MVTec’s HALCON makes use of superior subpixel-accurate matching technology which finds objects robustly and accurately in real-time. This works even if the objects are rotated, scaled, perspectively distorted, locally deformed, partially occluded or located outside of the image, or undergo non-linear illumination changes.
It has the ability to process images with 8 or 16 bits and also handles colour or multi-channel images. Objects can also be trained from images or from CAD-like data. HALCON’s unique component-based matching is able to locate objects that are composed of multiple parts that can move with respect to each other.
Check out the videos below which provide a step-by-step tutorial on using HALCON’s shaped-based matching.
An introductory overview showing the creation of shape models, finding objects in an image, and various parameters that can be adjusted in HDevelop to help improve robustness.
How to further process the results of HALCON’s shape based matching for counting, displaying the score, and aligning a region of interest based on the matching results. Applying transformation matrices to matching results are also demonstrated.
Learn about the advanced parameter settings for shape-based matching and their effects on speed and robustness.
Want to find out more about MVtec’s HALCON? Click here