A high-precision measurement method for the dimensions of ABS sheet molds

Abstract

The present application relates to the field of dimensional measurement technology, and specifically to a high-precision dimensional measurement method for an ABS sheet mold. The method comprises: scanning the surface of the ABS sheet mold twice to obtain all point cloud data and all sparse point cloud data, respectively, and constructing a three-dimensional mold model and a three-dimensional sparse model, respectively. The three-dimensional mold model is spatially calibrated and then divided into multiple component models; the spatial deviation, reflection deviation, adjustment coefficient, and adjusted filter radius of each component model are calculated, and the point cloud data on all component models are filtered using a filtering algorithm to obtain a reconstructed three-dimensional mold model, and the dimensions of the ABS sheet mold are measured. The present application effectively reduces the deviation caused by the reflectivity differences of different materials in the ABS sheet mold by dynamically adjusting the filtering radius of the filtering algorithm, thereby improving the precision and accuracy of the dimensional measurement of the sheet mold.

Description

Technical Field

[0001] The present application relates to the field of dimension measurement technology, and in particular to a method for high-precision dimension measurement of an ABS sheet mold. Background Art

[0002] The production process of ABS sheet materials uses ABS copolymer pellets as raw material. After heating in an extruder and rotating the screw, the molten ABS material is squeezed into a mold under a certain pressure to form a sheet blank, which is then cooled and shaped. During this process, the dimensional accuracy of the mold plays a decisive role in the accuracy and pass rate of the final sheet material.

[0003] When measuring the dimensions of a mold using traditional methods, the mold is reconstructed in 3D mainly through the 3D point cloud obtained by scanning, and then compared and aligned with the design CAD model. However, since ABS sheet molds contain components made of various materials, the surfaces of components made of different materials have different reflectivities, and the reflection intensity of the surface of components made of the same material is affected by factors such as processing accuracy and reflection angle, resulting in uneven reflection on the surface of components made of the same material. This difference in reflectivity and uneven reflection will cause distortion interference in the measurement of the 3D point cloud data, resulting in deviations in the measurement of the 3D point cloud, and thus a high reconstruction error during the 3D reconstruction of the mold, affecting the accuracy of the dimensional measurement of the sheet mold. Summary of the Invention

[0004] In order to solve the above technical problems, a high-precision measurement method for the dimensions of an ABS sheet mold is provided to solve the existing problems.

[0005] The solution to the technical problem of this application is to provide a high-precision dimensional measurement method for an ABS sheet mold, comprising the following steps:

[0006] Scan the ABS sheet mold surface twice to obtain all point cloud data and all sparse point cloud data, and construct a 3D mold model and a 3D sparse model respectively. The point cloud data includes position coordinates, color information, and reflection intensity.

[0007] Based on the centroid offset between the 3D sparse model and the 3D mold model, the 3D mold model is spatially calibrated using the 3D sparse model to obtain a calibrated 3D mold model; the calibrated 3D mold model is segmented into multiple component models based on the position coordinates and color information contained in all point cloud data on the calibrated 3D mold model;

[0008] Analyze the fluctuation and average level of the distance from the sparse point cloud distributed within the range of each component model to each component model, and calculate the spatial deviation of each component model;

[0009] The reflection deviation of each component model is determined by the extreme changes in the reflection intensity of all point clouds on each component model and the distance between the position coordinates of the extremely changed point clouds, combined with the discreteness of the reflection intensity;

[0010] The average level of reflection intensity of all point clouds on each component model, as well as the spatial deviation and reflection deviation, are analyzed to obtain the adjustment coefficient of each component model. The filter radius of the filtering algorithm is adjusted to obtain the adjusted filter radius corresponding to each component model. The point cloud data on all component models are filtered separately in combination with the filtering algorithm. The filtered point cloud data is used to obtain the reconstructed three-dimensional mold model, and the size of the ABS sheet mold is measured by point cloud thickness measurement.

[0011] Preferably, the steps of respectively acquiring all point cloud data and all sparse point cloud data, and respectively constructing a three-dimensional mold model and a three-dimensional sparse model include:

[0012] First, a global scan of the ABS sheet mold is performed to obtain the point cloud data corresponding to all scanning points on the mold surface, and a three-dimensional mold model is constructed using the point cloud data; multiple scanning points are selected from the mold surface as marking points, and the marking points are scanned point by point with high precision to obtain the sparse point cloud data corresponding to the marking points, and the sparse point cloud data is used to construct a three-dimensional sparse model.

