Querying Multimodal Robotics Data: A Gentle Introduction to Rerun (Part 5) скачать в хорошем качестве

Querying Multimodal Robotics Data: A Gentle Introduction to Rerun (Part 5)

Learn how to query and align multimodal sensor data from RRD and MCAP files using Rerun's powerful DataFusion-based query interface. This tutorial shows you how to extract time-series data from different sensor streams—IMU, GPS, and obstacle detection—and synchronize them programmatically for machine learning pipelines. What You'll Build: In this hands-on walkthrough, you'll learn to set up a Rerun server, connect via Jupyter notebooks, and execute SQL-like queries to pull forward the latest sensor values across misaligned timestamps. You'll see how to convert query results to Pandas DataFrames, unnest nested columns, filter by timeline indices, and prepare synchronized sensor data for ML model ingestion. This is essential for robotics developers working with asynchronous sensor data that needs temporal alignment. Code & Resources: GitHub Repository: https://github.com/rerun-io/rerun Getting Started Guide: https://rerun.io/docs/getting-started... Example Projects: https://rerun.io/examples Prerequisites: Part 4 of the Gentle Introduction series (generates the required RRD data file) Technical Stack: This tutorial uses Rerun's catalog client with Apache DataFusion for query execution, demonstrating compatibility with MCAP files, Jupyter notebooks, and Pandas workflows. The querying interface works seamlessly with ROS/ROS2 data pipelines and supports timeline-based data alignment for IMU, GPS, and sensor fusion applications. Try It Yourself: Clone the repo and run this example with your own robotics datasets. Star the repository if you find this useful for your computer vision or Physical AI projects. Integration Details: Compatible with: ROS2, MCAP, Apache DataFusion, Pandas Data formats: RRD files, MCAP recordings APIs covered: Catalog client, timeline indexing, DataFusion query interface Keywords: Rerun, robotics visualization, multimodal data, computer vision, Physical AI, sensor fusion, time-series data, IMU data, GPS telemetry, Apache DataFusion, ROS2, MCAP, Jupyter notebooks, machine learning pipelines Connect with Rerun: 🌐 Website: https://rerun.io 💬 Discord:   / discord   🐦 Twitter: https://x.com/rerundotio 💼 LinkedIn:   / rerun-io   Code: https://gist.github.com/bllchmbrs/057...