An automatic measuring device and method for the characteristic dimensions of irregular objects

Abstract

This invention relates to an automatic measurement device and method for the feature dimensions of irregular objects, belonging to the field of size measurement technology. The device includes a shape and size measuring instrument for performing point cloud scanning measurement on the object to be measured and obtaining point cloud data. It also includes: a robotic arm equipped with a vacuum suction cup; a conveyor belt positioned below the robotic arm; and a carrier plate for carrying the object to be measured, which is placed on the conveyor belt for transport. The method includes the following steps: Step 1: The robotic arm uses the vacuum suction cup to pick up the carrier plate carrying the object to be measured to send the object to the shape and size measuring instrument for point cloud scanning measurement and obtaining point cloud data; Step 2: The feature dimensions of the irregular object are defined as X, Y, and Z, where X, Y, and Z are the dimensions in the three orthogonal directions of the fragment; Step 3: After obtaining the point cloud of the irregular object, the point cloud is preprocessed to remove point cloud noise.

Description

Technical Field

[0001] This invention relates to the field of dimensional measurement technology, and in particular to an automatic measurement device and method for the characteristic dimensions of irregular objects. Background Technology

[0002] Traditional object size measurement often relies on manual methods, which are inefficient. For a large number of objects with completely irregular shapes and sizes, automated size measurement is not feasible. Currently, machine vision-based point cloud scanning measurement is gradually replacing manual methods in industrial settings. Current methods for scanning irregular objects require manually placing the object under the scanning system and then manually replacing it after scanning. After obtaining the 3D point cloud or mesh data of the object, the raw data must be imported into third-party software for manual surface fitting or intelligent size extraction, a cumbersome process. Furthermore, for completely irregular objects, such as fragments from explosions, it is often impossible to automatically fit their surfaces or edges, making it difficult to automatically calculate feature dimensions using third-party software.

[0003] In existing technologies, Wang Jiyang's "An Intelligent 3D Volume Measurement Device for Irregular Objects" from Shenyang University of Technology proposes an intelligent volume measurement device that can acquire a complete point cloud model through device operation, and then achieve volume measurement through subsequent segmentation, fitting, and calculation. However, the volume measurement calculation part is not described in detail and still requires manual fitting and calculation. Wang Tao et al.'s "A Method and Device for Measuring the Size of Irregular Objects" from Shandong Jierui Digital Technology Co., Ltd. establishes a corresponding identification algorithm for the target part through template parts, and measures its position and 3D model. This measurement algorithm is only for measuring a single part with a CAD model. Before each measurement, the corresponding identification algorithm for the part needs to be established. This method can only measure specific irregular objects and cannot measure random irregular objects, so it still has limitations.

[0004] In summary, current point cloud measurement methods for irregular objects often require manual assistance and cannot achieve automated dimensional measurement. This is especially true for large-scale industrial measurement of irregular objects, where it's impossible to obtain the object's feature dimensions through edge surface extraction and fitting, or identification methods. Manual measurement is inefficient and unsuitable for large-scale object measurement. Therefore, it is necessary to design an automated device and method for measuring the feature dimensions of irregular objects based on practical requirements, improving measurement efficiency and solving the challenge of automated measurement of large batches of irregular objects. Summary of the Invention

[0005] To address the technical problem that existing measurement techniques cannot meet the requirements for automatic measurement of point cloud dimensions of irregular objects, the present invention aims to provide an automatic measurement device and method for the feature dimensions of irregular objects. This device and method can automatically measure feature dimensions such as dimensions, cross-sectional area, and volume in three orthogonal directions, solving the problem of measuring the feature dimensions of irregular objects. Furthermore, by replacing manual fitting measurement with an automatic measurement algorithm, this device and method can achieve the measurement of the dimensions of a large number of random irregular objects, thereby improving measurement efficiency in industry.