[0013] Preferably, the further acquisition process of the calibrated three-dimensional mold model is: calculating the vector pointing from the center of mass of the three-dimensional mold model to the center of mass of the three-dimensional sparse model, recorded as the offset vector; performing vector addition on the position coordinates of each point cloud on the three-dimensional mold model and the offset vector, and spatially calibrating the three-dimensional mold model to obtain the calibrated three-dimensional mold model.

[0014] Preferably, the segmentation of the calibrated three-dimensional mold model into multiple component models includes: forming a feature vector from the position coordinates and color information of each point cloud on the calibrated three-dimensional mold model; clustering the feature vectors corresponding to all point clouds on the calibrated three-dimensional mold model, and segmenting the calibrated three-dimensional mold model into multiple component models based on the clustering results.

[0015] Preferably, the calculation of the spatial deviation of each component model includes: calculating the product of the discrete degree and the mean of the nearest distance from all sparse point clouds distributed in the area range contained in each component model in the calibrated three-dimensional mold model to each component model as the spatial deviation of each component model.

[0016] Preferably, determining the reflection deviation of each component model includes:

[0017] Calculate the discrete degree of reflection intensity of all point clouds on each component model, which is recorded as energy fluctuation;

[0018] The distance between the position coordinates of the point cloud corresponding to the maximum reflection intensity and the position coordinates of the point cloud corresponding to the minimum reflection intensity on each component model is recorded as the relative distance;

[0019] Calculating the range of reflection intensity of all point clouds on each component model; normalizing the ratio of the range to the relative distance and recording it as the relative change;

[0020] The reflection deviation is the product of the energy fluctuation and the relative change.

[0021] Preferably, obtaining the adjustment coefficient of each component model includes:

[0022] Calculating an average value of the reflection intensities of all point clouds on each component model, and normalizing the average value of all component models in the calibrated three-dimensional mold model;

[0023] Calculating a product of the spatial deviation and the reflection deviation;

[0024] The adjustment coefficient is a normalized value of the ratio between the product value of each component model and the normalized average value.

[0025] Preferably, the adjusted filter radius is the product of the adjustment coefficient of each component model and a preset maximum filter radius.

[0026] Preferably, obtaining the reconstructed three-dimensional mold model includes: performing three-dimensional reconstruction on all filtered point cloud data on all component models to obtain the reconstructed three-dimensional mold model.

[0027] Preferably, the measuring the size of the ABS sheet mold by point cloud thickness measurement includes: measuring the size of the reconstructed three-dimensional mold model by point cloud thickness measurement, and comparing and aligning it with the design CAD model of the ABS sheet mold to evaluate the dimensional accuracy of the ABS sheet mold.

[0028] This application has at least the following beneficial effects:

[0029] The present application scans the ABS sheet mold twice, once by performing a global scan of the point cloud, and the other time by selecting a sparse point cloud of some marked points for high-precision scanning, and constructs a three-dimensional mold model and a high-precision three-dimensional sparse model respectively. The beneficial effect is that the three-dimensional mold model can obtain the complete geometric details of the sheet mold surface, and the three-dimensional sparse model provides high-precision sparse point cloud data; the three-dimensional mold model is spatially calibrated using the high-precision three-dimensional sparse model, so that the three-dimensional sparse model and the three-dimensional mold model are at the same spatial coordinate position in space, and then the calibrated three-dimensional mold model is divided into multiple components according to different materials. The beneficial effect is that the coordinate system deviation or scanning cumulative error is corrected, and the size misjudgment caused by overall translation or rotation is avoided. Different materials are divided so that the point clouds on different components can be smoothed and corrected subsequently; the spatial deviation of each component model is calculated, and the beneficial effect is that the distance fluctuation between the high-precision sparse point cloud and the three-dimensional mold model is taken into account, reflecting the overall geometric distortion deviation of the component model, and evaluating the significance of local errors in the component model during the reconstruction process; the reflection deviation of each component model is determined, which is beneficial. The effect is that it considers the influence of uneven reflection on the surface of components of the same material on the point cloud position, reflects the dramatic change of reflection intensity under the component, and evaluates the distortion of uneven reflection on the surface of the component; secondly, the adjustment coefficient of each component model is obtained, the filter radius of the filtering algorithm is adjusted, and the adjusted filter radius corresponding to each component model is obtained. The point cloud data on all component models are filtered respectively in combination with the filtering algorithm. The filtered point cloud data is used to obtain a reconstructed three-dimensional mold model, and the size of the ABS sheet mold is measured by point cloud thickness measurement. Its beneficial effect is that by dynamically adjusting the filter radius of the filtering algorithm, the filtering intensity of the highly reflective component area is reduced to retain more detailed information, and the filtering intensity of the low-reflective component area is increased to make the surface reflection characteristics of the component model more uniform. By filtering and smoothing the point cloud data, the difference in reflection intensity and the distortion interference caused by uneven reflection are eliminated, making the reconstructed three-dimensional mold model closer to the actual mold geometry, which can effectively reduce the point cloud measurement deviation caused by the reflectivity difference of different materials in the ABS sheet mold, reduce the measurement impact of the mold size, and improve the precision and accuracy of the sheet mold size measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The following is a further detailed description of a high-precision dimensional measurement method for an ABS sheet mold of the present application in conjunction with the accompanying drawings.

