Course Name:Embedded Intelligence and Vision
Guided Autonomous Vehicles
| Course Code: S2293174
|
Semester: 4
| Credit: 3
|
Program: Electronics
|
Course Module: Compulsory
|
Responsible:Yan Xu, Hua Li
| E-mail: xuyan@tju.edu.cn
|
Department:Tianjin International Engineering Institute, Tianjin University
|
Time Allocation(1 credit hour = 45 minutes)
Exercise
| Lecture
| Lab-study
| Project
| Internship(days)
| Personal Work
| 12
| 24
| 12
| 0
| 0
| 0
|
|
Course Description
This is a graduate level specialization course designed to cover the state-of-the-art techniques in the field of embedded intelligent vision guided autonomous vehicles, early vision of ROI localization, Canny and LoG edge detection and segmentation, Hough Transforms and moments-based feature extraction, Laser Imaging Radar, Ultrasonic range finder and their integration with Stereo Vision Sensors. Also includes vision guided path extraction, driving obstacle/hazardous objects detection, vehicle nonlinear control modelling, PID, Kalman control with integration of intelligent controller design, OpenCV and OpenGL tool for implementations and GPU technology. Also requires four hands-on labs with embedded GPU platform.
|
Prerequisite
Image Processing Digital Signal Processing Control Theory Machine Learning (5) Computer Architecture (6) Embedded Systems (7) Operating Systems (8) System Programming.
|
Course Objectives
To be able to evaluate, design and implement vision based algorithms for feature extraction and segmentation for vehicle path planning.
To be able to analyze, design, and implement algorithm for integration of stereo vision and multi depth sensor input to form 3D depth mesh.
To evaluate, design, implement and improve PID and Kalman control with intelligent controller design techniques.
To understand machine learning techniques employed for vision application and for path extraction, driving obstacle/hazardous objects detection.
To be able to implement the design on embedded platform.
|
Course Syllabus
Introduction to digital image processing, edge detection techniques using C to manipulate digital images, tools for image processing, Matlab (or open source equivalent Octave), and OpenCV, OpenGL. Embedded Architecture for Computer Vision. Safety issues and procedures.
Image enhancement via histogram equalization techniques. 2D convolution technique. 2D convolution with LoG (Laplace of Gaussian) kernel and zero crossing for edge detection. Image binarization.
Binary image processing and Hough transform, binary image processing for pattern recognition. ROI localization, path extraction, traffic sign segmentation. Image/3D scene feature extractions. Image transformations, bird-eye view techniques. Image segmentation techniques and its applications in pattern recognition. Embedded platform with GPU processing capability.
Stereo Vision, depth map, point cloud map. Laser Image Radar principles and its implementation.
Laser Imaging Radar, Ultrasonic range finder and their integration with Stereo Vision Sensors. Vision guided path extraction, driving obstacle/hazardous objects detection. Embedded GPU implementation discussion.
FFT and its power spectrum. Motion estimation and optic flow computation.
Image tracking techniques and introduction to Kalman filter. Applications with Digital Image Processing. Implementation on embedded platform.
Nonlinear control system model of a vehicle motion trajectory control, PID controller design for vehicle control, Kalman filter techniques for motion trajectory tracking.
Intelligent controller design and integration/enhancement of modern controller design. Integration of vision technique with vehicle control.
Intelligent controller fine tuning and integration/enhancement with machine learning capability.
|
Textbooks & References
B.K. P. Horn.Robot Vision. The MIT Press, ISBN 0-262-08159-8, or 0-07-030349-5 (McGraw Hill).
Rafael C. Gonzalez and Richard E. Woods.Digital Image Processing(3rd ed). Prentice Hall, ISBN 0-201-18075-8.
Reference textbooks (optional)
Bradski and Kaebler.Learning OpenCV,Computer Vision with the OpenCV Library.O’Reilly Publisher, 2011. ISBN 978-0-596-51613-0.
Hearn Baker.Computer Graphics with OpenGL (3rd ed). Prentice Hall. ISBN 0-13-015390-7.
|
Grade Distribution
Attendance: 40% Final Exam:60%
|
Capability Tasks
CT1: To understand basic science, and to have analytical ability and the ability to integrate related knowledge.
CT2: To apply relevant professional knowledge to the field of science and technology: understanding of the basic concepts andits connotation, application of different methods and concepts which have been learned, capability of judging the scope and limitations of such applications.
CT3: To grasp methodologies and engineering tools: identifying, utilizing and solving problems. Even if the students are not familiar with the content, they can turn to computer tools for systematic analysis.
|
Achievements
To be able to evaluate, design and implement vision based algorithms for feature extraction and segmentation for vehicle path planning. - Level: A
To be able to analyze, design, and implement algorithm for integration of stereo vision and multi depth sensor input to form 3D depth mesh. - Level: M
To evaluate, design, implement and improve PID and Kalman control with intelligent controller design techniques. - Level: M
To understand machine learning techniques employed for vision application and for path extraction, driving obstacle/hazardous objects detection. - Level: A
To be able to implement design on embedded platform. - Level: A
|
Students: Electronics, Year 2
|