[0006] To achieve the above-mentioned objectives, the present invention provides an automatic measurement device for the feature dimensions of irregular objects, including a shape and size measuring device for performing point cloud scanning measurement on the object to be measured and obtaining point cloud data, and further including:

[0007] A robotic arm, wherein a vacuum suction cup is provided on the robotic arm;

[0008] A conveyor belt, wherein the conveyor belt is disposed below the robotic arm; and

[0009] A sample carrier for supporting the object to be tested, the sample carrier being placed on the conveyor belt for transport;

[0010] The robotic arm uses the vacuum suction cup to pick up the sample plate carrying the object to be tested and deliver the object to be tested to the shape and size measuring device.

[0011] Furthermore, the vacuum suction cups are four in a circumferential distribution and supported by a bracket connected to the robotic arm. The four vacuum suction cups simultaneously adhere to the edge of the planar carrier plate.

[0012] Furthermore, a debris guard is provided between every two adjacent vacuum suction cups along the circumferential distribution direction of the four vacuum suction cups, and the debris guard is arranged only in three consecutive directions between the vacuum suction cups, leaving one direction without the debris guard.

[0013] Furthermore, the vacuum suction cup is connected to a vacuum device via a connecting pipe.

[0014] Furthermore, the suction cup material of the vacuum suction cup is made of nitrile rubber.

[0015] An automatic measurement method for the characteristic dimensions of irregular objects, implemented using the aforementioned automatic measurement device for the characteristic dimensions of irregular objects, includes the following steps:

[0016] Step 1: The robotic arm uses the vacuum suction cup to pick up the sample plate carrying the object to be tested and delivers the object to be tested to the shape and size measuring device to perform point cloud scanning measurement and obtain point cloud data;

[0017] Step 2: Define the characteristic dimensions of the irregular object, defining them as X, Y, and Z, where X, Y, and Z are the dimensions of the fragment in three orthogonal directions;

[0018] Step 3: After obtaining the irregular point cloud of the object to be tested by scanning, the point cloud needs to be preprocessed to remove point cloud noise and downsample the point cloud to reduce the number of points.

[0019] Step 4: Before performing calculations on the point cloud data, perform noise reduction processing on the point cloud by outlier filtering.

[0020] Step 5: In step 3, the point cloud is denoised by outlier filtering. A voxel-based filtering method is used to divide the point cloud into several voxels. The average density of points in each voxel is calculated. Voxels with a density less than a certain threshold are considered outliers and are removed.

[0021] Step 6: Based on the voxel division in Step 4, perform voxel downsampling on the point cloud for subsequent calculations. Use the centroid of all points in a voxel as the substitute for that voxel, thereby achieving downsampling of the original point cloud.

[0022] Step 7: The preprocessed point cloud can be used to calculate the feature size of the point cloud through this automatic feature size extraction algorithm.

[0023] Furthermore, the feature size calculation process for the point cloud is as follows:

[0024] Step 1: Construct a kd-tree and query the k nearest neighbors for each point. After the kd-tree is constructed, it can be used to perform a K-nearest neighbor search to obtain the k nearest neighbors for each point cloud point, which facilitates the calculation of the X dimension.

[0025] Step 2: Roughly calculate the X dimension. Based on the kd-tree established in Step 1, calculate the k nearest neighbor points {P} for each group sequentially. ij The centroid p of |j=1…k} i As representatives of the current k points, this reduces the computational data to 1 / k of the total data N after downsampling. The set of centroids after downsampling is: {p i Calculate the Euclidean distance d between all centroids in sequence |i=1…N / k}. ij After each calculation, the current distance is compared with the temporary distance, and the larger value replaces the smaller value, until all calculations are completed, and the maximum distance d between the centroids is obtained. max The corresponding two centroids are p1 and p2, and the calculation formula is as follows:

[0026]

[0027] Step 3: Calculate the X dimension. Using the two centroids p1 and p2 obtained in Step 2, find two sets of k nearest neighbors. Calculate the Euclidean distance between each original point sequentially, and use the comparison method to obtain the maximum distance. At this point, the exact feature dimension X is obtained. The calculation formula is as follows:

[0028] d 1i2j =distance(p 1i ,p 2j ), i1=1…k, i2=1…k

[0029] X = max({d 1i2j |i1=1…k,i2=1…k});

[0030] Step 4: Using the line connecting the two endpoints of dimension X as the x-axis, and one endpoint of dimension X as o', draw a perpendicular plane to the x-axis through point o' to obtain the y'o'z' plane. Project each data point of the obtained fragment scanning result onto the y'o'z' plane, calculate its convex hull point, and use the convex hull point to calculate the minimum bounding rectangle and the farthest point pair to obtain the feature dimension Y. At the same time, obtain the y-axis direction through the two endpoints of the feature dimension Y.

