cyprus (5), Dominican Republic (5), Kenya (5), Nepal (5), Puerto Rico (5), Tanzania (5), India (4), Seychelles (4), Sint Maarten (4), Albania (3), Belarus (3), Belgium (3), Belize (3), Bosnia and Herzegovina (3), Czechia (3), El Salvador (3), Panama (3), Azerbaijan (2), Bermuda (2), Bolivia (2), Denmark (2), Egypt (2), Grenada (2), Guadeloupe (2), Honduras (2), North Macedonia (2), Serbia (2), Slovakia (2), Yemen (2), Armenia (1), Barbados (1), Bonaire (1), Christmas Island (1), Curaçao (1), Faroe Islands (1), Iran (1), Jordan (1), Lebanon (1), Madagascar (1), Martinique (1), Morocco (1), Oman (1), San Marino (1), Saudi Arabia (1), Senegal (1), Sri Lanka (1), Uruguay (1), Vatican City (1), Venezuela (1).
The regional picture
North America dominates on the strength of the US state DOT networks, and Asia is second almost entirely because of Japan and Taiwan. Europe is the broadest region — 42 countries — but thinner per country. Latin America, the Caribbean and Africa together hold 209 records across 32 countries: the network's growth frontier.
| Region | Records | Countries | Share |
|---|---|---|---|
| North America | 46,042 | 3 | 60.1% |
| Asia | 17,833 | 29 | 23.3% |
| Europe | 11,812 | 42 | 15.4% |
| Oceania | 679 | 3 | 0.9% |
| Latin America & the Caribbean | 178 | 24 | 0.2% |
| Africa | 31 | 8 | <0.1% |
Regions are this report's continental grouping of the catalogue's countries; country figures are unchanged by the grouping.
Inside the United States
US records are assigned to states by camera coordinates (41,312 of 41,355 assigned; 43 could not be placed). That makes this a geographic table, not a publisher table — a camera near a state line counts where it stands, whoever publishes it. State assignment generated October 26, 2026.
| State | Records | Share of US catalogue |
|---|---|---|
| Florida | 4,935 | 11.9% |
| Texas | 4,390 | 10.6% |
| Georgia | 4,288 | 10.4% |
| California | 3,879 | 9.4% |
| Utah | 2,135 | 5.2% |
| Virginia | 1,451 | 3.5% |
| Washington | 1,447 | 3.5% |
| Pennsylvania | 1,348 | 3.3% |
| New York | 1,337 | 3.2% |
| Minnesota | 1,282 | 3.1% |
Who publishes these cameras?
Public agencies, overwhelmingly — national river and road authorities, state transport departments and the FAA's weather-camera programme. The largest single publishing source in the catalogue is Japan's river.go.jp network (9,890 records, 12.9%). One aggregation pipeline recurs throughout: collections published as “OpenCCTV / …” contribute 34,040 records (44.5% of the catalogue) between them — the rows below are the individual networks as the catalogue names them.
| Publishing source | Records | Share |
|---|---|---|
| OpenCCTV / rivergojp | 9,890 | 12.9% |
| FDOT | 4,908 | 6.4% |
| TxDOT | 4,349 | 5.7% |
| GDOT | 4,282 | 5.6% |
| FAA WeatherCams | 3,528 | 4.6% |
| OpenCCTV / state511 | 3,358 | 4.4% |
| Caltrans | 2,936 | 3.8% |
| THB Highway Bureau | 2,200 | 2.9% |
| UDOT | 2,084 | 2.7% |
| OpenCCTV / castlerock | 2,079 | 2.7% |
| DGT | 1,932 | 2.5% |
| OpenCCTV / arcgis_cams | 1,871 | 2.4% |
| ASFINAG | 1,695 | 2.2% |
| OpenCCTV / virginia | 1,480 | 1.9% |
| OpenCCTV / ny511 | 1,352 | 1.8% |
The largest city clusters
City figures count the catalogue's own place labels. They are publisher labels, not gazetteer boundaries — and 13,424 records are filed under a state, province or country name instead of a city (US DOT feeds do this routinely), so those records are counted in the state and country tables, not here.
