introduction
Reverse engineering, also called reverse engineering or reverse engineering, is an engineering design model that constructs a product or part based on an existing product or part prototype. Based on this, the existing product is analyzed, understood and improved. There is a design redesign.
In a broad sense, reverse engineering can be divided into the following three categories:
(1) Real object reversal: It is recreated by mapping and folding under the condition of the existing product; it includes the reverse of functional reverse, performance reverse, scheme, structure, material and so on. Objects that are reversed in kind can be complete machines, components, and components.
(2) Software reversal: Product samples, technical documents, design books, instruction manuals, drawings, relevant specifications and standards, management specifications and quality assurance manuals are called technical software. There are three types of software reversal: both physical and full-service software; only physical and no technical software; no physical, only full or partial technical software.
(3) Image reversal: The designer has neither physical products nor technical software. Only the pictures of the products, advertisements or impressions after the visit, etc., the designer must use these image materials to conceive and design the products. Reverse the image.
At present, the research on reverse engineering at home and abroad mainly focuses on the reverse of geometric shape, that is, the reconstruction of the physical CAD of the product, which is called "physical reverse engineering". Reverse engineering and forward engineering are shown in Figure l below:
Reverse engineering data measurement technology
Data measurement is the geometrical coordinate data of discrete points on the surface of the product obtained by specific measuring equipment and measurement methods to digitize the geometry of the product. The measuring principle is: placing the product to be tested in the measurement space of the coordinate measuring machine, and obtaining the coordinate position of each measuring point on the tested product, according to the spatial coordinate value of these points, after computer data processing, fitting and forming the measurement Elements, such as circles, spheres, cylinders, cones, surfaces, etc., are mathematically calculated to derive their shape, positional tolerances, and other geometric quantities. Efficient and high-precision acquisition of digital information of products is the basis and key to reverse engineering.
The existing data collection methods are mainly divided into two categories:
(1) Contact data acquisition method The contact data acquisition method includes trigger data acquisition and continuous scan data acquisition, magnetic field method, and ultrasonic method using force-based firing principle. Contact data acquisition usually uses a three-coordinate measuring machine. When measuring, the probe and its direction can be selected according to the characteristics of the physical object and the measurement requirements. The number of measuring points and its distribution are determined, and then the measured path is determined, and sometimes the collision is checked. The trigger data acquisition method uses a trigger probe, which is also called a switch probe. When the probe of the probe contacts the surface of the product, the probe is triggered by the deformation of the probe, and the current probe is recorded by the data acquisition system. Coordinate values, moving the probe point by point, can obtain the coordinate data of the surface contour of the product. Commonly used contact trigger probes mainly include: mechanical trigger probe, strain gauge trigger probe, piezoelectric ceramic trigger probe. The advantages of using the touch probe are: it is suitable for the measurement of the space box type workpiece and the surface of the known product; the trigger type probe has strong versatility and is suitable for size measurement and online application; it is small in size and easy to be in a small space. Application; since the measuring machine is in a uniform linear low speed state when measuring data points, the dynamic performance of the measuring machine has little influence on the measurement accuracy. However, due to the limitations of the probe, some details of the part to be tested cannot be measured, and some fragile and easily deformable parts cannot be measured. In addition, the contact measuring probe is in contact with the surface of the part, and the measuring speed is slow. After the measurement, the probe compensation is performed, and the amount of data is small, and the shape of the solid cannot be truly reflected.
(2) Non-contact data acquisition method Non-contact data acquisition method mainly uses optical principle to collect data, including: laser triangle method, laser ranging method, structured light method and image analysis method.
The non-contact data acquisition speed is fast and the precision is high, and the measurement error caused by the measurement friction force and the contact pressure is eliminated, and the problem of the inferior point caused by the curvature interference of the contact type probe and the measured surface is avoided, and the dense point cloud information obtained is obtained. The large amount and high precision, the spot generated by the probe can also be made small, and it is possible to detect the position that is difficult to measure by the general mechanical probe, and to reflect the true shape of the surface to be measured to the utmost extent. The non-contact data acquisition method uses a non-contact probe, which is suitable for measuring soft objects due to no force; the non-contact probe has a high sampling rate between 50 times/second and 23000 times/second, suitable for surface shape. Measurement of unknown surfaces with complex and inaccurate requirements, such as wood molds and mud molds for automobiles and home appliances. However, the non-contact probe is greatly affected by the surface characteristics of the object (color, luminosity, roughness, shape, etc.). In most cases, the measurement error is larger than that of the contact probe, and is maintained above 10 micrometers. This method is mainly used for the measurement of easily deformable parts, parts with low precision requirements, parts requiring massive data, and supporting conditions without considering the measurement cost and related hardware and software.