[0031] Step 5: Take a point on the x-axis as the origin o, construct the y-axis according to the y-axis direction obtained in Step 4, and obtain the z-axis by cross product of the two axes. Project each data point of the obtained fragment scanning result onto the xoz plane, calculate the length of the line connecting the projection points parallel to the z-axis in turn, and take the maximum value, which is the feature size Z.

[0032] Step 6: Calculate other characteristic quantities, including volume calculation:

[0033] For the acquired irregular point cloud, the volume is calculated using the slicing method. The scanned data is divided into n blocks at equal intervals (l) along a defined direction. Points near the blocks are projected onto the blocks, and then the points on the blocks are simplified. B-spline curves with control vertices are used to connect the simplified points, and the area A enclosed by the curves is calculated. i The total volume of the object is the product of the sum of the areas of all slices and the distance between the slices, calculated using the following formula:

[0034]

[0035] Furthermore, in step two, X represents the maximum distance between any two points within the fragment, Y represents the maximum distance in the plane perpendicular to the aforementioned maximum distance direction, and Z represents the maximum distance in the third orthogonal direction.

[0036] Furthermore, in the first process: during the construction of the kd-tree, the two-dimensional image coordinate points with depth values ​​are arranged in a binary tree manner, and the coordinate variances of the row and column dimensions are calculated respectively. The median coordinate point with the larger dimensional variance is used for vertical partitioning, dividing the image coordinate points into left and right point sets. The left and right subset spatial partitioning is performed recursively according to each data point, and the nodes of each partition and the parent nodes of each node are recorded, thereby establishing a kd-tree that facilitates point querying, until every spatial point is contained in the kd-tree. Nodes that no longer have child nodes are recorded as "leaf nodes".

[0037] Furthermore, in process six, the specific steps for volume calculation are as follows:

[0038] a. Select the point cloud data to be used for volume calculation and store it sorted by Z coordinate;

[0039] b. Read the point cloud data corresponding to the slice and store it in array A;

[0040] c. Project the points in array A onto the slice, and connect the simplified points using a B-spline curve that controls the vertices to form a closed curve;

[0041] d. Calculate the area of ​​the closed curve. Since a B-spline curve is fitted by multiple small line segments, we can treat the closed curve as a polygon to calculate its area. First, calculate the centroid of the polygon, then connect the centroid to each vertex of the polygon to form multiple triangles. The sum of the areas of all the triangles is the area of ​​the closed curve.

[0042] e. If array A is not empty, it means that the data in A may be able to fit another curve, so go to c.

[0043] f. If the data has not been completely read, proceed to step b; otherwise, end this process. The sum of the areas corresponding to all slices and the product of the slice spacing is the volume of the object.

[0044] Compared with the prior art, the present invention has the following advantages:

[0045] This invention, based on point cloud scanning measurement, combines an automatic measurement device and method for irregular object feature dimensions to achieve rapid and automatic measurement of the feature dimensions of irregular objects. It is suitable for measuring large batches of irregular objects in industrial settings. The automatic measurement device and method for irregular object feature dimensions comprises a robotic arm, a stage, and a vacuum suction cup, forming an automated measurement chain for irregular object feature dimensions. This chain, combined with a point cloud feature dimension measurement algorithm, can automatically measure the feature dimensions of point clouds of irregularly shaped objects in specific scenarios, obtaining geometric quantities such as the feature dimensions and volume of the point cloud to be measured.

