Blob shape matching feature new machine vision version

The block analysis is the most common tool in the machine vision. Innomiles release the new version to enhance the function about aiming and filter the different shape by more speed.

The block function is important for the analysis the different characters. If we want to judge whether the two blocks are similar, we can usually distinguish them by the features of the area centroid.  To identify those area centroids are not sufficient to clearly distinguish between different blocks.   So if implement another feature can effectively picky up different block.   The Vision OVK Blob module proposes a new library of extended functions to assist the user in quickly extracting all the block shape features detected in the image and comparing them with the standard sample blocks one by one and sort it out even the size is enlarge or smaller in different are and similar degree.

The block shape feature has a rotation-independent feature, so the OVK Blob model can perform similarity evaluation regardless of the position or orientation of the block to be tested in the image.

Looking for the simple demonstration after we implement new algorithm

1) Various area and shape in omnidirectional angle.

2) Set up the standard one. It shows for match error=0

3) You could see the value of match error value are different for each shape.

The high value is 0.605. Then it could distinguish the different one.

Contact us for the new version release at