460-4131/01 – Optical Systems for Autonomous Driving (OSAJ)
Gurantor department | Department of Computer Science | Credits | 5 |
Subject guarantor | Ing. Radovan Fusek, Ph.D. | Subject version guarantor | Ing. Radovan Fusek, Ph.D. |
Study level | undergraduate or graduate | Requirement | Compulsory |
Year | 1 | Semester | summer |
| | Study language | Czech |
Year of introduction | 2019/2020 | Year of cancellation | 2022/2023 |
Intended for the faculties | FEI | Intended for study types | Follow-up Master |
Subject aims expressed by acquired skills and competences
The course acquaints with the topics of image analysis which accompany to the autonomous driving topics. In case of completing the course, students gain an overview of modern methods of image analysis which are used in the autonomous cars.
Teaching methods
Lectures
Tutorials
Project work
Summary
The following topics are covered: Image processing methods for object detection and recognition in the area of self-driving cars. Methods for object localization in the vehicle surrounding based on the machine learning (image descriptors, convolutional neural networks, deep learning, SVM). Principles of vehicle detection, pedestrian detection, traffic sign detection, road line detection. Object recognition based on lidar and depth images.
Compulsory literature:
Recommended literature:
1. Gonzalez, R., C., Woods, R., E.: Digital Image Processing, Prentice Hall, ISBN-10: 013168728X, ISBN-13: 978-0131687288, 2007
2. Michael Beyeler: Machine Learning for OpenCV, Packt Publishing, ISBN-13: 978-1783980284, 2017
Way of continuous check of knowledge in the course of semester
Conditions for granting the credit:
The tasks that form the program of exercises must be worked out.
Exam - written test.
E-learning
Other requirements
No further requirements are imposed on student.
Prerequisities
Co-requisities
Subject has no co-requisities.
Subject syllabus:
Lectures:
1. Main ideas behind object detection in images, a sliding window method.
2. Object detection methods, Haar-like features (Viola-Jones detector).
3. Local binary patterns for object detection.
4. Pedestrian and vehicle detection methods, histograms of oriented gradients.
5. Road line recognition in images.
6. Convolutional neural network.
7. Keypoint detectors and descriptors (SIFT, SURF).
8. AdaBoost and support vector machines for recognising the objects in images.
9. Traffic lights recognition in images.
10. Processing the images in IR spectrum and multispectral images.
11. Depth image processing (RealSense, Kinect).
12. LIDAR and spherical image processing.
13. Summary of lecture themes.
14. Reserve.
Exercises:
1. Implementation of basic template for object detection in images.
2. Implementation of a sliding window method.
3. Preparing data for training and testing.
4. Object detection using Haar-like features.
5. Object detection using local binary patterns.
6. Object detection using histograms of oriented gradient.
7. Convolutional neural network.
8. AdaBoost and SVM for recognising the objects in images.
9. Object recognition in IR images. Image enhancements and subsequent processing.
10. Depth image processing (RealSense, Kinect).
11. LIDAR and spherical image processing.
12. Combination of detectors for autonomous cars.
13. Reserve.
14. Credit.
Conditions for subject completion
Occurrence in study plans
Occurrence in special blocks
Assessment of instruction