[0046] This invention discloses an automatic measurement device and method for the feature dimensions of irregular objects. A novel automated measurement method is designed, which, through the automatic measurement device and irregular point cloud measurement algorithm, enables the automatic measurement of the feature dimensions of irregularly shaped objects of arbitrary shapes. It can automatically transport the object to be measured to the scanning point and directly calculate and output the feature dimensions of the irregular object. Compared with existing methods, the entire measurement process requires no manual intervention in placement and fitting measurements, and can obtain multiple dimensional data of the irregular point cloud in one go, saving manpower and time costs. Furthermore, compared with other methods, this measurement method is applicable to irregularly shaped objects of arbitrary shapes, has a wide range of applications, and is convenient and fast. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0048] Figure 1 This schematic diagram illustrates the structure of an automatic measurement device for the feature dimensions of irregular objects according to one embodiment of the present invention.

[0049] Figure 2 This schematic diagram illustrates the feature dimensions of a fragment in three orthogonal directions of an automatic measurement method for the feature dimensions of an irregular object according to an embodiment of the present invention.

[0050] Figure 3 This schematic diagram illustrates a flowchart of the feature dimension calculation method for an automatic measurement method of the feature dimensions of an irregular object in one embodiment of the present invention.

[0051] Figure 4 This diagram illustrates the convex hull point solution of an automatic measurement method for the feature dimensions of an irregular object according to one embodiment of the present invention.

[0052] Explanation of reference numerals in the attached figures:

[0053] 1. Robotic arm; 2. Vacuum suction cup; 3. Conveyor belt; 4. Loading plate; 5. Support frame; 6. Debris guard. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] See Figures 1 to 4 As shown; this embodiment discloses an automatic measurement device for the feature dimensions of irregular objects, including a shape and size measuring device for performing point cloud scanning measurement on the object to be measured and obtaining point cloud data, and further including:

[0056] Robotic arm 1, with a vacuum suction cup 2 mounted on it;

[0057] Conveyor belt 3 is positioned below robotic arm 1; and

[0058] The carrier plate 4 is used to support the object to be tested and is placed on the conveyor belt 3 for transportation.

[0059] In this process, the robotic arm 1 uses a vacuum suction cup 2 to pick up the sample plate 4 carrying the object to be measured, thereby sending the object to be measured to the shape and size measuring device.

[0060] During the measurement process, an irregular object is conveyed onto a carrier plate 4 via a conveyor belt 3. A robotic arm 1, in conjunction with a vacuum suction cup 2, adheres to the carrier plate 4, thus transferring the irregular object to be measured and forming an automated measurement chain. The robotic arm 1, by adsorbing the carrier plate 4 containing the object, automatically delivers the object to the shape and size measurement device. Then, the characteristic dimensions of the object are measured using an automatic irregular object size measurement method, thereby achieving automated measurement of irregular objects. The design of the vacuum suction cup 2 and the carrier plate 4 allows the measurement to be unrestricted by the shape of the object, enabling the measurement of objects of any shape.

[0061] Furthermore, there are four vacuum suction cups 2 arranged circumferentially and supported by a bracket 5. The bracket 5 is connected to the robotic arm 1. The four vacuum suction cups 2 simultaneously adhere to the edge of the planar carrier plate 4 to achieve balanced and stable adsorption of the carrier plate 4.

[0062] Furthermore, a debris guard 6 is provided between every two adjacent vacuum suction cups 2 along their circumferential distribution direction, and the debris guard 6 is arranged only in three consecutive directions between the vacuum suction cups 2, leaving one direction without a debris guard 6. The debris guard 6 avoids the risk of the test object falling during transportation. Leaving one direction without a debris guard 6 facilitates the recovery of the test object.

[0063] Furthermore, the vacuum suction cup 2 is connected to the vacuum equipment via a connecting pipe.

[0064] Furthermore, the suction cup material of vacuum suction cup 2 is made of nitrile rubber.

[0065] Vacuum chuck 2 is one of the actuators in the vacuum equipment. Made of nitrile rubber, it possesses significant tensile strength and is widely used in various vacuum holding devices. Before the system begins automated operation, vacuum chuck 2 is first connected to the vacuum equipment via a connecting pipe. Then, it comes into contact with the object to be lifted (in this case, the glass slide, i.e., slide 4, on which the fragment is placed). The vacuum equipment is activated, creating negative pressure within vacuum chuck 2, thus firmly holding the glass slide in place, allowing for the transfer of the glass slide. When the glass slide reaches its destination, air is smoothly injected into vacuum chuck 2, changing the negative pressure inside to zero pressure or a slightly positive pressure. Vacuum chuck 2 then detaches from the glass slide, completing the task of transferring the object to be measured.

