Development of Application Featuring Accumulation of Roadside Data Through Mobile Mapping Hardware and Sensor Data
Industry:
Automotive
The Problem Statement
The Guardian Angel data collection desktop application interfaces with mobile mapping hardware and sensor arrays to acquire, process, and save data sets for future use by machine learning platforms and geospatial workflows, focusing on capturing roadside asset information for condition assessment tools. It serves as the interface for collecting and processing data from an Ouster LiDAR sensor, a Movella XSENS 670G GNSS/IMU, and three Lucid Vision Labs industrial cameras.
THE SOLUTION
The application features a GUI interface showcasing live streams from the Ouster LiDAR sensor and three Lucid Vision Labs cameras, alongside current GPS information (timestamp, latitude, longitude, elevation), and allows sensor configuration adjustments like framerate and target brightness. Communication with the LiDAR sensor over Ethernet utilizes the Ouster SDK, while communication with the GPS sensor over Serial/COM port employs the Movella MT SDK, and communication with the Lucid Vision Industrial cameras over Serial/COM port utilizes the Lucid Vision Labs SDK.
In addition to displaying live feeds, the application supports processing streams and logging data in various formats: GPS/GNSS data is parsed and stored in RINEX and shapefile formats concurrently, LiDAR packets are stored in PCAP format with metadata in JSON, and further processing of Lucid Vision Industrial Camera streams includes applying a debayering filter to convert BayerRG8 images into compressed JPGs for storage. Displaying stream data and file logging operations are executed in a multi-threaded environment to maximize resource utilization.
Tech Stack Implemented:
- C++
- Qt
- Ouster SDK
- Movella MT SDK
- Lucid Vision Labs SDK
- GStreamer