Manual Β· about one afternoon

Measure the road outside your school.

Seven simple steps. You film the gate once, the tool measures every near-miss, and you get a report and a letter you can send to the road authority.

πŸ“±A phone that films
πŸ“A tape measure
🟨8 bright markers: A4 paper, chalk or cloth
πŸ’»A laptop with Python 3.11+ and ffmpeg
πŸ”‘A Meta Model API key
1

Ask first

The tool blurs every tracked face and pixelates number plates in the clips, but always ask before you film.
2

Mark the road

Diagram: what to measure with the tape at the gate
The plan seen from above: the 4 corner markers (red, N1 N2 F1 F2) and the 15 m check line (blue, A to B). Click to open it full size.

πŸ“ Open the field sheet

Print this sheet and write each number on it while you measure.

Only put markers on the road when it is safe, with someone watching the traffic.
3

Film from high up

Do not zoom or move during the video. A steady phone gives better measurements.
4

Mark the road in the app

ffmpeg -ss 5 -i YOUR-VIDEO.mp4 -frames:v 1 frame.jpg
Want to practise first? Open App, step 1 and press Try it with our gate: it loads our frame and our 4 real corners.
5

Run it

git clone https://github.com/shaunmortal/crossing-watch
cd crossing-watch
echo "MODEL_API_KEY=your-key-here" > .env
python3 test_cw.py        # checks that everything works, under a minute
python3 monday.py YOUR-VIDEO.mp4 site.json
python3 monday.py YOUR-VIDEO.mp4 site.json --budget 14 --yes
About $12 for an 8-minute video. It never spends more than your --budget.
6

Answer the agent's question

python3 agent.py YOUR-VIDEO.mp4 site.json --out out/YOUR-VIDEO-agent --answer R1 "15 m"
If the video and the tape agree, the agent writes the letter. Our gate: video 15.0 m, tape 15 m, 0% off.
7

Read the report and send it

python3 letter_pdf.py out/YOUR-VIDEO site.json
See what a finished report looks like: our gate's report.
Questions

Do I need to know coding?

No. You copy and paste the commands above.

Is it safe to share the clips?

Yes. Before it makes the clips, the tool blurs every tracked face and pixelates every number plate: SAM 3.1 finds the plates at full resolution, so cars and bikes still look like cars and bikes. A close vehicle whose plate SAM cannot find is pixelated whole. The clips keep the road's own sound, so you can hear the horn; Muse Voice Transcribe finds where anyone speaks, and those moments are silenced.

What is a near-miss?

A person and a vehicle reach the same spot within a few seconds. Under 1.5 s is serious; under 3 s is a close pass.

Can I try it without a key?

Yes: python3 pack.py sample sample/site.json makes a full report from made-up data, and the Explorer replays our real run in your browser.