SafeEdge: Multi-Sensor Mobile Mapping Platform
Desktop application for real-time LiDAR, GPS/IMU, and triple 4K camera data capture for roadside asset condition assessment.
The Challenge
Guardian Angel Technologies (USA) needed a robust desktop application to interface with advanced mobile mapping hardware — an Ouster LiDAR, Movella XSENS GNSS/IMU unit, and three 4K industrial cameras — for real-time roadside asset condition assessment. The system had to synchronize all sensors, stream live data, and log in multiple formats simultaneously.
Our Approach
We built a high-performance C++ desktop application using the Qt Framework, integrating three separate sensor SDKs into a unified data capture pipeline.
Multi-Camera Processing
Three Lucid Vision Labs 4K cameras capture at 100 FPS each. We implemented a custom GStreamer pipeline for hardware-accelerated encoding on an NVIDIA RTX 2000 Ada GPU, handling the massive data throughput without frame drops.
LiDAR & IMU Integration
The Ouster SDK provides real-time 3D point cloud data, while the Movella MT SDK delivers high-precision GNSS/IMU positioning. All sensor data streams are precisely time-synchronized for accurate geospatial alignment.
Data Logging
Multi-format recording supports simultaneous capture of point clouds, video streams, and positional data with sub-millisecond synchronization, enabling downstream AI-driven analysis workflows.
Technical Stack
- Languages: C++
- Framework: Qt
- SDKs: Ouster SDK, Movella MT SDK, Lucid Vision Labs SDK
- Video: GStreamer with NVIDIA GPU acceleration
- GPU: NVIDIA RTX 2000 Ada
Results
- Real-time capture from three 4K cameras at 100 FPS
- Synchronized multi-sensor data with minimal latency
- Precise roadside asset condition data
- Foundation for AI-driven geospatial analysis workflows