[0031] Figure 1 A flowchart of the steps of a high-precision measurement method for the dimensions of an ABS sheet mold provided in an embodiment of the present application;

[0032] Figure 2A flowchart of a method for obtaining a reconstructed three-dimensional mold model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] To make the objectives, technical solutions, and advantages of this application more clearly understood, the following describes in further detail a method for high-precision dimensional measurement of an ABS sheet mold proposed in this application, in conjunction with the accompanying drawings and implementation examples. It should be understood that the specific embodiments described herein are intended only to explain this application and are not intended to limit this application.

[0034] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0035] See also Figure 1 , which shows a flowchart of a method for high-precision measurement of the dimensions of an ABS sheet mold provided by one embodiment of the present application, the method comprising the following steps:

[0036] Step 1: Scan the surface of the ABS sheet mold twice to obtain all point cloud data and all sparse point cloud data respectively, and construct a three-dimensional mold model and a three-dimensional sparse model respectively. The point cloud data includes position coordinates, color information, and reflection intensity.

[0037] By performing high-precision inspection on ABS sheet molds, the overall error of the sheet mold can be controlled within a small range, enabling the production of high-precision, uniform ABS sheet. However, since laser scanners typically measure point by point, large sheet molds require scanning a large number of points to cover the entire surface. This is time-consuming and leads to low scanning efficiency, making it difficult to meet the rapid measurement requirements of large-scale production environments. In addition, the laser scanning equipment itself is relatively expensive. Therefore, after quickly collecting the overall point cloud of the ABS sheet mold, a laser scanner is used to perform high-precision measurement of parts of the ABS sheet mold, simultaneously taking into account the accuracy of both the overall and local areas.

[0038] Based on the above analysis, a 3D scanner is used to perform a global scan of the ABS sheet mold to obtain point cloud data corresponding to all scanned points on the surface of the ABS sheet mold. The point cloud data includes position coordinates, color information, and reflection intensity. The ABS sheet mold is then modeled using the point cloud data corresponding to all scanned points to construct a 3D mold model.

[0039] It should be noted that the point cloud data corresponding to each scanning point contains information in multiple dimensions, such as spatial position (X, Y, Z), color information (R, G, B), and reflection intensity.

[0040] Secondly, multiple scanning points are selected on the surface of the ABS sheet mold as marking points. A high-precision laser scanner is used to scan the marking points one by one to obtain high-precision sparse point cloud data corresponding to all marking points. The sparse point cloud data corresponding to all marking points are modeled to construct a 3D sparse model.

[0041] In this embodiment, 3D modeling technology is used for modeling. 3D modeling technology is well known and will not be described in detail here. Secondly, on the surface of the ABS sheet mold, no fewer than four marking points are deployed per square meter. As other implementation methods, the implementer can set the number according to actual conditions.

[0042] It should be noted that the selected marking points are equivalent to the sparse point cloud on the three-dimensional mold model. Therefore, the marking points are scanned by a high-precision laser scanner to obtain high-precision sparse point cloud data corresponding to each marking point, so as to be used for spatial calibration of the three-dimensional mold model constructed by the point cloud data corresponding to the scanning points.

[0043] At this point, the three-dimensional mold model and three-dimensional sparse model of the ABS sheet mold are obtained.