[0066] An automatic measurement method for the characteristic dimensions of irregular objects, implemented using an automatic measurement device for the characteristic dimensions of irregular objects, includes the following steps:

[0067] Step 1: The robotic arm 1 uses the vacuum suction cup 2 to pick up the carrier plate 4 carrying the object to be measured and send the object to be measured to the shape and size measuring device to perform point cloud scanning measurement and obtain point cloud data. This step is achieved by an automatic measuring device for the feature size of irregular objects.

[0068] Step 2: Define the feature dimensions of the irregular object. The selection of feature dimensions usually depends on the actual application. In most application scenarios, the size of the object to be measured is defined by the outer envelope size of the irregular object. The feature dimensions are defined as X, Y, and Z, where X, Y, and Z are the dimensions of the fragment in the three orthogonal directions.

[0069] Step 3: After obtaining the irregular point cloud of the object to be measured through scanning, the point cloud first needs to be preprocessed to remove point cloud noise and downsample the point cloud to reduce the number of points, which facilitates the subsequent calculation of feature dimensions and improves the accuracy of dimension calculation.

[0070] Step 4: In actual engineering, the 3D point cloud data obtained often contains a certain amount of noise. Therefore, before calculating the point cloud data, the point cloud is denoised by outlier filtering. Outliers are points in the point cloud that are significantly different from the surrounding points. These points are usually noise or other irrelevant points, so they need to be removed by filtering.

[0071] Step 5: In Step 3, outlier filtering is used to denoise the point cloud. A voxel-based filtering method is adopted to divide the point cloud into several voxels. The average density of points in each voxel is calculated. Voxels with a density less than a certain threshold are considered outliers and are removed. At the same time, the point cloud obtained by scanning is often very dense and has a large number of points, which is not conducive to the calculation of feature size and consumes a lot of computing time and resources. Therefore, based on the above voxel division, voxel downsampling is performed on the point cloud to facilitate subsequent calculations.

[0072] Step 6: Based on the voxel division in Step 4, perform voxel downsampling on the point cloud for subsequent calculations. Use the centroid of all points in a voxel as the substitute for that voxel, thereby achieving downsampling of the original point cloud.

[0073] Step 7: The preprocessed point cloud can be used to calculate the feature size of the point cloud through this automatic feature size extraction algorithm.

[0074] Furthermore, the feature size calculation process for point clouds is as follows:

[0075] Step 1: Construct a kd-tree and query the k nearest neighbors for each point. After the kd-tree is constructed, it can be used to perform a K-nearest neighbor search to obtain the k nearest neighbors for each point cloud point, which facilitates the calculation of the X dimension.

[0076] Step 2: Roughly calculate the X dimension. Based on the kd-tree established in Step 1, calculate the k nearest neighbor points {P} for each group sequentially. ij The centroid p of |j=1…k} i As representatives of the current k points, this reduces the computational data to 1 / k of the total data N after downsampling. The set of centroids after downsampling is: {p i Calculate the Euclidean distance d between all centroids in sequence |i=1…N / k}. ij After each calculation, the current distance is compared with the temporary distance, and the larger value replaces the smaller value, until all calculations are completed, and the maximum distance d between the centroids is obtained. max The corresponding two centroids are p1 and p2, and the calculation formula is as follows:

[0077]

[0078] Step 3: Calculate the X dimension. Using the two centroids p1 and p2 obtained in Step 2, find two sets of k nearest neighbors. Calculate the Euclidean distance between each original point in turn, and use the comparison method to obtain the maximum distance. At this point, the exact feature dimension X is obtained. The calculation formula is as follows:

[0079] d 1i2j =distance(p 1i ,p 2j), i1=1…k, i2=1…k

[0080] X = max({d 1i2j |i1=1…k,i2=1…k});

