Real-Time Sensor & Video Console

Summary. A cross-platform desktop application in C++17 and Qt6 that brings real-time data from a range of hardware sensors and video sources into one live dashboard. It builds on the open-source OpenCPN project and adds configurable sensor integrations. This describes the engineering only; I keep the company and project names private and leave the purpose out.

Sensor and video integration

The core of my work was integration, and I built most of these sensor connections myself: configurable, port-adjustable modules that connect to different sensors and feeds and bring their data in over serial (UART) and TCP/UDP. Across the system that covered GPS and GNSS receivers, AIS, other NMEA instruments, and live and recorded camera and video streams handled with FFmpeg. The same modules could be pointed at different ports and sources through configuration, without changing code.

Architecture

Data flows through a clear pipeline: acquisition from the serial or network port, parsing, then Qt signals and slots carry it to the live interface, to SQLite for storage, and to a logger. Video runs on background QThread workers so the interface stays responsive under real-time load, and a small TCP server streams it to multiple clients. A common integration layer means new sensor types and protocols slot in through configuration rather than special-casing each one, and the modules reconnect on their own when a sensor or feed drops, so the dashboard keeps running. To develop and test each part without the real hardware, I built standalone simulators that replay mock data.

Tooling

The application is built with CMake and runs on Linux and Windows. I worked across the codebase in C++, with Python for supporting utilities. Rather than starting from scratch, I integrated OpenCPN into the system and refactored the parts it needed.