In short, in the case where contact measurement can be applied, do not use non-contact measurement; contact measurement should be used as far as possible only when measuring the size and positional elements; if the measurement cost is met and the requirements are met, contact type should be used as much as possible. Measurement; non-contact scanning measurement for high requirements on product contour and dimensional accuracy; scanning for measurement of off-point; for products with low deformation and low precision, parts requiring large measurement data Non-contact measurement methods are used for measurements.
A total of 2 pages.
Reverse engineering data processing technology
Data processing is an important technical link in reverse engineering, which determines whether the subsequent CAD model reconstruction process can be carried out conveniently and accurately. According to the number of measurement points, the measurement data can be divided into general data points and massive data points; according to the regularity of the measurement data, the measurement data can be divided into scattered data points and rule data points; the measurement data obtained by different measurement systems The format is inconsistent, and almost all measurement methods and measurement systems inevitably have errors. Therefore, the measurement data must be processed before the CAD reconstruction using the measurement data. The data processing work mainly includes: data format conversion, multi-view cloud flattening, point cloud filtering, data reduction and point cloud blocking.
Each CAD/CAM system has its own data format. The current data structure and format of popular CAD/CAM software are different, which not only affects the data transmission and program connection between design and manufacturing, but also directly affects CMM. Data communication with CAD/CAM systems. The current approach is to use several major data exchange standards (IGES, STEP, AutoCAD DXF, etc.) to achieve data communication.
In the actual process of reverse engineering, since the coordinate measurement has its own measurement range, no matter what measurement method we use, it is difficult to completely measure the geometric data of the product in the same coordinate system. The digitization of the product cannot be done in the same coordinate system. When the model is reconstructed, the data in these different coordinates must be unified into one coordinate system. This data processing process is multi-view data positioning alignment (multi-view cloud flattening). . The alignment of multi-view data can be mainly divided into two types: direct alignment of measurement data by dedicated measurement software devices; and alignment of data processing afterwards. The use of post-mortem data processing alignment can be further divided into: direct alignment of data and graphics-based alignment. In the study of direct alignment of data, various algorithms have emerged, such as ICP algorithm; quaternion method; SVD method; alignment method based on three reference points.
The purpose of data smoothing is to eliminate the noise of the measured data to obtain accurate data and good feature extraction. Standard Gaussian, average or median filtering algorithms are currently used. Among them, Gaussian filtering can better maintain the shape of the original data, and the median filtering has better effect of eliminating data glitch. Therefore, the filtering algorithm should be flexibly selected according to the data quality and modeling method when selecting.
In the process of modeling using point cloud data, the storage and processing of these point cloud data has become an unbreakable bottleneck due to the existence of massive data points. In fact, not all data points contribute to the reconstruction of the model. Therefore, the amount of data can be reduced and the point cloud data can be reduced while ensuring a certain accuracy. The methods currently used include: methods for reducing data by using uniform meshes; methods for reducing data points by reducing multi-deformed triangles; and methods for reducing polyhedral data points by using error bands.
Data segmentation is based on the types of sub-surfaces that make up the shape of the physical shape. Groups of data belonging to the same surface type are grouped into different data fields, which facilitates subsequent model reconstruction. The data segmentation method can be divided into two methods: measurement based segmentation and automatic segmentation. The current segmentation methods include: data segmentation method based on parametric quadratic approximation; automatic segmentation method for scattered data points; and data segmentation method based on CT technology.
