FindBlobs

FindBlobs is used to filter and score blobs in a Blobs structure by a weighted set of ideal values.

FindBlobs Overview

Using the FindBlobs parameter settings, a formula is created that calculates a score to measure how closely the blobs referenced in the Blobs data structure match an "ideal" blob. Based upon the parameter settings, FindBlobs returns only those blobs that match the specified requirements.

Each blob in the referenced Blobs data structure undergoes a two-step process of analysis. First, each of the individual FindBlobs parameter settings creates a formula that evaluates that particular parameter value. For instance, the Angle score would be computed by dividing the differences of the Range and the Angle value of the Angle parameter and the Range and the Angle value of the referenced blob, multiplied by the Weight value. After the individual scores have been computed, each of the individual parameter scores are added together and divided by the sum of the individual Weight parameter values. This score is then compared against the Accept Thresh value to determine whether or not the blob meets the specified requirements.

Once FindBlobs has filtered and sorted the extracted blobs, a new Blobs data structure and accompanying result table is automatically inserted into the spreadsheet. The result table is sorted by how closely the found blobs match the configured requirements, in descending order.

FindBlobs can be extremely useful in classifying features for defect detection. For example, if the job were to detect scratches on a surface, FindBlobs could be configured to return only those blobs with higher Elongation values. Or, if the goal was to detect jagged edge defects, FindBlobs could be configured to return only those blobs with higher Perimeter or Spread values.

Inputs

Parameter

Description

Blobs

This parameter uses Cell References to a Blobs data structure that is to be filtered and sorted.

Number to Find

This parameter specifies the maximum number (1 to 4096; default = 3) of blobs to return. If there are more blobs in the image whose scores exceed theAccept Threshsetting, the blobs with the highest scores will be selected.

Note: If the value of the Number to Findparameter is greater than zero, FindBlobs will automatically insert a result table containing a maximum of 10 entries that correspond to the first 10 blobs in the Blobs data structure. If you wish to display more entries, you can expand the table by copying the last row and pasting additional rows at the end; make sure to increment the index number of any newly-pasted row so that the proper blob is indexed.

Accept Thresh

Specifies the minimum acceptable match score (0 to 100; default = 10). FindBlobs will return only those blobs with match scores exceeding the Accept Thresh. The match score is computed with the following formula:

(Angle score + Area score + Elongation score + Holes score + Perimeter score + Spread score) * (100/Sum of the parameters' Weight values)

Angle

This parameter is used to configure the scoring formula that identifies the ideal Angle score (+/- CCW degrees) of found blobs.

Angle Scoring Formula

The FindBlobs Angle score is computed by using the following formula:

Angle Score = [(Range + Angle) - BlobAngleVal]/Range * WeightVal

BlobAngleVal = the referenced blob's output Angle value

WeightVal = Weight/(sum of all weight values) * 100

Angle	The ideal blob angle (-360 to 360; default = 0).
Range	The blob angle tolerance, in similarity scoring (0 to 360; default = 0).
Weight	The relative importance of blob angle, in similarity scoring (0 to 100; default = 0).

Area

This parameter is used to configure the scoring formula that identifies the ideal Area score of found blobs.

Area Scoring Formula

The FindBlobs Area score is computed by using the following formulas, depending on the BlobAreaVal and the Area value:

If the BlobAreaVal is greater than the Area value, this formula is used:

Area Score = [(Range + Area) - BlobAreaVal]/Range * WeightVal
If the BlobAreaVal is less than the Area value, this formula is used:

Area Score = [BlobAreaVal - (Area - Range)]/Range * Weight Val

BlobAreaVal = the referenced blob's output Area value

WeightVal = Weight/(sum of all weight values) * 100

Area	The ideal blob area size, in pixels (0 to 900000; default = 1000).
Range	The blob area tolerance, in similarity scoring (0 to 900000; default = 20000).
Weight	The relative importance of blob area, in similarity scoring (0 to 100; default = 50).

Elongation

This parameter is used to configure the scoring formula that identifies the ideal Elongation score of found blobs.

Elongation Scoring Formula

The FindBlobs Elongation score is computed by using the following formula:

Elongation Score = [(Range + Elongation) - BlobElongationVal]/Range * WeightVal

BlobElongationVal = the referenced blob's output Elongation value

WeightVal = Weight/(sum of all weight values) * 100

Elongation	The ideal blob elongation (0 to 900000; default = 1).
Range	The blob elongation tolerance, in similarity scoring (0 to 900000; default = 2).
Weight	The relative importance of blob elongation, in similarity scoring (0 to 100; default = 50).

