Machine Vision Applications

1
Machine Vision Applications

• Case Study No. 3
• Analysing Images of Living Plants
• Visual control of a robot for micropropagation

2
Micropropagation

• Micropropagation
• Rapid non-fruiting copying of genetically
  identical plants
• How micropropagation is performed
• Human beings cannot do it properly
• Plants are fragile
• People infect growing medium
• Blind robots cannot do it
• Visually guided robot could
• Plant variation and structure
• Intra-species variation
• Open and closed plant structures

3
Open Structure Plant (Rose)
Leaves, stalks and stem are all discarded
Bifurcation leaf stalk attached to main
stem (nodal bud growing)
Replant Y-shaped piece
Planted in agar jelly
4
Lighting and Viewing

• Back-lighting (red to increase contrast)
• Two or three cameras, horizontal view
• Take care to avoid stray light

5
Preprocessing

• Enhance contrast and threshold
• Follow with binary closing operator
• Crack detector
• Applied to grey-scale image
• Finds stem and stalks removes leaves from the
  image)

6
Locating Bifurcations

• Skeletonising / thinning
• Find joints by counting white 8-neighbours
• Morphology / N-tuple
• V-shaped structuring element / kernel
• Place pastry cutter on each bifurcation
• Cutting uses a high-power scanning laser (minimal
  plant damage)

7
Handling Occlusion

• Birfucations identified from any camera
• OR results from all cameras
• Interpolate from one view to another when
  tracking the main stem - needs intelligence
• Ineffective for closed plant structures

8
Additional Measurements

• Height
• Width
• Symmettry and other aesthetic shape parameters
• Compactness (leggy plants are weak)
• Total leaf area - determines plant growth
• Leaf size distribution
• Stem / stalk widthr (measures physical strength)
• Leaf droop (indicates need for watering)

9
Reading

• B. G. Batchelor, Intelligent Image Processing in
  Prolog, Springer-Verlag, Berlin New York,
  1991, ISBN 0-540-19647-1 ISBN 0-387-19647-1,
  7.7.
• M. Graves B. G. Batchelor, Machine Vision for
  the Inspection of Natural Products , Springer
  Verlag, London, 2004, ISBN 1-852-33525-4, 3.5.2.