Vision OPS Monitoring System

Render Picture outside with LED's
Render on Cozy Desktop
Render Picture Outside
55 degree FOV

(WATCH ON YOUTUBE) ⬆️⬆️⬆️

Vision OPS began with a simple problem: fabrication machines often run for hours without anyone watching them. When a 3D print fails, a CNC machine encounters a problem, or a laser cutter becomes unsafe, the user may not notice until time, material, or equipment has already been lost. A normal camera can record the failure, but it cannot understand what is happening or help the user respond. We wanted to turn that passive camera feed into an intelligent monitoring system.

We started by building a local application that connects to a standard camera and analyzes live frames using a custom-trained AI model. Instead of relying on generic motion detection, Vision OPS is designed to recognize fabrication-specific problems, such as spaghetti failures and detached prints. The system evaluates the model’s confidence across multiple consecutive frames before confirming an incident, which helps reduce false alarms caused by a single unusual image.

Once a failure is confirmed, Vision OPS turns detection into action. It automatically changes the machine status, captures visual evidence, and generates a clear incident report containing the failure type, confidence level, probable cause, and recommended next step. We also created our own Discord webhook so critical alerts can reach the user immediately, even when they are away from the machine or not watching the dashboard.

The result is more than a monitoring camera. Vision OPS creates a complete process from observation, to detection, to confirmation, to reporting, and finally to real-time notification. Our goal is to help makerspaces, schools, labs, and fabrication facilities reduce wasted material, avoid unnecessary downtime, improve safety, and learn from previous failures instead of simply recording them.