Vision-based Measurement And Reconstruction Of Orthodontic Archwires

A good external appearance can make oneself show confidence,be respected,and even gain some other additional benefits.However,oral malocclusion not only damages human appearance,but also directly threatens human health and is a common oral disease.At present,the treatment of oral malocclusion is mainly completed by wearing the corresponding appliance.As part of the orthodontic appliance,the bending accuracy of the archwire not only affects the orthodontic treatment effect;but also has an important impact on the treatment cycle and the stability after recovery.Traditional archwire bending is done by professional orthodontists,which relies heavily on clinical experience,takes a long time,and is difficult to achieve high-precision preparation.With the advent of the era of automation,people turn their attention to the automatic bending of robots.Due to the ductile nature of metal,springback will occur during machine bending.The use of an archwire measurement system can reduce rebound errors during robotic bending operations.In addition,the three-dimensional reconstruction of the archwire can facilitate orthodontic treatment design and plan evaluation,and it can also improve treatment results and reduce patient harm.In this paper,binocular vision technology is proposed to provide a visual inspection method for the automatic bending of archwires.The main work of this paper is as follows:Firstly,two structural models of binocular vision and their errors were analyzed,and a suitable structural model of binocular was determined according to the size,shape,and expected accuracy of the archwire.And the selection of industrial optical lens,industrial camera and other related experimental equipment was completed according to the shooting distance and required field of view size.The hardware experimental platform for the measurement of archwire parameters based on binocular vision system was built.Secondly,a theoretical analysis of the camera calibration is conducted,and a circular calibration plate is used to overcome the vulnerability of the traditional checkerboard calibration plate to noise,thus improving the calibration accuracy.After the calibration,the global optimization algorithm is used to further optimize the calibration results,which further improves the accuracy of the calibration.Thirdly,in the study of corner point detection,stereo matching algorithm,considering the characteristics of the archwire itself is small and lacks texture,a corner point detection method based on Freeman chain code is proposed,which no longer performs corner point detection directly from the image,but encodes the archwire contour into a digital curve,determines the suspected feature points through the chain code difference,and filters out the pseudo-feature points according to the relevant threshold criterion,and then obtains the The real archwire feature points are obtained.The obtained feature points are matched using a matching method based on epipolar constraints and combined with the parallax range for search matching.It is compared with the existing corner detection algorithm and stereo matching algorithm.The accuracy and effectiveness of the proposed method are proved.Finally,A measurement platform based on binocular vision is built,and the measurement research is carried out on this basis.The design of the software platform was completed.Mainly relying on Python language and Open CV library development platform,the tasks of camera calibration,image correction,feature point detection,stereo matching and measurement were completed.Finally,all modules were integrated together based on Py Qt5 to form a system.At the same time,the 3D reconstruction of the archwire was explored,and the classical 3D reconstruction method based on SFM and the proposed 3D reconstruction method based on thin objects were used to reconstruct the archwire,and the experimental results were analyzed.