An analysis of hole detection schemes
E. R. Davits, Simon Peter Barker · 1990
This paper studies the intricacies of hole detection, with an emphasis on the efficient location of small holes- the ultimate goal being real-time recognition and location of objects from their features. First, methods of locating circles are examined, and then the study is extended to hole location. A new form of chord-bisection scheme is found to be faster than the standard Hough transform for circle location, but the lateral histogram approach turns out to provide the most efficient available vieans for hole location. Object recognition and location perform a central role in image analysis, and are particularly important in applications such as automated visual inspection and automated assembly. The most obvious means of performing object location is by template matching (correlation),