TALAHow every number
is worked out
TALA turns the CCTV cameras an LGU already has into automatic traffic counts and road-wear estimates. This page explains, in plain language, how it counts vehicles, sorts them by type, follows where they go, and judges how worn each road is.
1The TALA apps
TALA has two parts that work together. All the video work happens on your own computer; the web only ever sees totals.
TALA desktop app
Windows 10 / 11The counting station. Install it on the computer that has your camera feeds or recorded videos. It watches the video, counts every vehicle, keeps the vehicle log, and makes PDF and Excel reports.
Web dashboard
talaai.ph · any browserWhere the whole office sees the results: every site on a map, road status, busiest hours, where vehicles go, and reports for one site or all of them. Works on phones too.
2Getting set up
- Get an account. Your TALA administrator creates it. There is no public sign-up.
- Download the desktop app. Sign in at talaai.ph and open Account; the installer is there. On first start TALA downloads its graphics-card pack once, so keep the internet on.
- Add a camera. Choose a recorded video or a live camera, give the place a name (for example Rizal Ave, Pulilan, Bulacan), and pin it on the map.
- Draw the counting lines. Draw one line across each street that meets at the junction and name it after that street.
- Start counting. TALA works through the video and sends the totals to the dashboard.
| Computer needed | Details |
|---|---|
| System | Windows 10 or 11, 64-bit |
| Graphics card | NVIDIA graphics card recommended: many times faster. Without one TALA still works, just slower, with the same results. |
| Internet | Needed on first start and to upload results. Counting itself works offline. |
3How vehicles are counted
TALA looks at the video picture by picture, finds every vehicle, follows each one as it moves, and counts it when it crosses a counting line.
Finding and following vehicles
An AI model trained on Philippine road footage finds the vehicles in each picture and decides what type each one is. TALA then follows the same vehicle from picture to picture, so one vehicle is counted once even though it shows up in hundreds of pictures.
Counting lines
When you add a camera you draw a line across each street. A vehicle is counted when its path crosses a line. Because TALA knows which line a vehicle crossed coming in and which one going out, it also knows where it came from and where it went.
Speed setting
TALA can look at every picture or skip some to finish sooner. Looking at every picture is the most accurate and the slowest; skipping a few is a good balance for most roads. On very busy roads, look at more pictures.
Stopping and carrying on
You can stop or close TALA at any time. Next time, each video carries on where it stopped: nothing is counted twice and nothing is skipped.
4Vehicle types
TALA sorts vehicles into 10 types by size and number of axles, following the US FHWA vehicle classes. Each type has a road-wear value (see Road wear).
| Class | Vehicle type | Examples | Wear per vehicle |
|---|---|---|---|
| 01 | Motorcycles | Motorcycles, tricycles | 0.0007 |
| 02 | Passenger cars | Cars, sedans, hatchbacks | 0.0007 |
| 03 | Four-tire single unit | Jeepneys, vans, pickups, SUVs | 0.0007 |
| 04 | Buses | City and provincial buses | 0.57 |
| 05 | Two-axle, six-tire truck | Light delivery trucks | 0.25 |
| 06 | Three-axle single unit | Dump trucks, larger rigid trucks | 0.61 |
| 07 | Four or more axle single unit | Heavy rigid trucks | 0.61 |
| 08 | Trailer, four axles or fewer | Small tractor-trailers | 0.5965 |
| 09 | Five-axle semi-trailer | Container trucks | 0.99 |
| 10 | Trailer, six or more axles | Heavy haulers | 0.69 |
5Where vehicles go
For every vehicle TALA records the street it came from and the street it left by, for example Rizal Ave → Quezon Ave. The dashboard draws these as bands from left (came from) to right (went to); a thicker band means more vehicles. The busiest movements are listed with their share of all traffic.
6Busiest hours
TALA reads the clock printed on the CCTV picture, so every vehicle is put in the real hour it passed, not the time the video was processed. From that it shows traffic for each hour of the day and the peak hour, the busiest 60 minutes.
Counts are also kept in 15-minute blocks. The peak hour factor compares the busiest 15 minutes with the whole peak hour: close to 1 means traffic is steady through the hour, a lower number means it comes in bursts.
