Cabinet contour detection method, system and storage medium based on three-dimensional point cloud

Abstract

This invention relates to the field of image analysis technology, specifically to a method, system, and storage medium for rack contour detection based on 3D point clouds. The method acquires the normal vector of each point in the point cloud corresponding to the rack; based on the distribution of the normal vectors and positions of preset neighboring points of each point, suspected contour points are selected; based on the points contained in a specified plane under different preset directions, associated points of each suspected contour point are acquired; the associated points of each suspected contour point are divided into multiple associated point combinations; and the confidence level of each suspected contour point is obtained based on the angle between the plane formed by each suspected contour point and each of its associated point combinations, thus selecting contour points. This invention accurately selects contour points by acquiring the confidence level of each suspected contour point, avoiding misidentification of edge points near the rack's edge (such as ventilation openings or wiring ports) as contour points, thereby enabling more accurate analysis of whether the rack's shape is up to standard.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and in particular to a cabinet contour detection method, system and storage medium based on three-dimensional point cloud. Background Art

[0002] A cabinet is a structure used to store and protect various electronic devices. It is usually a regular cube made of metal plates with regular hollow vents on its surface to ensure heat dissipation of the electronic equipment in the cabinet. Since electronic equipment also needs to be connected to a variety of cables, there are also wiring ports on the cabinet. In cabinet production, the cabinet must be in a regular shape so that electronic equipment can be properly placed in and out of the cabinet.

[0003] In the existing methods, by acquiring the 3D point cloud data of the cabinet, establishing the point cloud model of the cabinet and combining it with machine vision, the cabinet contour can be detected quickly and accurately, and then the defects of the cabinet shape can be identified. However, in actual situations, when using 3D point cloud technology to detect the cabinet contour, it is easy to misjudge the edge of the hollow vents or wiring ports near the edge of the cabinet as the cabinet contour, resulting in errors in the cabinet contour detection results and misjudgment of the cabinet shape detection. Summary of the invention

[0004] In order to solve the technical problem that the edge of the hollow vent or wiring port near the edge of the cabinet is misjudged as the cabinet outline, resulting in errors in the cabinet outline detection result, the purpose of the present invention is to provide a cabinet outline detection method, system and storage medium based on three-dimensional point cloud. The technical solution adopted is as follows:

[0005] In a first aspect, an embodiment of the present invention provides a cabinet contour detection method based on a three-dimensional point cloud, the method comprising the following steps:

[0006] Get the normal vector of each point in the point cloud corresponding to the cabinet;

[0007] According to the distribution of normal vectors and positions of preset neighborhood points of each point, suspected contour points are screened out;

[0008] According to the points contained in the specified plane under different preset directions, the associated points of each suspected contour point are obtained; wherein the specified plane contains a line connecting the suspected contour point and the non-suspected contour point;

[0009] The associated points of each suspected contour point are divided into multiple associated point combinations, and the credibility of each suspected contour point is obtained according to the angle between each suspected contour point and the plane formed by the points in each associated point combination;

[0010] Filter out contour points based on their confidence level.

[0011] Furthermore, the method for obtaining the suspected contour points is:

[0012] According to the distribution of the normal vector and position of the preset neighborhood point cloud of each point, the possible degree of the contour of each point is obtained;

[0013] When the contour possibility degree meets the preset contour possibility degree range, the corresponding point is a suspected contour point.

[0014] Furthermore, the method for obtaining the possible degree of the contour is:

[0015] For any point in the point cloud, the plane perpendicular to the normal vector of the point is used as the projection plane;

[0016] Project the point and each of its preset neighboring points onto a projection plane, and on the projection plane, take the projection point of the point cloud as the center, the preset direction as the x-axis, and the direction perpendicular to the preset direction as the y-axis to form a two-dimensional coordinate system;

[0017] Get the projection point with the largest distance from the origin in each quadrant as the boundary reference point in the corresponding quadrant;

[0018] Get the number of projection points in each quadrant, all as the first number;

[0019] The possible contour path of the point is obtained based on the change in the distance between the boundary reference point and the origin and the minimum first number, as well as the fluctuation in the angle between the normal vector of the preset neighborhood point of the point and the projection plane; wherein the change in the distance between the boundary reference point and the origin and the fluctuation in the angle between the normal vector of the preset neighborhood point of the point and the projection plane are both positively correlated with the possible contour path, and the minimum first number is negatively correlated with the possible contour path.