Reverse model reconstruction
In the entire reverse engineering, the three-dimensional geometric model CAD reconstruction of the product is the most critical and complicated part. Because only the CAD model of the product can be obtained, we can carry out the subsequent manufacturing, rapid prototyping, virtual simulation manufacturing and redesign of the product. Before the model reconstruction, the designer not only needs to know the previous information such as the geometric characteristics of the product and the characteristics of the data, but also needs to understand the subsequent application problems such as structural analysis, processing and manufacturing molds, rapid prototyping and so on. The main modeling methods currently used are:
(1) Curve-fitting modeling: A polynomial function is used to approximate the original data by interpolation, and finally a sufficiently smooth surface is obtained. The curve is the basis of the surface. The commonly used model reconstruction method in reverse engineering is to first fit the data points into spline curves by interpolation or approximation, and then use the modeling software to complete the reconstruction of the surface patches. The advantage is that the principle is relatively simple, as long as the number of polynomials is high enough to obtain a satisfactory surface, but it is also easy to cause instability of the calculation, and the processing ability of the boundary is also relatively poor, generally used to fit a relatively simple surface.
(2) Directly fitting the surface of the curved surface This method directly fits the curved surface of the measured data points, and obtains the final curved surface model through transition, mixing and joining. Surface fitting modeling can handle both ordered and scattered data points. Algorithms are: B-spline surface interpolation based on ordered points; B-spline surface interpolation; B-spline surface approximation for arbitrary measurement points.
(3) Point data gridding The networked solid model usually connects data points into triangular patches to form a polyhedral solid model. Two simplified methods have been developed: based on the guaranteed initial geometry based on the given data points, the nodes and patches are repeatedly excluded, a new triangle is constructed, and finally the specified number of nodes is reached; the smallest node and patch are found. The smallest polyhedron.
Outlook
The research of reverse engineering has become increasingly attractive, and has made great achievements in data processing, surface piece fitting, geometric feature recognition, commercial professional software and coordinate measuring machine research and development. However, in practical applications, the whole process still requires a lot of human-computer interaction. The experience and quality of the operator directly affect the quality of the product. The smoothness of the automatically reconstructed surface is difficult to guarantee. The following key technologies will be the main development of reverse engineering. aspect:
(1) Data measurement: Develop special measurement equipment for reverse engineering, which can realize 3D digitization of product geometry with high speed and high precision, and can automatically measure and plan paths;
(2) Top processing of data: Develop and research a general-purpose data processing software for different types of measurement data, and improve and improve the current data processing algorithms;
(3) Surface fitting: capable of controlling the smoothness of the curved surface and enabling smooth splicing;
(4) Integration technology: Development of reverse engineering technologies including measurement technology, model reconstruction technology, network-based collaborative design and digital manufacturing technology. A total of 2 pages.
Reverse engineering data processing technology
Data processing is an important technical link in reverse engineering, which determines whether the subsequent CAD model reconstruction process can be carried out conveniently and accurately. According to the number of measurement points, the measurement data can be divided into general data points and massive data points; according to the regularity of the measurement data, the measurement data can be divided into scattered data points and rule data points; the measurement data obtained by different measurement systems The format is inconsistent, and almost all measurement methods and measurement systems inevitably have errors. Therefore, the measurement data must be processed before the CAD reconstruction using the measurement data. The data processing work mainly includes: data format conversion, multi-view cloud flattening, point cloud filtering, data reduction and point cloud blocking.
Each CAD/CAM system has its own data format. The current data structure and format of popular CAD/CAM software are different, which not only affects the data transmission and program connection between design and manufacturing, but also directly affects CMM. Data communication with CAD/CAM systems. The current approach is to use several major data exchange standards (IGES, STEP, AutoCAD DXF, etc.) to achieve data communication.
In the actual process of reverse engineering, since the coordinate measurement has its own measurement range, no matter what measurement method we use, it is difficult to completely measure the geometric data of the product in the same coordinate system. The digitization of the product cannot be done in the same coordinate system. When the model is reconstructed, the data in these different coordinates must be unified into one coordinate system. This data processing process is multi-view data positioning alignment (multi-view cloud flattening). . The alignment of multi-view data can be mainly divided into two types: direct alignment of measurement data by dedicated measurement software devices; and alignment of data processing afterwards. The use of post-mortem data processing alignment can be further divided into: direct alignment of data and graphics-based alignment. In the study of direct alignment of data, various algorithms have emerged, such as ICP algorithm; quaternion method; SVD method; alignment method based on three reference points.
