|
|
HOME
WHERE TO BUY
SITE MAP
LINKS
SEARCH
Open eVision
General Features
Evaluation
Learning
Development
EasyImage
EasyColor
EasyObject
EasyMatch
EasyFind
EasyGauge
EasyOCV
EasyOCR
EasyBarCode
EasyMatrixCode
Licensing System
|
New Acquisition Front Ends
|
ActiveGigE for Open eVision
ActiveDcam for Open eVision
Evaluation Programs
Data Sheet
|
|
|
Pattern Matching Library
|
|
Main Features
|
Typical Applications
|
|
|
Normalized correlation method
|
Presence, absence
|
|
Sub-pixel accuracy
|
Alignment
|
|
Rotation and scaling support
|
Pick and place
|
|
Multiple pattern occurrences
|
|
|
Gray-level and color images support
|
|
|
Non-square pixels management
|
|
|
Don't care areas
|
|
|
|
|
EasyMatch is a gray-level and color pattern matching library. It lets you train the system on a reference pattern and afterwards locate
its occurrences in other images. This tool is very convenient when the position of a given part is unknown in the field of view, or
if the presence of parts must be controlled. The library works by using normalized correlation method, i.e. measuring discrepancies
between the pattern and the target image.
|
|
|
|
Features
|
|
|
|
Multiple pattern occurrences
|
|
|
Several occurrences of a pattern, up to a user-defined number, are returned. Only the reliable ones are retained.
|
|
|
Standard, offset-normalized, gain-normalized and fully normalized correlation
|
|
|
The correlation is computed on continuous tone values (as opposed to binary). It is well known that
when the lighting conditions vary, as it is often the case, straight comparison of the pattern and
image behaves badly. To cope with this, automatically adjusting the contrast and/or intensity of the
pattern before comparison is very effective. This process is known as normalization. EasyMatch
provides four distinct normalization modes, depending on whether a gain and/or offset compensation
is used.
|
|
|
Normal, inverse or mixed contrast
|
|
|
Because of particular lighting effects, an object can appear with inverted contrast (white on black
instead of black on white or conversely). Depending on the application, it can be useful to keep
inverted instances or to disregard them. Three contrast modes are available: consider positive occurrences
only, negative occurrences only or both.
|
|
|
|
Translation, rotation and isotropic/anisotropic scaling
|
|
|
To find the best matches between the pattern and target image,
the target is allowed to translate horizontally and vertically. Additionally,
it can be allowed to rotate and/or to change its scale
in the X and Y directions simultaneously or independently. The
rotation angle and scale factors vary in a user-specified interval.
All degrees of freedom can be combined at will.
|
|
|
Variable accuracy, up to sub-pixel level
|
|
|
The accuracy with which the pattern is measured can be chosen
(the less accurate, the faster). A one tenth-of-a-pixel accuracy
can be achieved.
|
|
|
Don't care pixels
|
|
|
|
When the pattern cannot be inscribed in a rectangular ROI, the surrounding of the pattern can be
ignored by setting the pixels values below a threshold level. These pixels will not take part in the matching
process. The same feature can be used if parts of the template change from sample to sample.
|
|
|
Gray-level and color images
|
|
|
EasyMatch processes 8-bit gray-scale images as well as 24-bit RGB images.
|
|
|
Non-square pixels
|
|
|
When images are acquired with non-square pixels, rotated objects appear skewed. Taking the pixel
aspect ratio into account can compensate for this effect.
|
|
|
|
|
|