[0020] Furthermore, the method for obtaining the association points is:

[0021] For any suspected contour point and any non-suspected contour point, the number of all points in the specified plane containing the line connecting the suspected contour point and the non-suspected contour point in each preset direction is obtained, and all of them are used as reference numbers; wherein the specified plane can be extended;

[0022] The result of normalizing the accumulated reference quantities is used as the coplanarity value between the suspected contour point and the non-suspected contour point;

[0023] When the coplanarity value is greater than a preset coplanarity threshold, the non-suspected contour point is used as an associated point of the suspected contour point.

[0024] Furthermore, the method for obtaining the combination of associated points is:

[0025] For any suspected contour point, a preset number of associated points are randomly selected from the associated points of the suspected contour point with replacement each time to form a combination of associated points of the suspected contour point.

[0026] Furthermore, the method for obtaining the credibility is:

[0027] For any suspected contour point, a plane formed by the suspected contour point and any point in the combination of the suspected contour point and any associated point is obtained as the target plane of the suspected contour point;

[0028] Obtain the cosine similarity of the angle between any two target planes of the suspected contour point as the first eigenvalue;

[0029] The result of negatively correlating and normalizing the means of all the first eigenvalues ​​is used as the credibility value of the suspected contour point.

[0030] Furthermore, the method for screening out contour points based on credibility is:

[0031] When the credibility meets the preset credibility range, the corresponding suspected contour point is used as the contour point.

[0032] Furthermore, the method for obtaining the normal vector of each point in the point cloud is:

[0033] The normal vector of each point in the point cloud is obtained through the principal component analysis algorithm.

[0034] In the second aspect, another embodiment of the present invention provides a cabinet contour detection system based on three-dimensional point cloud, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.

[0035] In a third aspect, another embodiment of the present invention provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code is executed on a computer, the steps of any one of the above methods are implemented.

[0036] The present invention has the following beneficial effects:

[0037] The present invention selects suspected contour points that may be contour points according to the distribution of normal vectors and positions of preset neighborhood points of each point, thereby improving the efficiency of subsequent accurate analysis of contour points; in order to accurately analyze contour points, it is avoided that edge points of hollow vents or wiring ports close to the edge of a cabinet are mistakenly identified as contour points, and then, according to the points contained in a specified plane under different preset directions, the associated points of each suspected contour point are obtained, and the points coplanar with each suspected contour point are determined; the associated points of each suspected contour point are further divided into a plurality of associated point combinations, thereby improving the efficiency of analyzing the coplanarity of each suspected contour point and all its associated points; in order to more accurately analyze the possibility that each suspected contour point is a contour point, the credibility of each suspected contour point is obtained according to the angle between each suspected contour point and the plane formed by the points in each of its associated point combinations, thereby accurately reflecting the possibility that each suspected contour point is a true contour point; and then, based on the credibility, contour points are accurately selected, thereby avoiding the situation that edge points of hollow vents or wiring ports close to the edge of the cabinet are mistakenly identified as contour points, thereby making it possible to more accurately analyze the shape of the cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A schematic flow chart of a cabinet contour detection method based on three-dimensional point cloud provided by an embodiment of the present invention;

[0040] Figure 2 A flow chart of a method for obtaining suspected contour points provided by an embodiment of the present invention;

[0041] Figure 3 A structural diagram of a cabinet contour detection system based on three-dimensional point cloud provided by an embodiment of the present invention;

[0042] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the cabinet contour detection method, system and storage medium based on three-dimensional point cloud proposed by the present invention, its specific implementation method, structure, features and effects in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0044] 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 invention belongs.

[0045] The specific scheme of the cabinet contour detection method, system and storage medium based on three-dimensional point cloud provided by the present invention is described in detail below with reference to the accompanying drawings.