[0044] Step 2: Based on the offset of the center of mass between the 3D sparse model and the 3D mold model, the 3D mold model is spatially calibrated using the 3D sparse model to obtain a calibrated 3D mold model; the calibrated 3D mold model is divided into multiple component models according to the position coordinates and color information contained in all point cloud data on the calibrated 3D mold model; the distance fluctuation and average level of the sparse point cloud distributed in the range of each component model to each component model are analyzed to calculate the spatial deviation of each component model.

[0045] Furthermore, the 3D sparse model is a sparse reconstruction of the entire 3D mold model. Ideally, the 3D positions of the marker points should coincide with the corresponding points on the 3D mold model. However, in the actual modeling process, the two reconstructed models may drift due to acquisition errors from different scanners. Since the marker points are scanned by a high-precision laser scanner with high measurement accuracy, the 3D sparse model is used to perform spatial calibration of the 3D mold model. Specifically,

[0046] Calculate the vector from the center of mass of the 3D mold model to the center of mass of the 3D sparse model, which is recorded as the offset vector;

[0047] The three-dimensional mold model is spatially calibrated by performing vector addition on the position coordinates of each point cloud on the three-dimensional mold model and the offset vector to obtain a calibrated three-dimensional mold model;

[0048] It should be noted that the calibrated three-dimensional mold model is placed in the same spatial coordinate system as the three-dimensional sparse model, thus avoiding high-precision measurement deviation caused by overall offset.

[0049] Secondly, ideally, the position coordinates corresponding to the marked points in the three-dimensional sparse model should coincide with the position coordinates of the scanning points at the corresponding positions in the calibrated three-dimensional mold model. However, in practice, since sheet molds are often composed of components made of different materials, they will show different color differences. For example, components with rubber hard brush heads often appear black; metal components that present support structures appear bright silver; and plastic components embedded in model detail structures appear white. Therefore, due to the different materials of different components, the surfaces of different materials on the ABS sheet mold have different abilities to reflect energy, which may lead to differences in the signal strength and quality received by the measuring equipment, making the position of the scanning point inaccurate or lost, resulting in deviations in the measurement of the scanning point, and further causing deviations in the accuracy of the model reconstruction of the mold surface.

[0050] Based on the above analysis, the point cloud segmentation of the calibrated 3D mold model is performed to obtain the clusters, specifically:

[0051] The position coordinates and color information of each point cloud on the calibrated three-dimensional mold model are combined into a feature vector;

[0052] Clustering the feature vectors of all point clouds on the calibrated 3D mold model, and dividing the calibrated 3D mold model into multiple component models;

[0053] It should be noted that point cloud data contains multidimensional information. In this embodiment, the point cloud data on the calibrated 3D mold model includes position coordinates, color information, and reflection intensity. Therefore, the position coordinates and color information contained in the point cloud data are extracted for clustering. Clustering is performed using the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm. The DBSCAN clustering algorithm is a well-known technology and will not be described in detail here. As other implementation methods, implementers may adopt other methods of existing technologies, such as hierarchical clustering algorithms, etc. This embodiment does not impose any special restrictions on this. It should be noted that the clustering algorithm is used to segment the 3D mold model into different components, with each cluster corresponding to a component on the ABS sheet mold. Therefore, the calibrated 3D mold model is segmented into multiple component models.

[0054] Ideally, the scanning points and marking points on the surface of each component model should basically coincide with each other. However, due to the distortion of the reflected energy of different scanning points during the scanning process, the component as a whole may have a certain spatial deviation. In this case, the accuracy of the measurement data of the component is low, and it is seriously affected by noise interference or distortion. Therefore, by analyzing the coincidence of the scanning points and marking points on each component model, the spatial deviation is calculated. Specifically,

[0055] The spatial deviation of each component model is calculated by calculating the product of the discrete degree and the mean of the nearest distance from all sparse point clouds distributed in the area to each component model based on the area range of each component model in the calibrated 3D mold model.

[0056] In this embodiment, the degree of dispersion is measured by calculating the variance of the nearest distances from all sparse point clouds distributed in the area to each component model. As other implementation methods, the implementer may adopt other methods of the prior art, such as standard deviation, etc. This embodiment does not impose any special restrictions on this. Secondly, the distance is measured by calculating the nearest Euclidean distance from all marked points distributed in the area to each component model. The calculation of the Euclidean distance is a well-known technology and will not be repeated here.