7DPWH-format traffic counts
The Traffic Count report follows DPWH's 12-class system (DO 58 s. 2026, form QMSP-07-20), counted by direction for each hour.
| DPWH class | Vehicle | From TALA class |
|---|---|---|
| 1 & 13 | Motorcycle / tricycle | 01 |
| 2 | Passenger car | 02 |
| 3 & 4 | Utility (jeepney, van, pickup) | 03 |
| 5 & 6 | Bus (small and large) | 04 |
| 7 | Rigid truck, 2 axles | 05 |
| 8 | Rigid truck, 3 or more axles | 06, 07 |
| 9 | Truck semi-trailer, 3–4 axles | 08 |
| 10 | Truck semi-trailer, 5 or more axles | 09, 10 |
| 11, 12, 14 | Truck trailers and bicycles | Not counted |
Classes 1 & 13, 3 & 4 and 5 & 6 are shown together because the camera cannot reliably tell them apart.
ADT and AADT
ADT (average daily traffic) is given only from recordings that cover a full day, as the DPWH guideline asks. AADT (the yearly average) is ADT multiplied by the seasonal factor for that month, which you get from your DPWH district office; TALA does not guess it.
8Road wear and road status
Heavy vehicles wear a road out far faster than light ones. TALA adds up the wear from every counted vehicle and compares it with how much the road was designed to carry.
How wear is measured
Wear is measured in ESAL (equivalent standard axle loads): one ESAL is the wear caused by one standard 80 kN axle. A motorcycle adds about 0.0007; a loaded container truck close to 1. So one truck can wear the road as much as over a thousand motorcycles. The values per vehicle type are in the vehicle types table.
Road life used
Each site's total wear is compared with the DPWH design standard of 2,000,000 ESAL. The share used decides the road's status:
| Status | Road life used | What it means |
|---|---|---|
| Good | under 30% | Plenty of life left. |
| Watch closely | 30% to 75% | Plan an inspection and keep an eye on heavy traffic. |
| Needs repair soon | 75% or more | Close to its design limit. Put it near the top of the repair list. |
Road life forecast
The forecast estimates how many years are left before a road reaches its design limit if traffic carries on as recorded. You set how many hours a day traffic looks like the recordings (fewer if they were taken at busy times) and how fast traffic grows each year; the answer updates as you move the sliders.
9Vehicle log and privacy
The vehicle log lists every vehicle counted: a picture, the time in the video, the clock time on the CCTV picture, its type, and the streets it came from and went to. Use it to check the counts vehicle by vehicle; See in video shows the full frame with the vehicle boxed.
- Faces and plate numbers are blurred automatically before a picture is saved. The CCTV time stamp is never blurred, so rows can be checked against the footage.
- Pictures can be deleted after a set number of days. The counts stay.
- The log stays on your computer. It is never uploaded to the dashboard.
More in the Privacy Policy.
10Sending results to the dashboard
The desktop app sends each site's totals to the dashboard: vehicles by type, direction and hour, the recording date and length, and the road wear. Results are always saved on the computer first, so nothing is lost when the internet is down; TALA sends them once it is back.
Sending the same results again never counts anything twice. To remove a recording that should not count, remove it on the site's page in the dashboard; the site's totals are corrected at the same time.
11Reports
Reports can be made for one site or all sites, for any range of dates, as PDF or Excel.
| Report | What it contains |
|---|---|
| Traffic and pavement load | The full picture: traffic, vehicle mix, busiest hours, road wear and status |
| Traffic volume | Vehicles by type and by hour |
| Road wear and design life | Wear so far, road life used and years left |
| Where vehicles go | Every from-street → to-street movement |
| Traffic count (DPWH format) | DPWH classes by direction and hour, peak hour and ADT |
| Before / after | Two recordings compared, for example before and after a road change |
| Recording log | Every recording counted at each site |
12Accuracy and limits
TALA's numbers are automatic estimates. They are most accurate on clear daytime footage from a camera looking down on the road. They are less accurate:
- At night, and in rain, fog or strong glare: some vehicles are missed.
- When vehicles are far away or very small in the picture.
- When vehicles are hidden behind each other in heavy traffic.
- For vehicles that look alike, which is why some DPWH classes are shown together.
Check important counts against the vehicle log, and have an engineer review results before official use.