The Image Recognition Transform can operate in one of five modes.  Once you have placed the Image Recognition Transform into the MxTransIt Harness, select the operating mode in the properties panel for the Transform.  The modes are summarized below.



Usage Examples

Pattern Recognition

Used to determine if a particular pattern (as defined by a Reference Image) exists in the ROI of the Captured Image.

If the ROI of  has the potential of displaying three bitmaps, use this mode to determine which bitmap was displayed.

This mode is also used in cases where text is displayed, and where the text (or textual alternatives) are known when the transform is configured.  For example, the Door Ajar message may be one of just a few alternatives that could appear in a ROI.  There is no need to use text recognition in this case.  It takes longer to set up and is likely to be less reliable.

Pattern XY Position

Used to measure movement (vertical and/or horizontal) of a pattern within the ROI of the Captured Image.

Use, for example, to determine movement of a memory seat or up/down movement of a door window.

Needle Recognition

Used to determine the value indicated by a needle on a gauge.

Reference Images are provided for the minimum and maximum needle position, along with the two numeric values represented by those images.  When operating in this mode the Transform delivers the numeric value indicated by the needle by interpolating the actual (recognized) needle position between the two reference positions.

Typically used for the speedometer, tachometer, or other dials on an instrument cluster. Can also be used to acquire a value from test equipment, such as the temperature from an environmental chamber.

Color Recognition

Provides a result indicating whether the average color of an ROI matches (within a tolerance) a reference color.

May be used, for example, to determine if a tell-tale is illuminated as either red or green.

Color Pattern Matching

Provides a result indicating whether the color and pattern of an ROI matches (within a tolerance) a reference color and pattern.

May be used, for example, to determine which tell-tale is illuminated and if it is red or green.

Text Recognition

Used to recognize displayed characters.  Character recognition is performed on the ROI of the Captured Image.  The recognized characters may be delivered by the transform in the form of either a Message Transition (with ASCII format) or a Numeric Transition.

In cases where a reference image can be created, it is often best to use Pattern Recognition mode, even if the pattern could be recognized as text or digits.  So you might use Pattern Recognition to recognize the words 'Door Ajar'.

However, if the pass/fail criteria are defined in the TestCase (For example: If the test is to send Outside Air temperature over the CAN bus.  Verify that the display shows the correct Outside Temperature), then in this case you would use Text Recognition to determine a numeric value for the displayed outside Air Temperature.

Region of Interest (ROI) - A part of an image where a particular recognition mode is applied.

Reference Image - An image that is saved at the time the recognition mode is configured. It is used for comparison purposes while the recognition function is performed.

Captured Image - The image that is captured as the test is running, and upon when the recognition algorithm is run.

Related Topics:

Pattern Recognition

XY Position Recognition

Needle Recognition

Color Recognition

Color Pattern Matching

Text Recognition

Image Recognition Overview