Image Processing Reference
In-Depth Information
Fig. 12.2
The block-diagram of the image processing. The input is an image from the
camera—very noisy one might add. In the
center
of the image the lower part of a person is present).
The output is the position of each foot (indicated by the two
×
)
They quickly agreed that the first three blocks where no problem and therefore
focused their energy on the image processing block: Find Feet.
“Hmmm, any ideas on how to
...
” SB began, before he was interrupted by a big
smile on Mick's face.
“What!” SB started before he realized the answer himself and continued: “of
course, let's look in
the
topic”.
They ventured into Chap. 1 and soon found an overall structure for the image
processing. After studying the different book chapters they also managed to figure
out which algorithms to apply in the different sub-blocks, see Fig.
12.2
.Belowwe
shall have a closer look at how SB and Mick pulled it off.
12.2 Setup
12.2.1 Infrared Lighting
SB and Mick quickly realized that such a simple thing—for humans—as to find the
feet of a person can be quite complicated for a computer. The simplest solution they
came up with was to ask the user to wear distinctly colored socks/shoes and then
do thresholding in some color space. However, for some reason Mick didn't like
touching other people's shoes (!) and therefore came up with an alternative solution.
He had earlier in his life been puzzled about the fact that his TV remote control
sends out light in order to control the TV and that he could not see the light. He
learned that, this was because the remote was sending out infrared light, which the
human eye cannot see, i.e., humans can only see light of certain wavelengths, see
Fig. 2.2. He then tried to see if different cameras could actually see this light and
learned that some can and some can't—strange. He asked an expert who explained
that a camera is produced in order to capture the same as the human sees and hence a
filter is usually inserted into the camera that prevents the infrared light from entering
the sensing chip. In general, the better the filter the more expensive the camera. Mick
tried to take some cameras apart and actually found out that the filter is sometimes