MpatAllocModel

This function allocates a model, using data from the specified area of the model's source image.

You can change a model's search parameter settings at any time, using the appropriate MpatSet...() function. When the model is no longer required, you should release its memory, using MpatFree().

An M_NORMALIZED type model (without overscan data) is usually used for an image with an inconsistent surrounding region, such as an image of loose nuts and bolts lying on a metal sheet. To perform an angular search, define the angular range in which the model can appear using MpatSetAngle(). When preprocessing an M_NORMALIZED model (without overscan data) for which an angular search range is specified, rotated versions of the model are created assigning "don't care" pixels to regions that do not have corresponding data in the original model.

Angular search is fastest when performed with an M_CIRCULAR_OVERSCAN model; however, M_CIRCULAR_OVERSCAN should only be used when the region around the model is consistent, for example, when searching for a chip in the image of an integrated circuit.

When preprocessing an M_CIRCULAR_OVERSCAN model for which an angular search range has been specified, a set of models is extracted from rotated versions of the M_CIRCULAR_OVERSCAN model, creating models that would appear upright if the target image were rotated. For this type of model, a larger region than the one defined is extracted from the model's source image so as to allow creation of models at different angles.

You can only define models that respect the following condition: (maxvalue ² * SizeX * SizeY) < 2 ⁶³ where maxvalue is the maximum pixel value (typically, 255) in the target image and the model. This restriction is imposed to avoid overflows in the internal 64-bit accumulators.

The total area of the defined model must be greater or equal to 4 pixels (SizeX * SizeY >= 4).