[0081] Step 4: Using the line connecting the two endpoints of dimension X as the x-axis, and one endpoint of dimension X as o', draw a perpendicular plane to the x-axis through point o' to obtain the y'o'z' plane. Project each data point of the obtained fragment scanning result onto the y'o'z' plane and calculate its convex hull point. The convex hull of the point set refers to a minimum convex polygon, which satisfies that the points in the point set are either on the edge of the polygon or inside it. Since the two farthest points on a two-dimensional plane must be on the convex hull, use the convex hull point to calculate the minimum enclosing rectangle and the farthest point pair to obtain the feature dimension Y. At the same time, obtain the y-axis direction through the two endpoints of the feature dimension Y.

[0082] Step 5: Take a point on the x-axis as the origin o, construct the y-axis according to the y-axis direction obtained in Step 4, and obtain the z-axis by cross product of the two axes. Project each data point of the obtained fragment scanning result onto the xoz plane, calculate the length of the line connecting the projection points parallel to the z-axis in turn, and take the maximum value, which is the feature size Z.

[0083] Step 6: Calculate other characteristic quantities, including volume calculation:

[0084] For the acquired irregular point cloud, the volume is calculated using the slicing method. The scanned data is divided into n blocks at equal intervals (l) along a defined direction. Points near the blocks are projected onto the blocks, and then the points on the blocks are simplified. B-spline curves with control vertices are used to connect the simplified points, and the area A enclosed by the curves is calculated. i The total volume of the object is the product of the sum of the areas of all slices and the distance between the slices, calculated using the following formula:

[0085]

[0086] Furthermore, in step two, X represents the maximum distance between any two points within the fragment, Y represents the maximum distance in the plane perpendicular to the aforementioned maximum distance direction, and Z represents the maximum distance in the third orthogonal direction. The three dimensions are illustrated in the diagram below. Figure 2 As shown.

[0087] Furthermore, in process one: during the construction of the kd-tree, the two-dimensional image coordinate points with depth values ​​are arranged in a binary tree manner. The coordinate variances of the row and column dimensions are calculated separately. The median coordinate point with the larger dimensional variance is used for vertical partitioning, dividing the image coordinate points into left and right point sets. The left and right subset spatial partitioning is performed recursively according to each data point. At the same time, the nodes of each partition and the parent nodes of each node are recorded, thereby establishing a kd-tree that facilitates point querying, until every spatial point is contained in the kd-tree. Nodes that no longer have child nodes are recorded as "leaf nodes".

[0088] Furthermore, in process six: the specific steps for volume calculation are as follows:

[0089] a. Select the point cloud data to be used for volume calculation and store it sorted by Z coordinate;

[0090] b. Read the point cloud data corresponding to the slice and store it in array A;

[0091] c. Project the points in array A onto the slice, and connect the simplified points using a B-spline curve that controls the vertices to form a closed curve;

[0092] d. Calculate the area of ​​the closed curve. Since a B-spline curve is fitted by multiple small line segments, we can treat the closed curve as a polygon to calculate its area. First, calculate the centroid of the polygon, then connect the centroid to each vertex of the polygon to form multiple triangles. The sum of the areas of all the triangles is the area of ​​the closed curve.

[0093] e. If array A is not empty, it means that the data in A may be able to fit another curve, so go to c.

[0094] f. If the data has not been completely read, go to b; otherwise, end this process. The sum of the areas of all slices and the product of the slice spacing is the volume of the object.

[0095] In summary, this invention, based on point cloud scanning measurement and combined with an automatic measurement device and method for the feature dimensions of irregular objects, enables rapid and automatic measurement of the feature dimensions of irregular objects, making it suitable for the measurement of large batches of irregular objects in industrial settings. The automatic measurement device and method for the feature dimensions of irregular objects comprises a robotic arm, a stage, and a vacuum suction cup, forming an automated measurement chain for the feature dimensions of irregular objects. This chain can automatically measure the feature dimensions of point clouds of irregularly shaped objects in specific scenarios, obtaining geometric quantities such as the feature dimensions and volume of the point cloud to be measured.

