Deep encoder from Deep SORT for generating appearance descriptors for images
Sekonix GMSL Camera Drive for NVIDIA Drive AGX Xavier Developer Kit for DRIVE OS Linux 5.2.0 and DriveWorks 3.5
A set of projects aimed at working with usyd datsets. Includes a c++ implementation to read a set of bag + h.264 files and publish them as if it was running live
Stitch multiple camera images and generate a 180 FOV image
A library for transformation with uncertainty. It provides functions such as ego-motion prediction/correction and Gaussian/pointcloud transformation with or without uncertainty.
This project contains a ROS node that can take a pointcloud and project it into an image. The output is a pointcloud with RGB information, or a pointcloud with UV pixel coordinates
Tuatara package: Bounding box detection for multiple image sources based on tkdnn.
Project to take vehicle data and overlay on a video stream
This project is to perform motion correction on the raw lidar data. This is because the lidar scan takes 100ms to be completed, and during this time the vehicle may have moved. We apply a correction based on the odometry information from the vehicle