Blob
Search Tool
This tool allows the user to search an image for a specific
color, distance, or shape pattern, once it has been trained upon a
specific portion of the image as a blob.
Using the
Blob Search Tool
A Blob must be inputted into the Blob Search
tool's inputs page in order for the tool to work.
1. Drag and
drop the blob search tool from it's position into the ROT Main, then
"left click" onto it. This
will allow an ROI to be placed into the image in the viewer. It will be
located as a blue box in the upper left hand corner
of the CIF Viewer.
2.
Move the ROI to the desired place and select "train".
3.
Run the tool. Once the tool is run the information about the blobs will
be stored within the tool's results page.
Inputs:
A
blob set or a blob tool that has found blobs within it will need to be
inputted into this tool in order for the tool to work.
CIFViewer
The
Viewer that displays the "Blobs"
CIF
The input image.
CIF stands for "Common Image Format". This is Automation
Manager's native graphics format (like .jpg or
.bmp).
ROI
"Region of
Interest". By default this is the green rectangle that is placed on the
screen when the tool is selected. Supports most ROI shapes. Only used to Train. The Search lys searches all presented blobs.
Blobs
Where the blob set or other blob tool needs to be "dragged
and dropped" to for the Blob Search tool to find it's information.
Blob Compare:
This
portion of the tool takes comparisons from the original blob used and
compares it with the set parameters/areas that the user has set on the
blob compare page and checks those against the last blob
found.
Results:
Results:
The number selection
at the top will allow the user to change which blobs information they
are looking at.
Number
Blobs: The number of blobs found and measured.
Total Area: The combined area of all the blobs.
For each blob:
Area
Perimeter
Location
Number of
holes/children
Foreground or
background
Moments about each
axis
Width
Height
Roundness 1 being round
else < than 1
Squareness
1 being square else < than 1
Form factor Width/Height never exceeding 1
The
number selection at the bottom shows the colors of each blob in terms
of blue (1) green (2) and red (3).
Parameters:
Normalize Distance:
If this box is checked, It will normalize it's findings according to
the area. So if it finds a similar image but the image is smaller than
the original image, the tool will compare the area and
dimensions of the new blob
against the old to find if it is similar or the same.
Tolerance Distance:
The distance the user will tolerate the new image as compared to the
old image. The Tolerance expected will be found by finding the
difference and subtracting it.
Maximum Compared Error:
The max amount of error the user is willing to have between the images.
This will also be found by subtracting the difference.
Find Rotated: If
this box is checked, the tool will find blobs that are similar to the
original in any rotation.
Compare Method: This
allows the user to decide what parameters the tool will compare the
trained blob against the new blobs found. This tool has 4 options.
Color Distance Only: Compares findings according to color only. This
will only find exact matches to the colors found within the color
filter.
Difference: Compares the attributes of
the blobs found to those of the model. These will be listed on the Blob
Compare page. The comparison will be 0.00 if the finding is a perfect
match in that
category and will be 1.00 if it is not a perfect match. Any tool in the
Compare Method that compares attributes will demonstrate these results
within the Blob Compare page.
Normalized: Compares against the area and color of the first blob.
Ratio: Compares in terms of an attribute ratio, uses division to find a match.
Debug: This portion of the tool is for the programmers use, the user can ignore this option.
Sort Attribute:
The user may select an attribute by which to sort the information, or
the "no sort" option may be selected.
Sort Descending: If
checked, the blobs are sorted in descending order on the sort criteria.
Line Index Origin: Using
this part of the tool with the "line index vector" allows the user to
compare the blobs using the different portions of the blob or sets of
colors. For example it could measure from part 1 to part 2 or the
distance from part 1 to part 3, this would allow for more precise
measurements. The number set in the Line Index Origin is the starting
point position for these comparisons.
Line Index Vector: Use
this portion of the tool with the "Line Index Origin". The Vector
number is set to a number higher than zero to find a length as a
comparison. This will compare across the image for distance in
proportion to the rest of the image and area.
Count: The number of
new blobs found.
Points:
Indicates the number of points found.
Lines: Indicates
the number of lines found.
Render
Style Points: The user may choose what the appearance of
the newly found points will be.
Render Style Lines: The
user may choose what the appearance of the newly found lines will be.
Refresh Viewer: This
indicates how the image will appear in the viewer.
Example:
So, If these blob images were inputted into the Blob Search tool
and the tool was trained on the top pattern of red, black, and blue
blobs, then the image would be able to find the second blob image that
is located below the first set in out example picture. The user could
also set the parameters to measure across from teh red image to the
blue to get the distance from the origin by a vector (or to get other
measurements depending on how the user chooses to set up the
parameters). The tool would then allow comparisons to be done between
the two images for the user to get a complete picture of the blobs
found.