Biomedical Engineering Reference
In-Depth Information
5.5 Industrial applications of
image analysis
In the fi eld of plant sciences, image analysis provides a way to non-
subjectively measure many parameters specifi c to plants. Two examples are
discussed here that have different objectives and associated challenges, to
phenotype seeds and whole plants. The phenotype is the physical appearance
of the plant and is linked both to its genome and environment, so measuring
its parameters accurately is a core competency within plant sciences. Before
image processing or analysis can be done, the image has fi rst to be captured
consistently and this is as important as the subsequent
in silico
steps.
The fi rst example used a fl at bed scanner, which gave the advantage
of consistent lighting across the fi eld of view, and a known image scale
often expressed in dots per inch (DPI). The disadvantage of using
the scanner is that only small or 2D subjects are suitable for scanning.
The starting and fully processed and analysed image is shown in
Figure 5.7.
Imaging of seeds using a fl at bed scanner (left-hand
side). Image analysis can be used to measure many
morphological features of the seeds and return the
mean scores of longest length (shown on the false
colour image of the seeds as a white line in the
right-hand side), calliper width, circularity (in which a
value of 1 is a perfect circle and values are from 0-1),
area and, fi nally, perimeter
Figure 5.7
Search WWH ::
Custom Search