[0046] Embodiment 1:

[0047] The specific scenario of this embodiment is: in actual situations, there may be a situation where the edge of the hollow vent or wiring port near the edge of the cabinet is misjudged as the cabinet contour, resulting in errors in the cabinet contour detection result, and then the cabinet shape detection is inaccurate, and it is impossible to accurately analyze whether the shape of the cabinet is qualified. In order to accurately determine the contour points of the cabinet and avoid the points corresponding to the edges of the hollow vent or wiring port near the edge of the cabinet from being misidentified as contour points, this embodiment first screens out suspected contour points that may be contour points according to the distribution of normal vectors of preset neighborhood points of each point in the corresponding point cloud of the cabinet. In actual situations, there are situations where the edge points of the hollow vent or wiring port are misidentified as suspected contour points. Therefore, this embodiment further analyzes each suspected contour point to obtain associated points that are coplanar with each suspected contour point. For any suspected contour point, the more all associated points of the suspected contour point are in the same plane, the less likely the suspected contour point is to be a contour point, because the contour point is the intersection of different planes, and the contour point may belong to different planes, that is, all associated points of the contour point should not be in the same plane. Furthermore, this embodiment divides the associated points of each suspected contour point into multiple associated point combinations, and obtains the credibility of each suspected contour point according to the angle between each suspected contour point and the points in each associated point combination, and accurately determines the possibility of each suspected contour point being a contour point, thereby avoiding misjudging the edge points of the hollow vents or wiring ports close to the edge of the cabinet as cabinet contour points, thereby improving the accuracy of cabinet shape detection.

[0048] The present invention proposes a cabinet contour detection method based on three-dimensional point cloud, please refer to Figure 1, which shows a schematic flow chart of a cabinet contour detection method based on three-dimensional point cloud provided by an embodiment of the present invention, the method comprising the following steps:

[0049] Step S1: Obtain the normal vector of each point in the point cloud corresponding to the cabinet.

[0050] Specifically, this embodiment uses a cabinet as an example for analysis. It should be noted that all cabinets that appear later are this cabinet. A laser scanner is used to collect the three-dimensional point cloud data of the cabinet, wherein the cabinet is scanned from multiple angles to obtain a number of points, and then point cloud registration software, such as CloudCompare point cloud registration software and MeshLab point cloud registration software, is used to merge point clouds from different perspectives into a complete point cloud model, so that the point cloud data of the cabinet is accurately obtained.

[0051] The normal vector of each point in the point cloud corresponding to the cabinet is obtained by the principal component analysis algorithm. In the process of obtaining the normal vector of each point by the principal component analysis algorithm, the 20 points closest to each point are used as the neighborhood points of the corresponding point. The implementer can set the number of neighborhood points of each point according to the actual situation, which is not limited here. The method of obtaining the normal vector of each point in the point cloud by the principal component analysis algorithm is a well-known technology and will not be described in detail.

[0052] Step S2: Screen out suspected contour points based on the normal vectors and position distribution of the preset neighborhood points of each point.

[0053] It is known that the shape of the cabinet is a regular cube. For each contour point, the points around it will appear in different planes. It is known that the normal vectors of points in different planes are different. Therefore, this embodiment filters out suspected contour points according to the distribution of the normal vectors and positions of the preset neighborhood points of each point.

[0054] Preferably, in a manner that can be implemented in this embodiment, the method for obtaining the suspected contour points can be found in Figure 2 , which shows a flow chart of a method for obtaining suspected contour points provided by this embodiment, the method comprising the following steps:

[0055] Step S201: Obtain the contour likelihood of each point based on the normal vector and position distribution of the preset neighborhood point cloud of each point.

[0056] Specifically, in this embodiment, the 50 points with the shortest Euclidean distance from each point are used as the preset neighborhood points of the corresponding point. The implementer can set the number of preset neighborhood points of each point according to the actual situation, which is not limited here. Among them, the method for obtaining the Euclidean distance is a well-known technology and will not be repeated here. When the difference in the normal vector angle of the preset neighborhood points of a certain point is greater, it means that the point is more likely to be a contour point, because the preset neighborhood points of the contour point will be in different planes; at the same time, the point and the preset neighborhood points of the point are projected into the same plane, and a two-dimensional coordinate system is constructed with the projection point of the point as the origin. When the number and position distribution of the projection points in different quadrants in the two-dimensional coordinate system are significantly different, it means that the point is more likely to be a contour point, because the preset neighborhood points of the contour point will not be located in the same plane and the contour point is located at the edge. The preset neighborhood points of the contour point are all projected into the same plane, and there is a situation where the preset neighborhood points in a plane corresponding to the contour point overlap, resulting in fewer projection points of the preset neighborhood points in the plane projected into a quadrant and closer to the origin. Therefore, the contour probability of each point is obtained according to the distribution of the normal vector and position of the preset neighborhood point cloud of each point. The greater the contour probability, the more likely the corresponding point is to be a contour point.

