CN-122024217-A - Digital denture processing system and method

CN122024217ACN 122024217 ACN122024217 ACN 122024217ACN-122024217-A

Abstract

The application provides a digital denture processing system and method, which are characterized in that a plurality of partial point cloud views of the surface of a target denture workpiece are obtained, point cloud classification is further carried out based on surface curvature characteristics to obtain characteristic point clouds of complex occlusal surfaces and non-characteristic point clouds of flat axial surfaces, multi-view registration of geometric characteristics is respectively carried out on the characteristic point clouds and the non-characteristic point clouds to obtain complete three-dimensional point clouds of the target denture workpiece, fine registration is carried out on the complete three-dimensional point clouds and the three-dimensional point clouds of a target denture design model based on the overlapping rate of the point clouds to obtain registration residual errors between the target denture workpiece and corresponding point pairs of the target denture design model, and error thermodynamic diagrams of the processing precision of the target denture workpiece are generated by carrying out normalization and color mapping on all the registration residual errors. By adopting the scheme of the application, the false tooth processing piece with complex mixed geometric characteristics can be subjected to high-efficiency, high-precision and result-visualized automatic processing error detection.

Inventors

Wang Shidai
CHEN XIN
LI JIA
SUN YINGXIN
YU XIN
LI QINGCHAO
WANG CHUNYI
WANG SIYA
SHAO YAN
ZHANG HAO
LI NAN
YAN PEITAO

Assignees

朝阳市卫生学校

Dates

Publication Date: 20260512
Application Date: 20260115

Claims (10)

1. The false tooth processing error detection method based on point cloud registration carries out error on-line detection on false tooth processing pieces in a digital false tooth processing system and is characterized by comprising the following steps: Performing multi-view scanning on a target denture workpiece, obtaining a plurality of local point cloud views on the surface of the target denture workpiece, determining the surface curvature characteristics of a dental crown of the target denture workpiece according to the plurality of local point cloud views, and performing point cloud classification based on the surface curvature characteristics to obtain a characteristic point cloud set of a complex occlusal surface and a non-characteristic point cloud set of a flat axial surface; Performing multi-view registration of the adaptive geometric characteristics on the characteristic point cloud set and the non-characteristic point cloud set respectively, and further fusing to obtain a complete three-dimensional point cloud of the target denture workpiece; Performing fine registration on the complete three-dimensional point cloud and the three-dimensional point cloud of the target denture design model based on the point cloud overlapping rate to obtain a rigid transformation matrix with the point cloud overlapped, and further determining a registration residual error between corresponding point pairs of the target denture workpiece and the target denture design model according to the rigid transformation matrix; And (3) generating an error thermodynamic diagram of the processing precision of the target denture workpiece by normalizing and mapping the registration residual errors.
2. The method of claim 1, wherein determining the surface curvature characteristics of the target denture workpiece crown from the multiple local point cloud views specifically comprises: Constructing curvature characteristic fields of surface geometric relief for each of the plurality of partial point cloud views; performing adaptive curvature threshold filtering based on all curvature characteristic fields to separate out a high curvature candidate point set; And extracting the surface curvature characteristics of the dental crowns of the target denture workpieces from the Gao Qulv candidate points.
3. The method of claim 1, wherein performing point cloud classification based on the surface curvature features to obtain a feature point cloud of a complex occlusal surface and a non-feature point cloud of a flat axial surface specifically comprises: performing binary segmentation on each piece of local point cloud view in the plurality of pieces of local point cloud views based on the surface curvature characteristics, and extracting candidate characteristic area point clouds; detecting three-dimensional feature points of the candidate feature area point cloud to generate a feature description point cloud with scale and rotation invariance; and performing density threshold segmentation on the characteristic description point cloud to obtain a characteristic point cloud set of the complex occlusal surface and a non-characteristic point cloud set of the flat axial surface.
4. The method of claim 1, wherein performing multi-view registration of the adaptive geometric characteristics of the feature point cloud set and the non-feature point cloud set, respectively, to further fuse a complete three-dimensional point cloud of the target denture workpiece comprises: Performing group optimization registration based on the local feature histogram on the feature face point cloud set to obtain a feature face registration result; Performing seed point cloud form iterative registration on the non-characteristic face point cloud set to obtain a non-characteristic face registration result; The feature surface registration result and the non-feature surface registration result are integrally spliced and aligned to obtain a fusion point cloud after fine alignment; And reconstructing a curved surface based on the fusion point cloud to generate a complete three-dimensional point cloud of the target denture workpiece.
5. The method of claim 1, wherein performing fine registration on the complete three-dimensional point cloud and the three-dimensional point cloud of the target denture design model based on a point cloud overlap ratio, the obtaining a rigid transformation matrix with point cloud overlap specifically comprises: Establishing a bidirectional space division tree index between the complete three-dimensional point cloud and the three-dimensional point cloud of the target denture design model to perform point cloud matching, and obtaining an initial matching point pair set; Determining the point cloud overlapping rate of the current registration state according to the initial matching point pair set; and iteratively updating the spatial pose of the complete three-dimensional point cloud based on the point cloud overlapping rate to solve spatial transformation parameters, thereby obtaining a rigid transformation matrix with the point cloud overlapping.
6. The method of claim 1, wherein determining a registration residual between the target denture workpiece and the corresponding point pair of the target denture design model from the rigid transformation matrix comprises: acting the rigid transformation matrix on the complete three-dimensional point cloud to perform space pose normalization to obtain a workpiece point cloud which is in the same coordinate system as a target denture design model; Solving a spatial nearest neighbor point for each point in the point cloud of the machined part in the three-dimensional point cloud of the target denture design model, and further constructing a bidirectional point-to-point mapping relation field; and determining a registration residual error between the target denture machined piece and the corresponding point pair of the target denture design model according to the point pair mapping relation field.
7. The method of claim 1, wherein generating an error thermodynamic diagram of the target denture workpiece machining accuracy by normalizing and color mapping all registration residuals comprises: Performing piecewise linear normalization on all registration residuals based on tolerance to obtain normalized deviation indexes; Mapping the normalized deviation index to a corresponding color coordinate based on a preset chromatographic mapping table, and further obtaining a three-dimensional point cloud model after color attachment; rendering the three-dimensional point cloud model through a three-dimensional graphic engine to generate an error thermodynamic diagram of the machining precision of the target denture workpiece.
8. A digital denture processing system comprising an error detection unit, said error detection unit comprising: The acquisition module is used for carrying out multi-view scanning on the target denture workpiece, obtaining a plurality of local point cloud views on the surface of the target denture workpiece, determining the surface curvature characteristics of the dental crowns of the target denture workpiece according to the plurality of local point cloud views, and further carrying out point cloud classification based on the surface curvature characteristics to obtain characteristic point clouds of the complex occlusal surface and non-characteristic point clouds of the flat axial surface; the processing module is used for performing multi-view registration of the adaptive geometric characteristics on the characteristic point cloud set and the non-characteristic point cloud set respectively, and further fusing the characteristic point cloud set and the non-characteristic point cloud set to obtain a complete three-dimensional point cloud of the target denture workpiece; the processing module is used for carrying out fine registration on the complete three-dimensional point cloud and the three-dimensional point cloud of the target denture design model based on the point cloud overlapping rate to obtain a rigid transformation matrix with the point cloud overlapped, and further determining a registration residual error between corresponding point pairs of the target denture workpiece and the target denture design model according to the rigid transformation matrix; and the execution module is used for generating an error thermodynamic diagram of the processing precision of the target denture workpiece by normalizing and mapping the registration residuals.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the point cloud registration-based denture processing error detection method according to any one of claims 1 to 7 when the computer program is executed.
10. A computer-readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the point cloud registration-based denture processing error detection method according to any one of claims 1 to 7.

