Back to portfolio

TideEye: Underwater Visibility

Computer vision applied to underwater visibility at the Scripps pier.

TideEye: Underwater Visibility

The problem

A live underwater camera shows conditions, but a useful visibility estimate requires relating the image to known distances and ocean conditions.

What I built

I built a computer vision pipeline that fits a Beer–Lambert attenuation model across known-distance pier pylons. A serverless AWS backend feeds a React frontend alongside CDIP buoy and Open-Meteo data.

The result

TideEye presents camera-derived visibility estimates alongside oceanographic context. The estimate is model-based and depends on what can be measured in the camera image.