[0057] It should be noted that the large degree of discreteness indicates that the closest distance from the sparse point cloud corresponding to the marker point to the model of each component fluctuates violently, resulting in different offset deviations at different positions. The larger the spatial deviation, the larger the overall distortion deviation of the component model. The component has a large spatial deviation, which reflects that the more significant the geometric distortion caused by the difference in reflection intensity, the greater the reconstruction error of the sheet metal mold.

[0058] At this point, the spatial deviation of each component model is obtained.

[0059] Step 3: Determine the reflection deviation of each component model by combining the extreme changes in the reflection intensity of all point clouds on each component model and the distance of the position coordinates of the extremely changed point clouds with the discreteness of the reflection intensity.

[0060] Furthermore, due to the different surface processing precision of components made of the same material in different areas, uneven material composition, and different influences of reflection angles, the reflected energy at different locations on the surface of the same component will fluctuate. In other words, the more unstable the reflected energy at different locations on the same component, the more serious the interference of uneven reflection on the component. Therefore, by analyzing the changes in the reflection intensity of different point clouds on each component model, the reflection deviation is calculated, specifically:

[0061] Calculate the discrete degree of reflection intensity of all point clouds on each component model, which is recorded as energy fluctuation;

[0062] In this embodiment, the degree of discreteness is measured by calculating the variance of the reflection intensity of all point clouds on each component model. As other implementation methods, the implementer may adopt other methods of the prior art, such as standard deviation, etc. This embodiment does not impose any special restrictions on this.

[0063] Calculate the range of reflection intensity corresponding to all point clouds on each component model;

[0064] The distance between the position coordinates of the point cloud corresponding to the maximum reflection intensity and the position coordinates of the point cloud corresponding to the minimum reflection intensity on each component model is recorded as the relative distance;

[0065] In this embodiment, the distance is measured by calculating the Mahalanobis distance between the position coordinates of the point cloud corresponding to the maximum reflection intensity and the position coordinates of the point cloud corresponding to the minimum reflection intensity on each component model. The calculation of the Mahalanobis distance is a well-known technology and will not be repeated here. As other implementation methods, the implementer may adopt other methods of the existing technology, such as Euclidean distance, etc., and this embodiment does not impose any special restrictions on this.

[0066] The normalized result of the ratio of the range to the relative distance is recorded as the relative change;

[0067] In this embodiment, the sigmoid function is used for normalization processing, wherein the sigmoid function is a well-known technology and will not be described in detail here. As other implementation methods, the implementer may adopt other methods of the existing technology, such as the tanh function, etc. This embodiment does not impose any special restrictions on this.

[0068] It should be noted that the relative change reflects the degree of mutation of the reflection intensity per unit spatial distance. The larger the relative change, the greater the difference in reflection intensity within a small spatial range on the component model, and the greater the reflection deviation on the component model.

[0069] The product of the energy fluctuation and the relative change is used as the reflection deviation of each component model;

[0070] It should be noted that the greater the energy fluctuation, the more significant the fluctuation in the reflection intensity of different point clouds on the component model. The larger the obtained reflection deviation, the greater the fluctuation in the reflection intensity of the component model. In addition, this fluctuation is more dense in spatial distribution, reflecting that the greater the error in the point cloud data under the component model, the more concentrated the reflection distortion of the component.

[0071] At this point, the reflection deviation of each component model is obtained.

[0072] Step 4: Analyze the average level of reflection intensity of all point clouds on each component model, as well as the spatial deviation and reflection deviation, to obtain the adjustment coefficient of each component model, adjust the filter radius of the filtering algorithm, and obtain the adjusted filter radius corresponding to each component model. Combined with the filtering algorithm, filter the point cloud data on all component models separately. Use the filtered point cloud data to obtain the reconstructed 3D mold model, and measure the size of the ABS sheet mold through point cloud thickness measurement.

[0073] Furthermore, due to the influence of the reflection angle, a single component has a certain deviation in the reflection intensity at different positions, resulting in deviations in the measurement of the dimensions of a single component during the scanning and measurement process of the ABS sheet mold. At the same time, because different components are made of different materials, the energy reflection intensity of different materials varies during the scanning and measurement process. Often, the greater the reflection intensity of a component, the stronger its anti-interference ability, while the smaller the reflection intensity of a component, the more susceptible it is to interference from the reflectivity and the smaller its anti-interference ability.

