Machine Vision for Biological Systems

coatiarfAI and Robotics

Oct 17, 2013 (3 years and 11 months ago)

77 views

Instructor: Manoj Karkee
http://www.bsyse.wsu.edu/core/directory/faculty/karkee-manoj.htm
BSysE 530
Machine Vision for Biological Systems
What do you expect to Learn?
Bikram
Jianfeng
Jingjin
Meng
Overview of the Class
• Class Details
• Syllabus
• Quick Intro of Machine Vision
• Some Applications
• Challenges and Future Direction
• A Little Survey
Class Details
Meeting Time and Place:
Class meets on Tuesday and Thursday at 1:10PM to 2:25PM in
Room 125 Ag Tech Building, IAREC-Prosser, and via
Telecommunications
Instructor:Manoj Karkee
•Office Hours: Thursday 2:30PM to 3:30PM or by appointment.
•Office: 105 Ag Tech Building
•Phone: 509-786-9208
•Email: manoj.karkee@wsu.edu
Class Objective
• Enable students apply image processing and
machine vision techniques to solve resaerch
and engineering problems
– Introduce applications related to Agricultural
and Biological Systems
Text Books
Shapiro and Stockman:
Computer Vision
Gonzalez, Woods and Eddins:
Digital Image Processing with Matlab
+
One Article Per Week as Listed in Syllabus
Some References
Jain, Kasturi and Schunck –
Machine Vision
Gonzalez and Woods –
Digital Image Processing
Pratap, Rudra - Getting Started
with MATLAB: A Quick
Introduction for Scientists and
Engineers
+
GOOGLE :(
Assignments
• Use of Angel
• Class Participation
• Homework
– 8 Homework
– 4 Require Formal Report
– Report Format
• Class Project
• NO Exams
Final Project
• Each student will finish a final project related to the topics
covered in the class.
• Final projects must address a critical need or bridge some
knowledge gap in their respective field.
• Students are encouraged to think about potential topics for
final project as early as possible. A written project proposal
(8-10 pages) will be due on March 01.
• Each student will present the results of his/her project on
May 03.
• The final project report (20-25 pages) will be due on May
05.
Grading
Score Weightage
 Class Participation: 5%
 Homework Assignments: 50%
 Final Project Proposal: 10%
 Final Project Presentation: 15%
 Final Project Report:20%
Grading scheme:
 ≥85 – at least A-
 ≥75 – at least B-
 ≥65 – at least C-
 ≥55 – at least D-
Policies
Homework and Project
 Due at the beginning of class
 Late submission - 10% Penalty Each Day
Grading Disagreement:
 Visit me in person
 Within two days
Topics to be Covered
Syllabus, Report Format and
Other Materials are available
in
Angel
https://lms.wsu.edu/
Vision
• What is human vision?
• What are the components?
• What do eyes do?
• See
• What?
• Where?
• How do vision work?
• Capture Images - Eye
• Process/ Understand – Brain
• Color
• Geometry
• Texture
• Motion
• Stereo
• Shading
• Spatial Relationship
• Applications
• Actuation
Machine Vision
Making Sense from Images
What:Apples
Type:Red Delicious
Number:26
Large Class:15
Medium Class:11
Average Dia:8cm
Yield:3KG
3D Coordinates:
1:(20,12,30)
2:(21,5, 43)
3D Measurement
(e.g. Stereo)
Mid Level Processing:
-Segmentation
- Feature Extraction
Low Level Processing:
- Enhancement
- Filtering
High Level Processing:
- Recognition (e.g. Matching)
- Classification
Acquisition
Machine Vision in Relation to other Fields
Machine Vision and Computer Vision are often Interchangeable; Image
processing is prerequisite
machinevision.co.uk
Applications
caltech.edu
-Text/Language Recognition
-Medicine
-Biomedical Applications
-Supply Chain
-Image mining – Google goggles
-Manufacturing
-Mining
-Construction – traffic surveillance
-Crime Investigation
-Bio-Identification
-Transportation
-Military
-Archaeology
-Remote Sensing Applications
-Food and Agriculture
-Harvesting
-Sorting/Grading
-Packaging
-Loading/Uploading
militaryaerospace.com
alliedvisiontec.com
optoiq.com
http://imagegraphics
videosoftware.com
Some work from ISU
http://www.hci.iastate.edu/575x/doku.php
More ISU Work
Face tracking
Video: Blurring
Some CPAAS Projects
Kang et al.; Sucker Detection of Grapes for Targeted Spray
Using Optical Sensors
Sucker detection system on an
over-the-row frame
Raw image
Processed image
A Prosilica CCD camera
BermApplication
Targeted ApplicationBarrier Application
Some CPAAS Projects
Kang et al.; Automatic Barrier Application Systems for
Targeted Control of Cutworm in Vineyards
Sensor-1
measurement
area
Sensor-2
measurement
area
Vehicle
position
at t0
Correspon
ding trunk
v
Φ
Vehicle
position
at t1
Multi-nozzle array
Hokuyo Laser Range Finder
Some CPAAS Projects
Wang et al.; Automatic and rapid cherry rating in terms of
cherry color and size
Level 7 6
5 4 3 2 1
corresponding
color
Cherry chart
Camera
LCD Screen
Original
Image
Binary
Image
Cherry
Contour
Cherry
Size
Grape Pruner: Vision Robotics
http://www.youtube.com/watch?v=9GaGO9LIDEA&feature=pl
ayer_embedded#!
Some External Videos
Strawberry Harvester: Robotic Harvesting LLC
http://www.youtube.com/watch?v=RCBQqEGp8Go&feature=related
WSU Research:
http://www.youtube.com/watch?v=PHc7P2uE1co
Challenges
•Environmental Variability/Uncertainty
•Lighting/Illumination Variability
•Occlusion
•Global View
•Processing Speed
•Flexibility
•Application in Agricultural Robotics
•Complex Tree and Bushes
•Dedicate Nature of Produce
•Inconsistency in Crop Shape and Size
•Actuation Speed
•Robustness and Capacity

Cost
Future Need/Direction
• Increasing computational power
• Horticultural Modification
• Precision Applications
• Chemicals
• Nutrients
• Water
• Autonomous Monitoring/Scouting
• Crop Load Estimation
• In-field Sorting/Grading
• Canopy, branch/shoot, flower and fruit
detection and mapping
• Agricultural Robotics (fruits and
vegetables)
• Harvesting
• Pruning
• Thinning – flower and fruit
• Pollination
What do you expect to Learn 2?
Bikram
Jianfeng
Jingjin
Meng
Little Survey
Image
Processing
Artificial
Intelligence
(e.g. Fuzzy
Logic)
Calculus
(e.g. Maxima
and Minima)
Linear
Algebra
(Matrix etc.)
Matlab Other
Programming
(e.g. C)
Bikram
Jianfeng
Jingjin
Meng
*At least about 6 months or one semester of exposure
THE END

Office Hours: Thursday 2:30PM to 3:30PM or by appointment.
•Office: 105 Ag Tech Building
•Phone: 509-786-9208
•Email: manoj.karkee@wsu.edu