Part model geometric feature extraction method for neural network
A technology of geometric features and neural network, applied in the field of feature recognition of parts, can solve the problems of loss of fidelity and loss of key mapping of discretized representation, and achieve the effect of improving generalization ability and reliable recognition results.
Active Publication Date: 2022-04-08
ZHEJIANG UNIV
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Discretizing the representation loses fidelity and may lose key mappings back to the original B-rep entities
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0035] The present invention will be described in detail and clearly below with reference to the embodiments.
[0036] like figure 1 As shown, the embodiment of the present invention and its implementation process include the following:
[0037] The present invention includes the following steps:
[0038] Step 1: Establish a three-dimensional rectangular coordinate system for the part model. The x-axis direction of the three-dimensional rectangular coordinate system is horizontal to the right, the y-axis direction is vertical and upward, and the z-axis direction is determined by the right-hand rule. According to the data characteristics of the part model file, the geometric elements of the part model are divided into multiple lines and multiple surfaces;
[0039] Step 2: Sampling the lines and surfaces respectively for the lines and surfaces divided in step 1 to obtain discrete points corresponding to the lines and surfaces respectively, and perform feature parameter calcula...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a part model geometric feature extraction method for a neural network. Comprising the following steps: dividing geometric elements of a part model into a plurality of lines and a plurality of surfaces; performing line and surface sampling and characteristic parameter calculation on the divided lines and the divided surfaces to obtain corresponding geometrical characteristics, dividing the lines into straight lines, arcs and free curves according to the geometrical characteristics of the lines, and dividing the surfaces into planes, cylindrical surfaces, spherical surfaces and free curved surfaces according to the geometrical characteristics of the surfaces; and finally, performing feature parameter interpolation calculation and secondary sampling on the free curve and the free curved surface to obtain final geometric features of the free curve and the free curved surface, thereby obtaining geometric features of all lines and surfaces in the part model, and using the geometric features as input of a neural network to support subsequent application. According to the method, the local geometric features of the three-dimensional model are fully utilized, the generalization ability of the original neural network model is improved, and the recognition result of the free curve and the free curved surface is more reliable.
Description
technical field [0001] The invention relates to a feature extraction method in the field of part feature recognition, in particular to a method for extracting geometric features of a part model for a neural network. Background technique [0002] The industry standard for entity model representation is the boundary representation (B-rep). A B-rep is a versatile data structure consisting of faces, edges, and vertices, glued together by topological relationships between them. B-rep allows various parametric curves and surfaces to be accurately represented by CAD modeling operations. Part designers interact directly with B-rep faces, edges and vertices through the software to select, align and modify 3D shapes. To take advantage of recent advances in deep neural networks in CAD software, proper representation of B-rep data is required. [0003] Although B-rep data is widely used in the industry, the research that directly combines B-rep data and deep neural networks is still ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G05B19/4097
Inventor 冯毅雄岑鸿晋洪兆溪吴轩宇
Owner ZHEJIANG UNIV