Smart Human Detection System Development

I need a reliable human-detection system built from the ground up. The core requirement is straightforward: accurately recognise when a person enters the monitored zone and trigger a software event that I can later tie into alerts, recording, or access-control logic.

Deployment conditions are still flexible—I may end up installing the solution indoors, outdoors, or both—so the code and hardware choices must tolerate varied lighting and weather without extensive retuning. I am equally open to camera-based visual recognition (e.g., OpenCV, YOLO, TensorFlow), infrared arrays, or motion-sensor fusion; the deciding factor will be detection accuracy and ease of maintenance rather than any single technology.

What you will deliver
• A working prototype—hardware schematics (if any), source code, and a clear build/run guide.
• API or messaging hook that fires whenever a human target is confirmed.
• A short report summarising test methodology and measured accuracy (precision/recall) on at least two contrasting scenes.

Clean, well-commented code, no proprietary lock-ins, and instructions that let me redeploy on a different machine or microcontroller will complete the job.