[0057] In a method that can be implemented in this embodiment, the method for obtaining the possible degree of contour is as follows: for any point in the point cloud, the plane perpendicular to the normal vector of the point is used as the projection plane; the point and each of its preset neighboring points are projected onto the projection plane, and on the projection plane, the projection point of the point cloud is the center, the preset direction is the x-axis, and the direction perpendicular to the preset direction is the y-axis, forming a two-dimensional coordinate system; in this embodiment, the preset direction is set to the horizontal direction, and the implementer can set the preset direction according to the actual situation, which is not limited here. The projection point with the largest distance from the origin in each quadrant is obtained as the boundary reference point in the corresponding quadrant; the number of projection points in each quadrant is obtained, all as the first number; when the difference between the Euclidean distances of the boundary reference points in different quadrants from the origin is greater, and the first number is smaller, it means that the point is more likely to be a boundary point, and the point is more likely to be a contour point. At the same time, when the angle between the normal vector of the preset neighboring point of the point and the projection plane is more different, it means that the preset neighboring point of the point is less likely to be in the same plane, and the point is more likely to be a contour point. Then, according to the change of the distance between the boundary reference point and the origin and the minimum first quantity, as well as the fluctuation of the angle between the normal vector of the preset neighborhood point of the point and the projection plane, the possible contour course of the point is obtained; wherein, the change of the distance between the boundary reference point and the origin and the fluctuation of the angle between the normal vector of the preset neighborhood point of the point and the projection plane are both positively correlated with the possible contour course, and the minimum first quantity is negatively correlated with the possible contour course.

[0058] In order to accurately represent the possibility of each point being a contour point, this embodiment specifically quantifies the contour possibility degree as a contour possibility degree value. The larger the contour possibility degree value, the more likely the corresponding point is a contour point. The calculation formula of the contour possibility degree value is: ; In the formula, is the possible degree value of the contour of the i-th point; is the maximum Euclidean distance between the boundary reference point corresponding to the i-th point and the origin; is the minimum Euclidean distance between the boundary reference point corresponding to the i-th point and the origin; is the smallest first quantity corresponding to the i-th point; is the angle between the normal vector of the jth preset neighborhood point of the ith point and the projection plane; J is the number of preset neighborhood points of the ith point; is the mean of the angles between the normal vectors of all preset neighborhood points of the i-th point and the projection plane; is the first preset constant, where is a positive integer; norm is the normalization function.

[0059] This embodiment will Set to 1 to avoid the denominator being 0. Implementers can set it according to actual conditions. The size is not limited here.

[0060] At this point, the contour probability value of each point is obtained.

[0061] Step S202: When the contour possibility degree satisfies a preset contour possibility degree range, the corresponding point is a suspected contour point.

[0062] It is known that the greater the contour possibility, the more likely the corresponding point is a contour point. Therefore, this embodiment sets the preset contour possibility threshold to 0.7. The implementer can set the preset contour possibility threshold according to the actual situation, and it is not limited here. When the contour possibility value is greater than the preset contour possibility threshold, the corresponding point is regarded as a suspected contour point.

[0063] At this point, the suspected contour points of the cabinet are determined.

[0064] Step S3: obtaining the associated points of each suspected contour point according to the points contained in the designated plane under different preset directions; wherein the designated plane contains the line connecting the suspected contour point and the non-suspected contour point.

[0065] In actual situations, when the three-dimensional point cloud technology is used to detect the cabinet contour, it is easy to misjudge the edge of the hollow vent or wiring port near the edge of the cabinet as the cabinet contour. Therefore, the suspected contour point may be the edge point of the hollow vent or wiring port near the edge of the cabinet. If the shape of the cabinet is analyzed based on the suspected contour point, it will cause errors in the cabinet contour detection result, and thus it is impossible to accurately judge whether the shape of the cabinet is qualified. In order to avoid misjudging the edge point of the hollow vent or wiring port near the edge of the cabinet as a cabinet contour point, this embodiment obtains the possible coplanar associated points of each suspected contour point. When the possibility that a suspected contour point is not coplanar with all its associated points is greater, the suspected contour point is more likely to be a true contour point, because the contour points are located in different planes, and then by analyzing the coplanarity of each suspected contour point with all its associated points, the possibility of each suspected contour point being a true contour point is determined.

