課程名稱 |
電腦視覺應用於工程 Computer Vision in Construction |
開課學期 |
111-2 |
授課對象 |
工學院 營建工程與管理組 |
授課教師 |
林之謙 |
課號 |
CIE5141 |
課程識別碼 |
521EU9280 |
班次 |
|
學分 |
3.0 |
全/半年 |
半年 |
必/選修 |
選修 |
上課時間 |
星期四6,7,8(13:20~16:20) |
上課地點 |
土研402 |
備註 |
本課程以英語授課。 限本系所學生(含輔系、雙修生) 總人數上限:30人 |
|
|
課程簡介影片 |
|
核心能力關聯 |
本課程尚未建立核心能力關連 |
課程大綱
|
為確保您我的權利,請尊重智慧財產權及不得非法影印
|
課程概述 |
This course introduces 2D and 3D visual sensing (Computer Vision) and how it can solve civil engineering problems. This is an application-driven and project-based course that also covers basic understanding of the theoretical concepts of computer vision and image processing (e.g. feature detection and matching, 3D scene reconstruction, object detection and tracking, and SLAM).
Learning is expected to occur by:
● Lectures/discussions in the classroom,
● Hands-on sessions
● Readings
● Class project |
課程目標 |
By the end of the course, students will have full understanding of the following concepts and will be prepared for further vision-related investigations with engineering and management applications:
1. Basics of image formation and processing: digital images and video streams, camera models and camera calibration techniques
2. Fundamental concepts of single-view metrology, multiple view geometry and structure-from-motion and their application for 3D site reconstruction and recognition
3. Basics of image processing, filters, detectors and descriptors
4. Concepts of object classification, localization and detection
5. Range, scope and advantages of computer vision techniques for monitoring construction progress, productivity, safety, and quality of operations in addition to structural health monitoring and stability analysis
6. Basics of machine learning, deep learning techniques for interpreting visual data |
課程要求 |
Pre-class Learning Materials
● Please review the pre-class learning material beforehand
● The progress will follow the schedule in the next section
Hands-on Session
● Submission of hands-on session assignments is required, or it will be considered as an absence.
● Once your hands-on session assignments are submitted and reviewed by the TA, you may leave the class early.
|
預期每週課後學習時數 |
|
Office Hours |
|
指定閱讀 |
|
參考書目 |
● Computer Vision: Algorithms and Applications, R. Szeliski, Springer, 2011. http://szeliski.org/Book/HZ
● Multiple View Geometry in Computer Vision, by R. Hartley and A. Zisserman, Academic Press, 2004. http://www.robots.ox.ac.uk/~vgg/hzbook/FP
● Computer Vision, A Modern Approach, by D.A. Forsyth and J. Ponce, Prentice Hall, 2003.http://luthuli.cs.uiuc.edu/~daf/book/book.html.
● Related journal papers and proceedings. |
評量方式 (僅供參考) |
|
週次 |
日期 |
單元主題 |
Week 1 |
|
● Course Introduction
● Computer Vision Application in Construction
● Software Installation |
Week 2 |
|
● Linear Algebra
● Image Warping
● Hands-on (Transform image with transformation matrix)
|
Week 3 |
|
● Light, Shading and Color
● Hands-on (Convert RGB image to black & white)
|
Week 4 |
|
● Linear Filtering
● Hands-on (Edge detection)
|
Week 5 |
|
● Feature Detection and Description
● Hands-on (Feature detection and matching using OpenCV)
|
Week 6 |
|
● Camera models
● Single-view geometry and calibration
● Hands-on (Camera calibration, measuring height)
|
Week 7 |
|
● Epipolar geometry
● Stereo
● Hands-on (Homography)
|
Week 8 |
|
● Structure from motion I
● Hands-on (Epipolar geometry)
|
Week 9 |
|
● Clustering
● K-means
● Hands-on (OpenCV, OpenSfM, or Visual SfM)
|
Week 10 |
|
● Categorization and Classifiers
● Labeling
● Hands-on (Final Project discussion)
|
Week 11 |
|
● Neural Network
● Hands-on (Perceptron, multi-layer perceptron)
|
Week 12 |
|
● Convolutional Neural Network
● Hands-on (CNN classifying construction equipment)
|
Week 13 |
|
● Object Detection
● Hands-on (RCNNs)
|
Week 14 |
|
● Object Tracking
● SLAM
|
Week 15 |
|
● Guest Lecture
● Activity Recognition
● Summary
|
Week 16 |
|
● Final Project Presentation |
|