[MathewTech254]

What this does

Turns the detect.py stub into an actual working detector. It pulls frames from your webcam (or a video file), runs them through YOLO, draws the boxes, and puts an FPS counter and a helmet/no-helmet tally in the corner.

pip install -r requirements.txt
python detect.py                                       # webcam
python detect.py --model helmet-yolov8n.pt --conf 0.5   # custom weights
python detect.py --source campus-footage.mp4           # video file

Press q to quit.

Closes #2

Things I ran into

The default yolov8n.pt is a COCO model and doesn't have a helmet class, so out of the box it will always say NO DETECTIONS no matter what you point it at. That's expected — you need weights trained on helmet classes, which you pass with --model. I mentioned it in the --help text so nobody thinks it's broken.

Class names came back in more shapes than I expected, so classify_label normalises them first. Helmet, no-helmet, No_Helmet, helmetless and without helmet all end up in the right bucket. A plain "no" in label check would have worked too, but it also matches things like Nose helmet, which is not a thing, but you get the idea. Also worth knowing: Ultralytics gives you names as a dict on most models and a list on others, so both are handled.

One more: the first version stopped the whole loop the first time a frame failed to read. Webcams drop frames, so now it puts up with 10 in a row before deciding the stream is actually over.

Testing

No webcam here, so I ran the loop against generated video and real photos with the GUI calls stubbed out.

• python -m py_compile detect.py clean, --help shows all three flags, yolov8n.pt loads
• 40-frame video end to end: every frame processed, correct window title, valid BGR frames, HUD drawn, q quits early, end of stream exits clean
• Bad video path returns exit 1 without drawing anything
• Real detections come through with boxes, labels and confidence
• Fed it helmet-classed boxes to check the counters and the COMPLIANT / VIOLATION banner actually flip

Note

Only detect.py changed. requirements.txt and README.md untouched.

[github-actions]

🎉 Thank you for your Pull Request, @MathewTech254!

Welcome to AIML Club OCT! 🚀

A maintainer will review your code shortly. In the meantime:

• 📋 Please ensure your branch follows our Contributing Guide.
• 💬 Showcase your work or ask questions in our Student Project Showcase.
• 📌 Track ongoing milestones on our Roadmap Board.

Happy coding! 💻✨

[UmeshCode1]

Hi @MathewTech254! 👋

Thank you so much for this PR! The real-time webcam loop, FPS calculation, normalized multi-label handling (\helmetless, \No_Helmet, etc.), and 10-frame drop tolerance are fantastic additions to make the detector actually deployable in live environments.

There is currently a small git conflict with the latest \main\ branch in \�eginner/safety-helmet-detector/detect.py\ (due to recent fallback updates).

Could you please sync your fork with upstream \main\ and rebase / resolve the conflict?
\\�ash
git fetch origin main
git rebase origin/main

resolve conflicts in detect.py, then:

git push --force-with-lease
\\

As soon as that's rebased, we are ready to approve and merge this! Appreciate your awesome work! 🚀

[MathewTech254]

Am on it. Will push when done. Thank you for considering my work.

…

On Fri, Oct 23, 2026, 12:44 PM Umesh Patel ***@***.***> wrote: *UmeshCode1* left a comment (AIMLCLUBOCT/Projects#6) <#6 (comment)> Hi @MathewTech254 <https://github.com/MathewTech254>! 👋 Thank you so much for this PR! The real-time webcam loop, FPS calculation, normalized multi-label handling (\helmetless, \No_Helmet, etc.), and 10-frame drop tolerance are fantastic additions to make the detector actually deployable in live environments. There is currently a small git conflict with the latest \main\ branch in \�eginner/safety-helmet-detector/detect.py\ (due to recent fallback updates). Could you please sync your fork with upstream \main\ and rebase / resolve the conflict? \\�ash git fetch origin main git rebase origin/main resolve conflicts in detect.py, then: git push --force-with-lease \\ As soon as that's rebased, we are ready to approve and merge this! Appreciate your awesome work! 🚀 — Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=CM4VKUMSVH5Q6EXZJHVUIA35VMSBVA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMMZWGIYDQOBQHA2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-6362088084>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/CM4VKUOWLG5DD4HSHPG6DFD5VMSBVAVCNFSNUABGKJSXA33TNF2G64TZHMYTAMJRG42DKMBQGY5US43TOVSTWNJWGU4TKNRZHE3DTILWAI> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/CM4VKULYYI52ZNOH3O3RGYL5VMSBVA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMMZWGIYDQOBQHA2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJKTGN5XXIZLSL5UW64Y> and Android <https://github.com/notifications/mobile/android/CM4VKUJ4YUM37NCXB5OIDNT5VMSBVA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMMZWGIYDQOBQHA2KM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLTGN5XXIZLSL5QW4ZDSN5UWI>. Download it today! You are receiving this because you were mentioned.Message ID: ***@***.***>