| Place label | Country | Records |
|---|---|---|
| Houston | United States | 1,174 |
| London | United Kingdom | 887 |
| Dallas | United States | 843 |
| Orange County | United States | 585 |
| Hillsborough County | United States | 503 |
| Seattle | United States | 499 |
| Fort Worth | United States | 489 |
| Broward County | United States | 477 |
| Ottawa | Canada | 444 |
| Phoenix | United States | 429 |
| Miami-Dade County | United States | 421 |
| Toronto | Canada | 413 |
| Madrid | Spain | 370 |
| Stockholm | Sweden | 350 |
| Busan | South Korea | 321 |
What the feeds look like
The catalogue classifies cameras by publisher and feed technology rather than by subject. 21,423 records (28% of the catalogue) declare a feed format; HLS video is the largest, ahead of refreshing JPEG stills. Counting access paths instead: 63,599 records expose a direct feed URL, 13,235 a stream URL and 7,431 a link to the publisher's own page — many records carry more than one.
| Declared feed format | Records | Share of typed records |
|---|---|---|
| HLS video streams | 11,121 | 51.9% |
| JPEG still images (refreshing) | 8,189 | 38.2% |
| MP4 streams | 1,658 | 7.7% |
| Embedded publisher players (iframe) | 273 | 1.3% |
| MJPEG streams | 182 | 0.8% |
Edge of the map
The northernmost record in the catalogue is Arctic Bay — Camera 1 (SouthEast) in Arctic Bay, Canada, at 73.01°N. The southernmost is a research camera at 54.94°S near Puerto Williams, at the southern tip of South America — filed by its publisher under a neighbouring country's label, a reminder (counted in the limitations below) that country labels originate with publishers.
Methodology — and what this edition does not claim
Population. Every figure in this report was computed by a kept analysis script from the committed catalogue files: one file per country (109 files, 76,575 records), the catalogue's attribution build (generated October 20, 2026: 77,376 indexed records, 798 near-duplicates dropped, 3 unattributed) and the US state assignment (generated October 26, 2026). Records were deduplicated by camera ID across files — the script found 0 duplicate IDs and 0 records with invalid coordinates. The full methodology — indexing, deduplication, feed checks, geographic coordinates and how outdated cameras are removed — is documented on the camera-data methodology page.
No availability percentage. The catalogue carries no per-camera status or last-checked field, so this report publishes no uptime figure. A record was present when the catalogue was assembled; no edition can guarantee a given feed is streaming at the moment you read it. The live headline count is the network's own presented figure.
No “new this month” figure — yet. The catalogue carries no per-camera added date either. This first edition is the baseline: the snapshot behind it is kept, and the November edition will report the measured change in records, countries and sources against it, rather than estimating additions.
Limitations
- This report measures the cameras WorldEye's loaders collect, not every public camera on Earth. Cameras no loader reaches are not counted, in either direction.
- Country and city labels originate with publishers and their aggregators; normalisation corrects codes and known variants, but a mislabelled record can misplace a camera.
- City labels are feed labels, not boundaries: county and metro labels appear alongside city names, and feeds filed under a state or province are counted at that level.
- Companion figures on the Live Camera Index come from its 23 October 2026 snapshot of the production catalogue API — the third measure defined above. That assembly is wider than the committed directory this report counts, so the two instruments are not expected to agree exactly.
Download the dataset — and cite this report
The dataset is the exact file this page renders from — totals, all 109 countries, regions, US states, publishing sources, city clusters and feed formats, generated by the same script from the same catalogue files.
WorldEye (2026). Camera Network Report — October 2026. WorldEye. https://worldeye.app/reports/camera-network-2026-10
Counts may be cited and republished with attribution to WorldEye and a link to this page. The camera feeds themselves remain the property of their publishers.