The purpose of data smoothing is to eliminate the noise of the measured data to obtain accurate data and good feature extraction. Standard Gaussian, average or median filtering algorithms are currently used. Among them, Gaussian filtering can better maintain the shape of the original data, and the median filtering has better effect of eliminating data glitch. Therefore, the filtering algorithm should be flexibly selected according to the data quality and modeling method when selecting.
In the process of modeling using point cloud data, the storage and processing of these point cloud data has become an unbreakable bottleneck due to the existence of massive data points. In fact, not all data points contribute to the reconstruction of the model. Therefore, the amount of data can be reduced and the point cloud data can be reduced while ensuring a certain accuracy. The methods currently used include: methods for reducing data by using uniform meshes; methods for reducing data points by reducing multi-deformed triangles; and methods for reducing polyhedral data points by using error bands.
Data segmentation is based on the types of sub-surfaces that make up the shape of the physical shape. Groups of data belonging to the same surface type are grouped into different data fields, which facilitates subsequent model reconstruction. The data segmentation method can be divided into two methods: measurement based segmentation and automatic segmentation. The current segmentation methods include: data segmentation method based on parametric quadratic approximation; automatic segmentation method for scattered data points; and data segmentation method based on CT technology.
Reverse model reconstruction
In the entire reverse engineering, the three-dimensional geometric model CAD reconstruction of the product is the most critical and complicated part. Because only the CAD model of the product can be obtained, we can carry out the subsequent manufacturing, rapid prototyping, virtual simulation manufacturing and redesign of the product. Before the model reconstruction, the designer not only needs to know the previous information such as the geometric characteristics of the product and the characteristics of the data, but also needs to understand the subsequent application problems such as structural analysis, processing and manufacturing molds, rapid prototyping and so on. The main modeling methods currently used are:
(1) Curve-fitting modeling: A polynomial function is used to approximate the original data by interpolation, and finally a sufficiently smooth surface is obtained. The curve is the basis of the surface. The commonly used model reconstruction method in reverse engineering is to first fit the data points into spline curves by interpolation or approximation, and then use the modeling software to complete the reconstruction of the surface patches. The advantage is that the principle is relatively simple, as long as the number of polynomials is high enough to obtain a satisfactory surface, but it is also easy to cause instability of the calculation, and the processing ability of the boundary is also relatively poor, generally used to fit a relatively simple surface.
(2) Directly fitting the surface of the curved surface This method directly fits the curved surface of the measured data points, and obtains the final curved surface model through transition, mixing and joining. Surface fitting modeling can handle both ordered and scattered data points. Algorithms are: B-spline surface interpolation based on ordered points; B-spline surface interpolation; B-spline surface approximation for arbitrary measurement points.
(3) Point data gridding The networked solid model usually connects data points into triangular patches to form a polyhedral solid model. Two simplified methods have been developed: based on the guaranteed initial geometry based on the given data points, the nodes and patches are repeatedly excluded, a new triangle is constructed, and finally the specified number of nodes is reached; the smallest node and patch are found. The smallest polyhedron.
Outlook
The research of reverse engineering has become increasingly attractive, and has made great achievements in data processing, surface piece fitting, geometric feature recognition, commercial professional software and coordinate measuring machine research and development. However, in practical applications, the whole process still requires a lot of human-computer interaction. The experience and quality of the operator directly affect the quality of the product. The smoothness of the automatically reconstructed surface is difficult to guarantee. The following key technologies will be the main development of reverse engineering. aspect:
(1) Data measurement: Develop special measurement equipment for reverse engineering, which can realize 3D digitization of product geometry with high speed and high precision, and can automatically measure and plan paths;
(2) Top processing of data: Develop and research a general-purpose data processing software for different types of measurement data, and improve and improve the current data processing algorithms;
(3) Surface fitting: capable of controlling the smoothness of the curved surface and enabling smooth splicing;
(4) Integration technology: Development of reverse engineering technologies including measurement technology, model reconstruction technology, network-based collaborative design and digital manufacturing technology.
Roll Up Door Servo Motor,Automatic Roller Shutter Door Opener,Roller Shutter Door Servo System,Glass Sliding Door Motor
Shenzhen Hongfa Automatic Door Co., Ltd. , https://www.spiralfastdoor.com