Holes

This parameter is used to configure the scoring formula that identifies the ideal Holes score of found blobs.

Holes Scoring Formula

The FindBlobs Holes score is computed by using the following formula:

Holes Score = [(Range + Number) - BlobHolesVal]/Range * Weight Val

BlobHolesVal = the referenced blob's output Holes value

WeightVal = Weight/(sum of all weight values) * 100

Number	The ideal blob hole count (0 to 10000; default = 0).
Range	The blob hole count tolerance, in similarity scoring (0 to 10000; default = 2).
Weight	The relative importance of blob hole count, in similarity scoring (0 to 100; default = 0).

Perimeter

This parameter is used to configure the scoring formula that identifies the ideal Perimeter score of found blobs.

Perimeter Scoring Formula

The FindBlobs Perimeter score is computed by using the following formula:

Perimeter Score = [(Range + Perimeter) - BlobPerimeterVal]/Range * WeightVal

BlobPerimeterVal = the referenced blob's output Perimeter value

WeightVal = Weight/(sum of all weight values) * 100

Perimeter	The ideal blob perimeter (0 to 100000; default = 1000).
Range	The blob perimeter tolerance, in similarity scoring (0 to 100000; default = 10000).
Weight	The relative importance of blob perimeter, in similarity scoring (0 to 100; default = 50).

Spread

This parameter is used to configure the scoring formula that identifies the ideal Spread score of found blobs.

Spread Scoring Formula

The FindBlobs Spread score is computed by using the following formula:

Spread Score = [(Range + Spread) - BlobSpreadVal]/Range * WeightVal

BlobSpreadVal = the referenced blob's output Spread value

WeightVal = Weight/(sum of all weight values) * 100

Spread	The ideal blob spread (0 to 900000; default = 1).
Range	The blob spread tolerance, in similarity scoring (0 to 900000; default = 2).
Weight	The relative importance of blob spread, in similarity scoring (0 to 100; default = 50).

Show

Specifies the display mode for FindBlobs graphical overlays on top of the image.

0 = hide all (default)	All graphics will be hidden, except when the cell containing the FindBlobs function is highlighted in the spreadsheet.
1 = result graphics only	Blob outlines will be displayed at all times.

Outputs

Returns

A Blobs data structure containing the matching blobs sorted by score, or #ERR if any of the input parameters are invalid.

Results

When FindBlobs is initially inserted into a cell, a results table is created in the spreadsheet using the following Blobs Vision Data Access functions.

Note:

If the value of the Number to Find parameter is greater than zero, FindBlobs will automatically insert a result table containing a maximum of 10 entries that correspond to the first 10 blobs in the Blobs data structure. If you wish to display more entries, you can expand the table by copying the last row and pasting additional rows at the end. Make sure to increment the index number of any newly-pasted row so that the proper blob is indexed.
Additional data elements can be accessed by using the Blobs Vision Data Access functions.

Spreadsheet Label	Function	Description
Index		The indexed position of a blob(s).
X	GetX(Blobs, Blob Index)	Returns the x-coordinate of the center of mass (centroid) of the referenced blob.
Y	GetY(Blobs, Blob Index)	Returns the y-coordinate of the center of mass (centroid) of the referenced blob.
Angle	GetAngle(Blobs, Blob Index)	Returns the angle of the center of mass of the referenced blob, relative to the center of the ROI.
Color	GetColor(Blobs, Blob Index)	Returns the color value (0 = black, 1 = white) of the referenced blob.
Score	GetScore(Blobs, Blob Index)	A measure of how closely the blob matches the criteria of a FindBlobs function.
Area	GetArea(Blobs, Blob Index)	Returns the area of the referenced blob (measured in pixels). Note: If there are holes in the blob, the Area value will vary depending on whether or not the Fill Holes checkbox is enabled or disabled.
Elongation	GetElongation(Blobs, Blob Index)	Returns a value that represents how the pixels of the referenced blob are stretched out from the center of mass (centroid) of the blob. For instance, a circle reports 0 Elongation, while a nail has a high Elongation value.
Holes	GetHoles(Blobs, Blob Index)	Returns the number of holes contained within the referenced blob.
Perimeter	GetPerimeter(Blobs, Blob Index)	Returns the length of the boundary around the referenced blob; the perimeter is calculated by counting the external edges of the pixels that form the boundary of the blob.
Spread	GetSpread(Blobs, Blob Index)	Returns a value that represents how the pixels of the referenced blob are distributed away from the center of mass (centroid). For instance, a circular shaped object would have a lower Spread value than an oval shaped object.

See the Blobs Vision Data Access functions for details on the result table.