[0074] Secondly, if the anti-interference ability of each component model is smaller during the scanning measurement process, a larger filter radius should be set when filtering the point cloud data on the component through the filtering algorithm to smooth out the sharp fluctuations in reflection intensity and make the reflection characteristics of the component model surface more uniform. Conversely, a smaller filter radius should be set to retain more detailed information. Therefore, based on the average level of reflection intensity corresponding to all point cloud data on each component model, as well as the spatial deviation and the reflection deviation, the adjustment coefficient is determined, specifically:

[0075] Calculating the average value of the reflection intensity of all point clouds on each component model; normalizing the average value of all component models in the calibrated three-dimensional mold model;

[0076] In this embodiment, the Z-Socre normalization method is used for normalization processing, wherein the Z-Socre normalization method is a well-known technology and will not be described here in detail. As other implementation methods, the implementer can adopt other methods of the existing technology, such as the maximum and minimum normalization method, etc. This embodiment does not impose any special restrictions on this.

[0077] Calculating a product of the spatial deviation and the reflection deviation, and using a normalized value of a ratio between the product of each component model and the normalized average value as an adjustment coefficient of each component model;

[0078] In this embodiment, the sigmoid function is used for normalization processing, wherein the sigmoid function is a well-known technology and will not be described in detail here. As other implementation methods, the implementer may adopt other methods of the existing technology, such as the tanh function, etc. This embodiment does not impose any special restrictions on this.

[0079] It should be noted that, the larger the average value, the greater the reflection intensity of different scanning points on the component, the stronger the anti-interference ability of the component, and the smaller the filtering radius of the filtering algorithm should be set; the larger the product value, the larger the measurement error of the point cloud data caused by the reflection distortion and geometric deviation of the component, and the larger the obtained adjustment coefficient, indicating that the suppression of the reflection distortion of the component should be increased, thereby indirectly correcting the point cloud data.

[0080] Furthermore, based on the adjustment coefficient, the filter radius of the filter algorithm corresponding to each component model is adjusted, specifically:

[0081] The product of the adjustment coefficient of each component model and the preset maximum filter radius is used as the adjusted filter radius corresponding to each component model;

[0082] In this embodiment, the preset maximum filtering radius is 0.2 m. For other implementations, the implementer may set it according to actual conditions.

[0083] Based on the adjusted filter radius, a filtering algorithm is used to filter all point cloud data on each component model to obtain filtered point cloud data;

[0084] In this embodiment, a PCL (Point Cloud Library) bilateral filtering algorithm is used for filtering. The PCL bilateral filtering algorithm is a well-known technology and will not be described in detail here.

[0085] Perform 3D reconstruction on all filtered point cloud data on all component models to obtain a reconstructed 3D mold model; measure the dimensions of the reconstructed 3D mold model through point cloud thickness measurement;

[0086] In this embodiment, a Poisson reconstruction algorithm is used for three-dimensional reconstruction. The Poisson reconstruction algorithm and point cloud thickness measurement are both well-known technologies and will not be described in detail here.

[0087] The flowchart of the method for obtaining the reconstructed three-dimensional mold model provided in the embodiment of the present application is as follows: Figure 2 shown.

[0088] The reconstructed 3D mold model is compared and registered with the designed CAD model of the ABS sheet mold to evaluate the dimensional accuracy of the ABS sheet mold.

[0089] It should be noted that the design CAD model is a 3D model created using CAD software based on the part's design drawings. It contains the part's precise design dimensions and shape information. By comparing the reconstructed 3D mold model with the design CAD model, the error between the actual sheet metal mold produced and the design drawings can be assessed.

[0090] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0091] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0092] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the present application. It should be noted that a person skilled in the art can make various modifications and improvements without departing from the spirit of the present application. Therefore, any simple modifications, equivalent variations, and modifications to the above embodiments made in accordance with the technical essence of the present application without departing from the content of the present application's technical solution fall within the scope of protection of the present application's technical solution.

