Catalog Information Image Processing Laboratory for Remote Sensing (3 units) Description: Techniques and applications of digital image processing in remote sensing, multispectral image enhancement and analysis, classification, feature extraction for cartography, rule-based systems for mapping from imagery. Grading: Regular grades A B C D E Additional Information Use of ECE 531 units in an ECE graduate study program may be restricted. See the current Graduate Handbook or the ECE Graduate Advisor for current policy. ECE/OPTI 531 is a core course requirement for the Ph.D. minor program in Remote Sensing and Spatial Analysis. This course was designed and taught by Prof. Robert A. Schowengerdt, who retired from teaching in June, 2004. The Committee on Remote Sensing and Spatial Analysis has decided to continue this course indefinitely to accommodate students who have it in their study plans. For Fall 2005, Dr. Kurtis Thome will be the official instructor, and Dr. Daniel Filiberti will conduct most of the lectures as an adjunct faculty member. Dr. Filiberti is a recent graduate of ECE under Prof. Schowengerdt, and has over a decade of industry experience. Any concerns or grading questions should be referred to Dr. Thome, who has ultimate responsibility for this course. Prerequisites: consent of instructor Instructor Dr. Kurtis Thome Meinel 414 kurt.thome@opt-sci.arizona.edu 621-4535, 621-4929 Lecturer Dr. Daniel Filiberti ECE408A DIAL dan@ece.arizona.edu 621-4554 Office hours (tentative) Dr. Thome, Meinel 414 Mon: 1:30-3:30 Wed: 9-12 Or by appointment
Course Information Textbook The course is largely based on material in Remote Sensing - Models and Methods for Image Processing, R. A. Schowengerdt, 1997. This textbook is required. Course Website Handouts, notes and assignments will be available at http://www.ece.arizona.edu/~dial/ece531/ece531.html Topics (chapter # in parentheses; not all topics will be covered in depth) The Nature of Remote Sensing (1) overview of remote sensing, sensors, images and data systems Optical Radiation Models (2) optical physics of atmosphere, surface reflectance Sensor Models (3) imaging, quantization, resolution Data Models (4) multivariate statistical descriptions, spatial statistics, influence of the atmosphere, terrain and sensor on data Spectral Transforms (5) multispectral ratios, principal components, vegetation indices Spatial Transforms (6) filtering, Fourier transforms, resolution pyramids Correction and Calibration (7) noise removal, geometric rectification, radiometric calibration Image Registration and Fusion (8) spatial correlation, stereo analysis, multisensor fusion Thematic Classification (9) statistical classifiers, neural networks, supervised and unsupervised training Laboratory Assignments Computer programming skills are not required for this course. We discourage the use of low-level programming languages (e.g., FORTRAN, C/C++, etc.) for completing the labs. High-level interpreted environments for image processing (e.g., IDL, MATLAB) can be used, but some of the labs will be difficult to perform even in these due to the user interaction (GUI) requirements. We will make available tclsadie, a relatively easy-to-use image processing system for remote sensing, developed by DIAL students. There are several ways you can access tclsadie (see handout ACCESSING tclsadie), and all of the labs are written for use with tclsadie. You may also choose to use other software to complete the lab assignments. Dr. Thome expects to have several computing systems installed with ENVI (Research Systems, Inc.) for student use in his lab. If you use software other than
tclsadie, you accept responsibility for completing the assignment as specified. The following lab exercises (subject to change) will be assigned: Lab #1: Introduction to Remote Sensing Images and Image Processing Lab #2: Image Statistics/Scatterplots Lab #3: Spectral Transforms Lab #4: Spatial Transforms Convolution Lab #5: Spatial Transforms Fourier Transform Lab #6: Geometric Processing Lab #7: Multispectral Classification Unsupervised and Supervised You should view the labs as homework, and devote an equivalent amount of time to them. You can do your lab work individually, or in small groups (2 or 3 students only). I encourage you to pair up with another student in the class; my experience indicates that working in pairs is beneficial to you in learning the material and in overcoming computer-related problems ( Two heads are better than one! ). For a group, one lab report is submitted and each person in the group will receive the same grade. Late Assignments Assignments will be due at the beginning of class on the specified date. Late lab reports, homeworks or missed exams will receive zero scores, unless it has been pre-arranged with instructor approval. Homeworks, Exams and Grading Your semester grade will consist of 2 exams (15% each): 30% 7 lab reports (lab 1 is 5%, 2-6 are 10% each, 7 is 15%): 70% 2 homework assignments (up to 5% lab report extra credit each, see below): 0-10% Lab reports will be graded into one of the following categories: Excellent: 90, 95, 100% Good: 75, 80, 85% Fair: 55, 60, 65, 70% Poor: 35, 40, 45, 50% Fail: 10, 15, 20, 25, 30% None: 0% The optional homework assignments are intended to provide examples of questions that could be asked on an exam. Each homework assignment will add up to 5% extra credit to your lab report score, not to exceed the total value (70%) of the lab reports. There will be an optional, comprehensive final exam for students who would like an
opportunity to improve their grade. If you take the final and your score is higher than your lowest score on the previous two exams, I ll replace that score with the final score and recalculate your semester average and grade. If your score on the final is lower than either of the two previous exams, the original semester grade will remain as your grade. The optional final will be on December 13th, 2005 from 11:00 AM to 1:00 PM. I will have your final semester grade by the last day of class. This grade will be assigned based on the class distribution, i.e. "on the curve. If you achieve a total score of 92% or higher, you will receive an A regardless of the class distribution. I will not regrade papers on the basis of a reason such as I think I deserve more points. I will consider regrading, if you have objective reasons (error in scoring, alternative solution, etc). Attendance You are responsible for obtaining all information disseminated during lectures. Attendance is not monitored nor graded. Academic Integrity Integrity is expected of every student in all academic work. The guiding principle of academic integrity is that a student s submitted work must be the student s own. The entire Code of Academic Integrity, with its responsibilities, rights, and procedures is available through the Dean of Students office and at http://w3.arizona.edu/~dos/uapolicies. Disability Accommodations Students with disabilities: If you anticipate the need for reasonable accommodations to meet the requirements of this course, you must register with the Disability Resource Center (DRC) as soon as possible. After your registration, you can provide the instructor with a letter of identification and discuss with the instructor how the course activities and requirements will impact your ability to fully participate in the class. The DRC will work with you and the instructor to establish appropriate accommodations. Their web address is http://drc.arizona.edu/. Email Check your email regularly (once a day at a minimum). Important information will sometimes be broadcast to the whole class over email. Q&A Sessions To improve communication and facilitate the learning experience, we may meet near campus around lunchtime a couple of times during the semester. The intent is to encourage informal discussions and questions, hopefully related to the class. This is not an official part of the course and attendance is not mandatory, but encouraged. Each student will pay for his/her own lunch, i.e. this is not a faculty treat!