[0096] This invention discloses an automatic measurement device and method for the feature dimensions of irregular objects. A novel automated measurement method is designed, which, through the automatic measurement device and irregular point cloud measurement algorithm, enables the automatic measurement of the feature dimensions of irregularly shaped objects of arbitrary shapes. It can automatically transport the object to be measured to the scanning point and directly calculate and output the feature dimensions of the irregular object. Compared with existing methods, the entire measurement process requires no manual intervention in placement and fitting measurements, and can obtain multiple dimensional data of the irregular point cloud in one go, saving manpower and time costs. Furthermore, compared with other methods, this measurement method is applicable to irregularly shaped objects of arbitrary shapes, has a wide range of applications, and is convenient and fast.

[0097] The contents not described in detail in this invention are common knowledge to those skilled in the art.

[0098] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0099] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 a process or multiple processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment to implement the process. Figure 1a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0101] It should also be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0102] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. An automatic measurement method for the characteristic dimensions of irregular objects, characterized in that, The following steps are involved: Step 1: The robotic arm (1) uses a vacuum suction cup (2) to pick up the sample plate (4) carrying the object to be measured, so as to send the object to be measured to the shape and size measuring device, perform point cloud scanning measurement and obtain point cloud data; Step 2: Define the characteristic dimensions of the irregular object. Define the characteristic dimensions as X, Y, and Z, where X, Y, and Z are the dimensions of the fragment in three orthogonal directions; X represents the maximum distance between any two points within the fragment, Y represents the maximum distance in the plane perpendicular to the above maximum distance direction, and Z is the maximum distance in the third orthogonal direction. Step 3: The preprocessed point cloud is used to calculate the feature size of the point cloud through an automatic feature size extraction algorithm; The feature size calculation process for the point cloud is as follows: Step 1: Construct a kd-tree and query the k nearest neighbors for each point. After the kd-tree is constructed, it can be used to perform a K-nearest neighbor search to obtain the k nearest neighbors for each point cloud point, which facilitates the calculation of the X dimension. Step 2: Roughly calculate the X dimension. Based on the kd-tree established in Step 1, calculate the k nearest neighbor points {P} for each group sequentially. ij The centroid p of |j=1…k} i As representatives of the current k points, this reduces the computational data to 1 / k of the total data N after downsampling. The set of centroids after downsampling is: {p i |i=1…N / k};Calculate the Euclidean distance d of all centroids in sequence. ij After each calculation, the current distance is compared with the temporary distance, and the larger value replaces the smaller value, until all calculations are completed, and the maximum distance d between the centroids is obtained. max The corresponding two centroids are p1 and p2, and the calculation formula is as follows: Step 3: Calculate the X dimension. Using the two centroids p1 and p2 obtained in Step 2, find two sets of k nearest neighbors. Calculate the Euclidean distance between each original point sequentially, and use the comparison method to obtain the maximum distance. At this point, the exact feature dimension X is obtained. The calculation formula is as follows: d 1i2j =distance(p 1i ,p 2j ),i1=1…k,i2=1…k X=max({d 1i2j |i1=1…k,i2=1…k}); Step 4: Using the line connecting the two endpoints of dimension X as the x-axis, and one endpoint of dimension X as o', draw a perpendicular plane to the x-axis through point o' to obtain the y'o'z' plane. Project each data point of the obtained fragment scanning result onto the y'o'z' plane, calculate its convex hull point, and use the convex hull point to calculate the minimum bounding rectangle and the farthest point pair to obtain the feature dimension Y. At the same time, obtain the y-axis direction through the two endpoints of the feature dimension Y. Step 5: Take a point on the x-axis as the origin o, construct the y-axis according to the y-axis direction obtained in Step 4, and obtain the z-axis by cross product of the two axes. Project each data point of the obtained fragment scanning result onto the xoz plane, calculate the length of the line connecting the projection points parallel to the z-axis in turn, and take the maximum value, which is the feature size Z. Step 6: Calculate other characteristic quantities, including volume calculation: For the acquired irregular point cloud, the volume is calculated using the slicing method. The scanned data is divided into n blocks at equal intervals *l* along a defined direction. Points near the blocks are projected onto the blocks, and then the points on the blocks are simplified. The simplified points are connected using B-spline curves with control vertices, and the area A enclosed by the curves is calculated. i The total volume of the object is the product of the sum of the areas of all slices and the distance between the slices, calculated using the following formula:

