Open drzraf opened 6 years ago
find-object is very features2d oriented, as it was the goal at the beginning. For blob detector, I think there are some examples to use out of the box, even in python: https://www.learnopencv.com/blob-detection-using-opencv-python-c/
If someone comes with an idea on how it could work with Find-Object, we may consider adding it. For example in this mode, "objects" would be "blobs" and detection would happen in FindObject::detect(): https://github.com/introlab/find-object/blob/1abda367d6bd4486aa6a8f332f05bc7e0d448296/src/FindObject.cpp#L1356 Instead of using features2d detector, SimpleBlobDetector may be used directly, returning a rectangle around the blob. This would be then compatible with all the code for ROS and TF.
Was it considered? That would be a good candidate to add if find-object detectors parameters management is flexible enough.