Frequently Asked Questions (FAQs) What is the course all about? - Physics of optical remote sensing - Image processing techniques for remote sensing - How these two topics are related for information extraction/interpretation What software will be used? tclsadie, an image processing library for UNIX. Windows and Macintosh versions are also available but not supported. At your own risk, other software can be used, but it may not be able to handle some assignments. Every student is urged to have the UNIX version up and running as a fallback if other software is inadequate. Why don't we use some popular software like ERDAS Imagine? Because my intent is to teach you the techniques of image processing, not how to use a particular computer program. Imagine and similar commercial programs have a long learning curve that takes too much time in one semester; tclsadie has a relatively short learning curve and more flexibility. Will any programming be required? No. What is the math level in this course? Less than a typical ECE graduate course. Calculus, basic vector-matrix operations, and multivariate statistics are required. Will there be any homework? The lab assignments should be considered homework, as well as the optional homework assignments. Can I work alone on the lab assignments? Not recommended; you may team with one or two other students. Diverse pairing such as engineering students with non-engineering students, can produce the most benefit from the assignments. If your initial team doesn't work out, you can re-team for a subsequent assignment. What is expected in the lab reports? - Complete description of what you did and how you did it, and why. - Your results, including images, numerical results, graphs - Some tclsadie session log records, as appropriate for your documentation - A clear, logical organization in your approach to the assignment - Neatness and clarity count! Can lab reports be turned in outside of class? No - Email inbox quotas prevent electronic submission.
Mailboxes are open pigeon holes and are not secure. DIAL/ECE408 may be unattended with no one to receive the homework. Extenuating circumstances: pre-arrange submission with instructor. Are there any restrictions on use of this course for an ECE major? - MS: cannot be used for ECE "core" (9 units w/thesis, 15 units w/o thesis) - PhD: cannot be used for 9-unit ECE "core"; can be used for some minors including Remote Sensing and Spatial Analysis minor. (see http://www.ece.arizona.edu/academics/advising/grad/handbook/handbook.pdf)
Course Schedule (estimate) Date Chapter Section Topic Assignment Due 8/23 Overview Introduction 8/25 1.1-1.4 Remote Sensing Systems Signatures 8/30 1.5-1.6 Image Display Data Systems 9/1 Optical Radiation Models 2.1-2.2 VNIR-SWIR Lab 1 9/6 2.3 MWIR-LWIR 9/8 Sensor Models 3.1-3.8 Resolution Spectral Response Spatial Response Signal Amplification Sampling and Quantization 9/13 Data Models 4.1-4.3 Univariate Statistics 9/15 4.4-4.6 Multivariate Statistics 9/20 4.7-4.9 Topographic and Sensor Effects Lab 2 9/22 Spectral Transforms 5.1-5.3 Introduction Feature Space Multispectral Ratios 9/27 5.4-5.5 Principal Components Tasseled-Cap 9/29 5.6-5.8 Contrast Enhancement Lab3, HW 1 10/4 Review 10/6 Exam 1 10/11 Spatial Transforms 6.1-6.3.2 Linear Filters Statistical Filters 10/13 6.3.3 Gradient Filters 10/18 6.4 Fourier Transforms Lab 4 10/20 6.5 Scale-Space Transforms 10/25 Correction and Calibration 7.1-7.2 Noise Reduction 10/27 7.3 3.9.1-3.9.6 Radiometric Calibration Distortion Correction 11/1 Thematic Classification 9.1-9.3 Introduction Process Feature Extraction 11/3 9.4-9.5 Training Nonparametric Lab 5 Lab 6 11/8 9.6 Parametric 11/10 9.7-9.8 Spatial-Spectral Segmentation Subpixel Classification 11/15 9.9 Hyperspectral Analysis Lab 7, HW 2 11/17 Review 11/22 Exam 2 11/24 Thanksgiving 11/29 8.1-8.4 Registration GCP Location Hierarchical Warp Stereo 3.9.7 12/1 8.5 Multi-image Fusion 12/6 OPEN TOPIC