udacity/self-driving-carArchived
Self Driving Car
This is an open-source autonomous driving perception pipeline that processes camera and lidar sensor data to detect, track, and fuse objects in real-world driving environments. The project integrates an end-to-end perception workflow combining sensor calibration, deep learning object detection, Kalman filter tracking, and sensor fusion for robust scene understanding.
The pipeline includes camera calibration tools to remove lens distortion from raw images, deep learning model training for object classification and detection, and multi-object tracking using Kalman filters with data association techniques. It also provides 3D point cloud processing for lidar-based object identification and a sensor fusion framework that combines camera and lidar detections to improve perception robustness and reliability.
The project covers the full perception stack from sensor calibration through detection, tracking, and fusion, with capabilities for both image-based and point cloud-based object detection.
Features
- Perception Pipelines - Provides an open-source pipeline for processing camera and lidar data to detect, track, and fuse objects.
- Object Detection and Tracking - Locates and maintains tracks for multiple objects using deep learning models and Kalman filters.
- Object Detection - Locates multiple objects within a single image by applying a trained object detection model.
- Object Detection - Applies trained convolutional neural networks to locate multiple objects within images.
- Autonomous Driving Detections - Locates and classifies objects in images and lidar point clouds for real-time autonomous driving scene understanding.
- Kalman Filter Trackers - Maintains and updates tracks for multiple objects using Kalman filters and data association techniques.
- Multi-Object Trackers - Maintains and updates tracks for multiple objects using Kalman filters and data association techniques.
- LiDAR-Camera Detection Fusions - Combines camera and lidar detections to improve perception robustness and reliability through sensor fusion.
- LiDAR-Camera Fusions - Combines camera and lidar detections to create robust object perception for autonomous vehicles.
- Coordinate Alignment Systems - Provides camera-lidar coordinate alignment tools essential for accurate sensor fusion in autonomous driving.
- Detection Fusion Frameworks - Combines camera and lidar detections to improve perception robustness and reliability in autonomous driving.
- Point Cloud Detections - Identifies objects from lidar sensor data using deep learning on three-dimensional point clouds.
- Lidar Object Identifications - Identifies objects from lidar point cloud data using deep learning for three-dimensional perception.
- Camera Calibration - Removes lens distortion from raw camera images to improve accuracy of downstream perception algorithms.
- Lens Distortion Removal - Removes lens distortion from camera images to improve accuracy of downstream perception algorithms.
- Multi-Object Trackers - Implements Kalman filter-based multi-object tracking with data association for autonomous driving perception.
- Camera Calibration Tools - Removes lens distortion from raw camera images to improve downstream perception algorithm accuracy.
- Convolutional Neural Network Training - Trains convolutional neural networks for object classification and detection in autonomous driving scenarios.
- Vision Model Training - Trains convolutional neural networks to classify and detect objects in images using a deep learning framework.
- Deep Learning and Computer Vision - Trains convolutional neural networks to classify and detect objects in images for autonomous driving applications.
- Educational Courses - Educational content for training autonomous vehicle software engineers.
Star history
How this analysis was created: This summary and feature list were written by an AI model that read the project's README and public documentation pages. Each feature links to the documentation it came from; stars, license and language come straight from the GitHub API. The model does not read the source code, and the analysis is refreshed when the project is re-analysed. Learn more on our About page.
