View on GitHub


Civil Maps Hardware Abstraction Layer




This series is about working with sensor data for autonomous vehicles and is based on Civil Maps’ real world experiences with these technologies. We started working with LiDAR data in 2013. We were providing 3D mapping for heavy industries such as railroads and pipeline projects. For the purpose of these sessions, we are focused on LiDAR, IMU, and sensor fusion, although other sensors may be discussed in future webinars.


Session One: Here we cover some basics of how LiDAR works with an emphasis on how Civil Maps uses LiDAR. Via a code walkthrough, we show you how to read a sample LiDAR dataset and how to visualize it in a visualizer. We are open-sourcing our hardware abstraction layer (HAL). The discussion includes transforming spherical coordinates into the cartesian system. We talk about packet structure and provide a quick look at our Atlas DevKit and Atlas DevKit Lite, which are low-cost hardware/software development kits for localization, 3D mapping, and data collection.

LINKS Github Video Playlist Slides

Session Two: We discuss how GPS and IMU work together in the context of capturing vehicle motion and a simple technique for creating a trajectory from a sample set of IMU data. After part 1 & 2 you will be able to generate a point cloud by fusing the IMU trajectory and the LiDAR data.

LINKS Github Video Playlist Slides

Session Three (Thursday, May 25th 1:30 PM PST (4:30 PM EST)): This conversation covers general sensor fusion concepts. It begins with a discussion about using 3D semantic maps in sensor fusion. The talk also includes a look at hardware prerequisites (spatial and temporal). Near the end of the seminar, attendees will download some code. They’ll use that to work with Civil Maps cognition.

LINKS Github Video Playlist Slides

Session Four (Thursday, June 29th 1:30 PST (4:30 PM EST)): For the last segment, we will work with a synthetic dataset to explore additional concepts in fusion and cognition.

LINKS Video Playlist Slides


Scott Harvey is a Co-founder and Computer Vision Engineer at Civil Maps. He leads engineering efforts for the Localization team, which has developed innovative compression technology and localization in 6 dimensions for autonomous vehicles. Scott has technical expertise in Computer Vision and Control Systems. He earned engineering degrees from Brown University and Stanford University. When he’s not working on localizing self-driving cars in six degrees of freedom, Scott plays in a band called “Tourist Club.”