Goal: detection of dogs crossing certain area on 2 specific spots (start/finish of 2 Flyball lanes)
Preferred solution: "budget" / robust / set and forget / weatherproof (drizzle rain conditions) solution
Current Solution: Vision Cam (Android Phone running iVCAM), transmitting to (local) host through WiFi (5Ghz)
Results vary (depending on running processes as all processing is done on a "budget" notebook) between 0.3 and 2 seconds.
Notebook (Windows 11) running Minimal Api .net 8 app, for visuals on the field (web interface for key users to start runs, reset etc) (can also be used as power supply, wireless / self contained preferred for the sensor modules)
Current knowledge: Mainly high level languages, e.g. C# Full Stack Development
Current Budget: 100 to 200 euro (for the sensing devices)
Current Experience:
IR route (MK120), alignment issues (sender / receiver), to fragile
hardware issue, 5v to (local) webapi
LiDAR route (D200), measurement to slow (turning speed of the sensor)
works in low light conditions, trouble with rain
Vision route, processor to slow (> 300ms sec to 2000ms)
opportunities, direction detection, dog or human detection etc
ended up with ArUco markers with diff check on predetermined surface for "fast" processing.
My current idea is to try the Arduino route, as in 2x Arduino UNO R4 Wifi with multiple GP2-0215K distance sensors. And to determine the High Low state and only send the flip switch state from the Arduino to the API for further processing.
From my understanding PIR sensors are typically too slow to register movement and the professional systems work with a "Safety" Light Curtain setup.
Furthermore the actual time needs to be synced between the lanes, <100ms will be sufficient and though an api call after booting will be enough.
Are there any ideas or suggestions that can help?