Claims

1. A high-precision measurement method for the dimensions of an ABS sheet mold, characterized in that: The method comprises the following steps: Scan the ABS sheet mold surface twice to obtain all point cloud data and all sparse point cloud data, and construct a 3D mold model and a 3D sparse model respectively. The point cloud data includes position coordinates, color information, and reflection intensity. Based on the centroid offset between the 3D sparse model and the 3D mold model, the 3D mold model is spatially calibrated using the 3D sparse model to obtain a calibrated 3D mold model; the calibrated 3D mold model is segmented into multiple component models based on the position coordinates and color information contained in all point cloud data on the calibrated 3D mold model; Analyze the fluctuation and average level of the distance from the sparse point cloud distributed within the range of each component model to each component model, and calculate the spatial deviation of each component model; The reflection deviation of each component model is determined by the extreme changes in the reflection intensity of all point clouds on each component model and the distance between the position coordinates of the extremely changed point clouds, combined with the discreteness of the reflection intensity; The average level of reflection intensity of all point clouds on each component model, as well as the spatial deviation and reflection deviation, are analyzed to obtain the adjustment coefficient of each component model. The filter radius of the filtering algorithm is adjusted to obtain the adjusted filter radius corresponding to each component model. The point cloud data on all component models are filtered separately in combination with the filtering algorithm. The filtered point cloud data is used to obtain the reconstructed three-dimensional mold model, and the size of the ABS sheet mold is measured by point cloud thickness measurement.

2. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The steps of respectively acquiring all point cloud data and all sparse point cloud data, and respectively constructing a three-dimensional mold model and a three-dimensional sparse model include: First, a global scan of the ABS sheet mold is performed to obtain the point cloud data corresponding to all scanning points on the mold surface, and a three-dimensional mold model is constructed using the point cloud data; multiple scanning points are selected from the mold surface as marking points, and the marking points are scanned point by point with high precision to obtain the sparse point cloud data corresponding to the marking points, and the sparse point cloud data is used to construct a three-dimensional sparse model.

3. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The further acquisition process of the calibrated three-dimensional mold model is as follows: calculating the vector pointing from the center of mass of the three-dimensional mold model to the center of mass of the three-dimensional sparse model, recorded as the offset vector; performing vector addition on the position coordinates of each point cloud on the three-dimensional mold model and the offset vector, and spatially calibrating the three-dimensional mold model to obtain the calibrated three-dimensional mold model.

4. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The method of segmenting the calibrated three-dimensional mold model into multiple component models includes: forming a feature vector from the position coordinates and color information of each point cloud on the calibrated three-dimensional mold model; clustering the feature vectors corresponding to all point clouds on the calibrated three-dimensional mold model, and segmenting the calibrated three-dimensional mold model into multiple component models based on the clustering results.

5. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The calculation of the spatial deviation of each component model includes: calculating the product of the discrete degree and the mean of the nearest distance from all sparse point clouds distributed in the area contained in each component model in the calibrated three-dimensional mold model to each component model as the spatial deviation of each component model.

6. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: Determining the reflection deviation of each component model includes: Calculate the discrete degree of reflection intensity of all point clouds on each component model, which is recorded as energy fluctuation; The distance between the position coordinates of the point cloud corresponding to the maximum reflection intensity and the position coordinates of the point cloud corresponding to the minimum reflection intensity on each component model is recorded as the relative distance; Calculating the range of reflection intensity of all point clouds on each component model; normalizing the ratio of the range to the relative distance and recording it as the relative change; The reflection deviation is the product of the energy fluctuation and the relative change.

7. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The step of obtaining the adjustment coefficients of the component models includes: Calculating an average value of the reflection intensities of all point clouds on each component model, and normalizing the average value of all component models in the calibrated three-dimensional mold model; Calculating a product of the spatial deviation and the reflection deviation; The adjustment coefficient is a normalized value of the ratio between the product value of each component model and the normalized average value.

8. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The adjusted filter radius is the product of the adjustment coefficient of each component model and the preset maximum filter radius.

9. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The obtaining of the reconstructed three-dimensional mold model includes: performing three-dimensional reconstruction on all filtered point cloud data on all component models to obtain the reconstructed three-dimensional mold model.

10. The high-precision measurement method for the size of an ABS sheet mold according to claim 1, characterized in that: The measuring of the size of the ABS sheet mold by point cloud thickness measurement includes: measuring the size of the reconstructed three-dimensional mold model by point cloud thickness measurement, and comparing and aligning it with the design CAD model of the ABS sheet mold to evaluate the dimensional accuracy of the ABS sheet mold.

Images