Description

Digital denture processing system and method Technical Field The application relates to the technical field of digital denture processing, in particular to a digital denture processing system and a digital denture processing method. Background With the deep application of the digital manufacturing technology in the field of stomatology, denture processing is gradually changed from traditional manual manufacturing to a CAD/CAM-based digital design and manufacturing mode, in the mode, how to rapidly and accurately detect three-dimensional geometric precision of a finished denture physical part becomes a key link for ensuring clinical suitability and functionality of the denture, and currently, the industry generally relies on a three-dimensional scanning and point cloud registration technology to realize the comparison between a machined part and a design model so as to evaluate the processing error of the denture, thereby promoting the denture processing to develop to an intelligent and high-precision direction. The common denture processing precision detection method mainly depends on a point cloud registration technology, particularly an iterative nearest point algorithm and a variation thereof, the methods generally carry out global registration on the denture integral point cloud obtained by scanning and a design model point cloud, realize alignment and calculate deviation by minimizing corresponding point distances, however, as the denture geometric form is special, the surface of the denture is provided with a high-curvature complex characteristic area such as an occlusal surface and a low-curvature flat area such as an axial surface, the geometric characteristics of different areas are difficult to be considered in the traditional single registration strategy, the complex characteristic area is easy to cause registration failure or precision reduction due to insufficient characteristic points or noise interference, the flat area is possibly subject to mismatching or iterative convergence slow due to lack of significant characteristics, and meanwhile, the traditional method is mostly not used for carrying out characteristic distinction, so that the calculation burden and the efficiency of the registration process are heavy, the final error evaluation result is lack of visual and clear positioning visual expression, and the process correction is difficult to be guided rapidly, and therefore, the automatic processing error detection on the denture workpiece with complex mixed geometric characteristics is difficult to be faced in the industry. Disclosure of Invention The application provides a digital denture processing system and method, which can carry out high-efficiency, high-precision and result-visualized automatic processing error detection on denture workpieces with complex mixed geometric characteristics. In a first aspect, the present application provides a method for detecting an error in denture processing based on point cloud registration, the method comprising: Performing multi-view scanning on a target denture workpiece, obtaining a plurality of local point cloud views on the surface of the target denture workpiece, determining the surface curvature characteristics of a dental crown of the target denture workpiece according to the plurality of local point cloud views, and performing point cloud classification based on the surface curvature characteristics to obtain a characteristic point cloud set of a complex occlusal surface and a non-characteristic point cloud set of a flat axial surface; Performing multi-view registration of the adaptive geometric characteristics on the characteristic point cloud set and the non-characteristic point cloud set respectively, and further fusing to obtain a complete three-dimensional point cloud of the target denture workpiece; Performing fine registration on the complete three-dimensional point cloud and the three-dimensional point cloud of the target denture design model based on the point cloud overlapping rate to obtain a rigid transformation matrix with the point cloud overlapped, and further determining a registration residual error between corresponding point pairs of the target denture workpiece and the target denture design model according to the rigid transformation matrix; And (3) generating an error thermodynamic diagram of the processing precision of the target denture workpiece by normalizing and mapping the registration residual errors. In some embodiments, determining the surface curvature characteristics of the target denture workpiece crown from the multi-slice partial point cloud view specifically includes: Constructing curvature characteristic fields of surface geometric relief for each of the plurality of partial point cloud views; performing adaptive curvature threshold filtering based on all curvature characteristic fields to separate out a high curvature candidate point set; And extracting the surface curvature characteristics of the