[0066] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the associated points is: for any suspected contour point and any non-suspected contour point, obtain the number of all points in the specified plane containing the line connecting the suspected contour point and the non-suspected contour point in each preset direction, all of which are used as reference numbers; for example, taking the xth suspected contour point and the yth non-suspected contour point as an example for analysis, the xth suspected contour point is connected to the yth non-suspected contour point, and the line connecting the xth suspected contour point and the yth non-suspected contour point is obtained. The line connecting the xth suspected contour point and the yth non-suspected contour point is a line segment. In this embodiment, the preset directions are set to directions of 0°, 45°, 90°, 135° and 180° respectively. The implementer can set the size of the preset direction according to actual conditions, and it is not limited here. A plane containing the line connecting the xth suspected contour point and the yth non-suspected contour point is constructed as the specified plane, and the specified plane is rotated according to the preset direction to obtain the points of the cabinet contained in the specified plane in each preset direction, all of which are used as reference numbers. Among them, the specified plane can be extended to ensure that the specified plane can include all the points of the cabinet in different preset directions. When the number of references is larger, it means that the suspected contour point and the non-suspected contour point are more likely to be located in the same plane, and then all the reference numbers are accumulated and normalized as the coplanarity value of the suspected contour point and the non-suspected contour point; wherein, this embodiment normalizes the accumulated results of all reference numbers through the normalized norm function. When the coplanarity value is larger, the non-suspected contour point is more likely to be coplanar with the suspected contour point. Therefore, this embodiment sets the preset coplanarity threshold to 0.5. The implementer can set the size of the preset coplanarity threshold according to the actual situation, and it is not limited here. When the coplanarity value is greater than the preset coplanarity threshold, the non-suspected contour point is used as the associated point of the suspected contour point, and then all the associated points of the suspected contour point are obtained.

[0067] At this point, the associated points of each suspected contour point are obtained.

[0068] Step S4: Divide the associated points of each suspected contour point into multiple associated point combinations, and obtain the credibility of each suspected contour point according to the angle between each suspected contour point and the plane formed by the points in each associated point combination.

[0069] It is known that when the possibility that a suspected contour point is coplanar with all its associated points is greater, the suspected contour point is less likely to be a contour point, because the real contour point is located in different planes, and the contour point and all its associated points are not in the same plane. In actual situations, each suspected contour point will correspond to many associated points. In order to improve the analysis of the coplanarity of the associated point and the suspected contour point, this embodiment divides the associated points of each suspected contour point into multiple associated point combinations, and then constructs a plane composed of each suspected contour point and the points in each associated point combination, that is, for any suspected contour point, all the points in the suspected contour point and any associated point combination are jointly constructed as a plane. When the angle between a suspected contour point and the plane composed of the points in each associated point combination is larger, it means that the suspected contour point is more likely to be non-coplanar with all its associated points, which indirectly means that the suspected contour point is more likely to be a contour point. Therefore, this embodiment obtains the credibility of each suspected contour point according to the angle between each suspected contour point and the plane composed of the points in each associated point combination. When the credibility is larger, the corresponding suspected contour point is more likely to be a contour point.

[0070] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the combination of associated points is: for any suspected contour point, a preset number of associated points are randomly selected from the associated points of the suspected contour point with replacement each time to form a combination of associated points of the suspected contour point. In this embodiment, the preset number is set to 10, and the implementer can set the size of the preset number according to actual conditions, which is not limited here. So far, each combination of associated points of each suspected contour point is obtained.

[0071] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the degree of credibility is: for any suspected contour point, obtain the plane formed by the points in the combination of the suspected contour point and any of its associated points as the target plane of the suspected contour point; when the angle between the different target planes corresponding to the suspected contour point is larger, it means that the possibility that the suspected contour point is coplanar with all its associated points is smaller, and the suspected contour point is more likely to be a contour point, and then obtain the cosine similarity of the angle between any two target planes of the suspected contour point as the first eigenvalue, the smaller the first eigenvalue, the larger the angle between the corresponding two target planes, and the more likely the suspected contour point is to be a contour point. In order to accurately express the possibility that the suspected contour point is a contour point, the mean of all the first eigenvalues ​​are negatively correlated and normalized, and the result is used as the credibility value of the suspected contour point. Among them, the calculation formula of the credibility value is: ; In the formula, is the credibility value of the a-th suspected contour point; is the total number of target planes with the a-th suspected contour point; is the angle between the mth target plane and the kth target plane corresponding to the ath suspected contour point; is the cosine function; exp is the exponential function with a natural constant as the base.

