MpatFindModel

This function finds occurrences of the specified model in the given image and returns the position of each occurrence. The search is performed using the model's current search parameters. Note, this function assumes that the model has been preprocessed (MpatPreprocModel()).

You can limit the operation to a region of the image buffer, using a rectangular region of interest (ROI) set using MbufSetRegion(). Doing so will cause an increase in processing time, but it is more flexible than setting a search region using MpatSetPosition() because it allows a search at an angle and it is possible to define your search region in world units.

This function assumes that the model was extracted from a source image with the same scaling as the target image. In addition, if using a normalized grayscale non-rotational model, it assumes that the source image and the model's target image are of the same orientation (with approx. 5° tolerance). For angular search, use MpatSetAngle() to set the angular search range of the model.

The model's search parameters specify the maximum number of occurrences for which to look in the target image (MpatSetNumber()).

If a correlation has a match score greater than or equal to the certainty level (MpatSetCertainty()), it is automatically considered an occurrence (default 80%). The remaining occurrences will be the best of those greater than or equal to the acceptance level (MpatSetAcceptance()).

Results are written to the specified result buffer. Use the MpatGet...() functions to read the results. See MpatAllocResult() for more information on allocating result buffers.

This function returns results in decreasing match-score order. This means that the most-likely occurrence is always returned first.

Patterns that are very close to the edge of the image might be found with lower match scores than usual due to edge effects. For this reason, you should use an ROI set using MbufSetRegion() or a search region set using MpatSetPosition(), rather than a child buffer, to restrict the search to a portion of the image.