Docker Image Details
The Ultralytics docker image was used as the base image for this project. It includes the requried GPU tools to run yolo and was easy to modify, so it was used as a base image. It also offers base images for the Jetson-5 board which is ideal for our purposes.
The configuration, data, output and tracking directories are added to the image after it's built for this project. Each of those sections is described in below.
-
configuration - Directory that contains the required configuration files for the application to run. Many settings can also be overwritten through the cli interface.
-
data - Directory recomended for the user to mount pre-recorded data/samples to run the tracking algorithm against. This is configurable, but this is shown as a placeholder.
-
output - Output directory where files from tracking will be saved, it is recommeded to mount this directory to a local volume to grab the data easily.
-
tracking - directory that contains the additional radar tracking and object tracking pipeline
Container Structure After Additions
|-- bin/ |-- boot/ |-- configuration/ | |-- RadarConfig.yaml | |-- VideoConfig.yaml | |-- TrackingConfig.yaml |-- data/ | |-- radar | |-- video |-- dev/ |-- etc/ |-- home/ |-- media/ |-- output | |-- 2024-09-15_21-45-01 | |-- other-outputs |-- tracking/ | |-- UAV-Object-Tracking/ | | |-- tracking.py | | |-- PROJECT FILES |-- ultralytics/ | |-- CITATION.cff | |-- CONTRIBUTING.md | |-- LICENSE | |-- README.md | |-- README.zh-CN.md | |-- calibration_image_sample_data_20x128x128x3_float32.npy | |-- docker/ | |-- docs/ | |-- examples/ | |-- mkdocs.yml | |-- pyproject.toml | |-- tests/ | |-- ultralytics/ | |-- ultralytics.egg-info | `-- yolov8n.pt |-- var/ | other system folders