[0072] At this point, the credibility value of each suspected contour point is obtained.

[0073] Step S5: Filter out contour points based on the degree of credibility.

[0074] It is known that when the credibility is greater, the corresponding suspected contour point is more likely to be a contour point. Therefore, this embodiment sets the preset credibility threshold to 0.7. The implementer can set the preset credibility threshold according to the actual situation, which is not limited here. When the credibility value is greater than the preset credibility threshold, the corresponding suspected contour point is used as a contour point. At this point, the contour point is accurately determined to avoid misjudging the edge point of the hollow vent or wiring port near the edge of the cabinet as a cabinet contour point, so that the shape of the cabinet can be analyzed more accurately.

[0075] In summary, this embodiment obtains the normal vector of each point in the point cloud corresponding to the cabinet; according to the distribution of the normal vector and position of the preset neighborhood points of each point, the suspected contour points are screened out; according to the points contained in the specified plane under different preset directions, the associated points of each suspected contour point are obtained; the associated points of each suspected contour point are divided into multiple associated point combinations, and the credibility of each suspected contour point is obtained according to the angle between each suspected contour point and the plane formed by the points in each associated point combination, and the contour points are screened out. The present invention accurately screens out the contour points by obtaining the credibility of each suspected contour point, avoiding the edge points of the hollow vents or wiring ports close to the edge of the cabinet from being mistakenly identified as contour points, so as to more accurately analyze whether the shape of the cabinet is qualified.

[0076] Embodiment 2:

[0077] The present invention also proposes a cabinet contour detection system based on three-dimensional point cloud, see Figure 3 , which shows a structure diagram of a cabinet contour detection system based on three-dimensional point cloud provided by an embodiment of the present invention, the system includes: a data acquisition module 10, a suspected contour point acquisition module 20, an associated point acquisition module 30, a credibility acquisition module 40 and a data processing module 50.

[0078] The data acquisition module 10 is used to obtain the normal vector of each point in the point cloud corresponding to the cabinet.

[0079] The suspected contour point acquisition module 20 is used to screen out suspected contour points according to the normal vector and position distribution of the preset neighborhood points of each point.

[0080] The associated point acquisition module 30 is used to acquire the associated points of each suspected contour point according to the points contained in the designated plane under different preset directions; wherein the designated plane contains the line connecting the suspected contour point and the non-suspected contour point.

[0081] The credibility acquisition module 40 is used to divide the associated points of each suspected contour point into multiple associated point combinations, and acquire the credibility of each suspected contour point according to the angle between each suspected contour point and the plane formed by the points in each associated point combination.

[0082] The data processing module 50 is used to filter out contour points based on the degree of credibility.

[0083] It should be noted that: the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer equipment is divided into different functional modules to complete all or part of the functions described above. In addition, the cabinet contour detection system based on three-dimensional point cloud and the cabinet contour detection method based on three-dimensional point cloud provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0084] Embodiment 3:

[0085] The present invention also proposes a cabinet contour detection device based on three-dimensional point cloud, which includes a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to execute a cabinet contour detection method based on three-dimensional point cloud provided in an embodiment of the present application. The device can be a chip, component or module, and the chip may include a connected processor and memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute a cabinet contour detection method based on three-dimensional point cloud provided in the above embodiment.

[0086] In addition, the present application embodiment also protects a computer device, see Figure 4 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402, wherein when the processor 402 executes the computer program 403, the computer device can execute any one of the cabinet contour detection methods based on three-dimensional point clouds introduced above.

[0087] Embodiment 4:

[0088] This embodiment also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to implement a cabinet contour detection method based on a three-dimensional point cloud provided in the above embodiment.

[0089] Embodiment 5:

[0090] This embodiment further provides a computer program product. When the computer program product is run on a computer, the computer is caused to execute the above-mentioned related steps to implement a cabinet contour detection method based on three-dimensional point cloud provided in the above embodiment.