2. The method for automatically measuring the characteristic dimensions of irregular objects according to claim 1, characterized in that, Following step two, the following steps are also included: Step A. After obtaining the irregular point cloud of the object to be tested by scanning, the point cloud first needs to be preprocessed to remove point cloud noise and downsample the point cloud to reduce the number of points. Step B. Before performing calculations on the point cloud data, the point cloud is denoised by outlier filtering. Step C. In step A, the point cloud is denoised by outlier filtering. A voxel-based filtering method is used to divide the point cloud into several voxels. The average density of points in each voxel is calculated. Voxels with a density less than a certain threshold are considered outliers and are removed. Step D. Based on the voxel division in Step B, voxel downsampling is performed on the point cloud for subsequent calculations. The centroid of all points in a voxel is used as the substitute for that voxel, thereby realizing the downsampling of the original point cloud.

3. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 2, characterized in that, In the first process, during the construction of the kd-tree, the two-dimensional image coordinate points with depth values ​​are arranged in a binary tree manner. The coordinate variances of the row and column dimensions are calculated separately. The median coordinate point with the larger dimensional variance is used for vertical partitioning, dividing the image coordinate points into left and right point sets. The left and right subset spatial partitioning is performed recursively according to each data point. At the same time, the nodes of each partition and the parent nodes of each node are recorded, thereby establishing a kd-tree that facilitates point querying, until every spatial point is contained in the kd-tree. Nodes that no longer have child nodes are recorded as "leaf nodes".

4. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 2, characterized in that, In process six, the specific steps for volume calculation are as follows: a. Select the point cloud data to be used for volume calculation and store it in sorted order by Z coordinate; b. Read the point cloud data corresponding to the slice and store it in array A; c. Project the points in array A onto the slice, and connect the simplified points using a B-spline curve that controls the vertices to form a closed curve; d. Calculate the area of ​​the closed curve. Since the B-spline curve is fitted by multiple small line segments, the closed curve is treated as a polygon to calculate its area. First, calculate the centroid of the polygon, and then connect the centroid to each vertex of the polygon to form multiple triangles. The sum of the areas of all the triangles is the area of ​​the closed curve. e. If array A is not empty, it means that the data in A may be able to fit another curve, so go to c; f. If the data has not been completely read, proceed to step b; otherwise, end this process. The sum of the areas corresponding to all slices and the product of the slice spacing is the volume of the object.

5. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 2, characterized in that, The method employs an automatic measurement device for the feature dimensions of irregular objects. This device includes a shape and size measuring instrument for performing point cloud scanning measurements on the object to be measured and obtaining point cloud data. It also includes: A robotic arm (1) is provided with a vacuum suction cup (2); Conveyor belt (3), said conveyor belt (3) being disposed below said robotic arm (1); and A carrier plate (4) is used to support the object to be tested, and the carrier plate (4) is placed on the conveyor belt (3) for transport. The robotic arm (1) uses the vacuum suction cup (2) to pick up the sample plate (4) carrying the object to be tested, so as to send the object to be tested to the shape and size measuring device.

6. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 5, characterized in that, The vacuum suction cups (2) are four in a circumferential distribution and are supported by a bracket (5). The bracket (5) is connected to the robotic arm (1). The four vacuum suction cups (2) are simultaneously adsorbed on the edge of the planar carrier plate (4).

7. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 6, characterized in that, In each of the four vacuum suction cups (2), a debris guard (6) is provided between every two adjacent vacuum suction cups (2) along its circumferential distribution direction, and the debris guard (6) is arranged only in three consecutive directions between the vacuum suction cups (2), so that one direction is left without the debris guard (6).

8. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 7, characterized in that, The vacuum suction cup (2) is connected to the vacuum equipment via a connecting pipe.

9. The automatic measurement method for the characteristic dimensions of irregular objects according to claim 5, characterized in that, The suction cup material of the vacuum suction cup (2) is made of nitrile rubber.

Images