SafeEdge App
The Guardian Angel desktop application for data collection connects with advanced mobile mapping equipment and sensor arrays to gather, process, and store datasets for subsequent use by machine learning systems and geospatial processes. It is specifically tailored to collect information about roadside assets to aid in condition assessment. The system functions as a powerful interface for managing data from an Ouster LiDAR sensor, a Movella XSENS 670G GNSS/IMU, and three industrial cameras from Lucid Vision Labs, all operating at 100 FPS on an edge device powered by an RTX 2000 Ada. Features of the application include real-time sensor setup and control (pan, tilt, zoom), as well as the synchronization of data streams from various sensors.
Tech Stack: C++ Qt Framework Ouster SDK Movella MT SDK Lucid Vision Labs SDK GStreamer Nvidia RTX 2000 Ada
Results: The Guardian Angel application exhibited outstanding performance by integrating multiple high-performance sensors into an intuitive interface. It effectively captured and processed real-time data from three 4K cameras at 100 FPS, along with data from the Ouster LiDAR sensor and the Movella XSENS GNSS/IMU, all while running smoothly on an RTX 2000 Ada edge device. The application provided synchronized data from all sensors, including pan, tilt, and zoom measurements from the LiDAR, facilitating accurate and prompt data collection. Its multi-threaded architecture enhanced resource use, enabling smooth live streaming with low latency during data acquisition and processing. This system delivered highly accurate roadside asset information, establishing a new benchmark for efficient, AI-driven geospatial operations.