[0091] Among them, the device, computer-readable storage medium, computer program product or chip provided in this embodiment is used to execute the corresponding method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here.

[0092] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0093] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A cabinet contour detection method based on three-dimensional point cloud, characterized in that: The method comprises the following steps: Get the normal vector of each point in the point cloud corresponding to the cabinet; According to the distribution of normal vectors and positions of preset neighborhood points of each point, suspected contour points are screened out; According to the points contained in the specified plane under different preset directions, the associated points of each suspected contour point are obtained; wherein the specified plane contains a line connecting the suspected contour point and the non-suspected contour point; The associated points of each suspected contour point are divided into multiple associated point combinations, and the credibility of each suspected contour point is obtained according to the angle between each suspected contour point and the plane formed by the points in each associated point combination; Filter out contour points based on the degree of credibility; The method for obtaining the suspected contour points is: According to the distribution of the normal vector and position of the preset neighborhood point cloud of each point, the possible degree of the contour of each point is obtained; When the contour possibility degree meets the preset contour possibility degree range, the corresponding point is a suspected contour point; The method for obtaining the possible degree of the contour is: For any point in the point cloud, the plane perpendicular to the normal vector of the point is used as the projection plane; Project the point and each of its preset neighboring points onto a projection plane, and on the projection plane, take the projection point of the point cloud as the center, the preset direction as the x-axis, and the direction perpendicular to the preset direction as the y-axis to form a two-dimensional coordinate system; Get the projection point with the largest distance from the origin in each quadrant as the boundary reference point in the corresponding quadrant; Get the number of projection points in each quadrant, all as the first number; The possibility degree of the contour of the point is obtained based on the change of the distance between the boundary reference point and the origin and the minimum first number, as well as the fluctuation of the angle between the normal vector of the preset neighborhood point of the point and the projection plane; wherein the change of the distance between the boundary reference point and the origin and the fluctuation of the angle between the normal vector of the preset neighborhood point of the point and the projection plane are both positively correlated with the possibility degree of the contour, and the minimum first number is negatively correlated with the possibility degree of the contour.

2. A cabinet contour detection method based on three-dimensional point cloud as claimed in claim 1, characterized in that: The method for obtaining the associated points is: For any suspected contour point and any non-suspected contour point, the number of all points in the specified plane containing the line connecting the suspected contour point and the non-suspected contour point in each preset direction is obtained, and all of them are used as reference numbers; wherein the specified plane can be extended; The result of normalizing the accumulated reference quantities is used as the coplanarity value between the suspected contour point and the non-suspected contour point; When the coplanarity value is greater than a preset coplanarity threshold, the non-suspected contour point is used as an associated point of the suspected contour point.

3. The cabinet contour detection method based on three-dimensional point cloud according to claim 1, characterized in that: The method for obtaining the combination of associated points is: For any suspected contour point, a preset number of associated points are randomly selected from the associated points of the suspected contour point with replacement each time to form a combination of associated points of the suspected contour point.

4. The cabinet contour detection method based on three-dimensional point cloud according to claim 1, characterized in that: The method for obtaining the credibility is: For any suspected contour point, a plane formed by the suspected contour point and any of its associated points is obtained as the target plane of the suspected contour point; Obtain the cosine similarity of the angle between any two target planes of the suspected contour point as the first eigenvalue; The result of negatively correlating and normalizing the means of all the first eigenvalues ​​is used as the credibility value of the suspected contour point.

5. The cabinet contour detection method based on three-dimensional point cloud according to claim 1, characterized in that: The method for screening out contour points based on the degree of credibility is: When the credibility meets the preset credibility range, the corresponding suspected contour point is used as the contour point.

6. The cabinet contour detection method based on three-dimensional point cloud according to claim 1, characterized in that: The method for obtaining the normal vector of each point in the point cloud is: The normal vector of each point in the point cloud is obtained through the principal component analysis algorithm.

7. A cabinet contour detection system based on three-dimensional point cloud, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When executing the computer program, the processor implements the steps of a cabinet contour detection method based on three-dimensional point cloud as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program codes, and when the computer program codes are executed on a computer, the steps of the cabinet contour detection method based on three-dimensional point cloud described in any one of claims 1 to 6 are implemented.

Images