Visual-based Method For Shape Feature Parameter Estimation Of Orthodontic Archwire During The Bending Process

Orthodontic archwire is an important appliance that affects orthodontic treatment,the continuous development of robot technology promotes the transition of orthodontic archwire appliance from traditional manual bending to automatic bending.In the process of automatic bending of archwire,due to its shape and material characteristics,archwire is prone to produce large shape error due to springback.In this paper,the shape parameter calculation based on machine vision technology is proposed to solve the above problems,and the shape parameter calculation method of orthodontic archwire in the process of bending operation is studied.In this paper,binocular vision is used to calculate the shape parameters of the archwire.Specifically,the obtained binocular image is preprocessed,and the skeleton of the archwire is extracted after segmentation and thinning.Then an improved feature point extraction algorithm is used to detect the feature points of the archwire.After detecting the feature points of the archwire,an improved random Hough transform method is proposed,which can accurately extract the feature segments of the archwire from the detected feature points.Finally,based on the polar constraint,a line feature matching method with matching similar discriminant function is used to realize the stereo matching of the line feature of the archwire binocular image,and the threedimensional parameters of the archwire shape are calculated according to the matching results.In this study,a binocular vision experimental platform was built,and the shape parameters of 10 groups of archwires were calculated by the above methods.The results show the effectiveness of this method.