Aiming at the problem that the accuracy of traditional matching methods is low due to the different positions and directions of parts in image acquisition

in the template matching of metal parts

RTMM

a template matching algorithm using circular feature template

is proposed to solve. Firstly

the template is divided into several ring regions by a group of concentric circles

and the relationship between ring regions is used as the positioning feature of pre-selected regions quickly screen several pre-selected regions in matching images. Then

Sobel operator is used to search the stable points in each ring region

and according to the number of stable points contained in each ring region

the ring region that can replace the whole template image is selected. Finally

exact matching features are extracted from the ring region

and the pre-selected region is screened to get the exact matching results.