Halcon Assistants

Image Acquisition

Matching

• Shape-Based Matching The shape-based matching describes the model by the shapes of contours instead of using the gray values of pixels and their neighborhood as template. In particular, shape-based matching should be chosen if occlusions or clutter can not be avoided or if a matching of objects with changing color is applied.
• Deformable Matching Like shape-based matching, the local deformable matching extracts contours and matches their shapes against the shapes of previously created models. But in contrast to shape-based matching, even slightly deformed objects are found and the deformations are returned as an additional result. This matching approach also allows to rectify the image part containing the de-formed object. Therefore, only the objects position is determined, whereas the orientation and scale
are interpreted as part of the deformation.
• Correlation-Based Matching The correlation-based matching approach is based on gray values. It uses a normalized cross correlation to evaluate the correspondence between a model and a test image.It can compensate both additive as well as multiplicative variations in illumination. In contrast to the shape-based matching, also objects with slightly changing shapes, lots of texture, or objects in blurred images (contours vanish in blurred images, e.g., because of defocus) can be found.
• Descriptor-Based Matching In contrast to the perspective deformable matching, the template is not built by the shapes of contours but by a set of so-called interest points. These points are first extracted and then classified according to their location and their local gray-value neighborhood. Similar to the perspective deformable matching, the descriptor-based matching is able to find objects even if they are perspectively deformed. Descriptor-based matching can be applied for images from a calibrated
camera as well as from an uncalibrated camera. This matching method provides a calibrated version with which also a 3D pose instead of 2D transformation parameters can be derived.

 

Measure

 

OCR