Insulator appearance parameter detection method and system based on RGB point cloud data

Abstract

The invention relates to the technical field of insulator measurement, in particular to an insulator appearance parameter detection method and system based on RGB point cloud data. According to the method, denoising is performed on an original RGB point cloud image set of the side surface of a target insulator through a bilateral filtering algorithm, feature extraction is performed through an FPFH algorithm, and point cloud registration processing is performed through an ICP algorithm, so that a high-quality side surface global point cloud image is obtained; then, carrying out classification processing on the side global point cloud image by adopting a PointNet algorithm, and only retaining point clouds belonging to the insulated blade part; performing curved surface reconstruction on the point cloud of the insulating blade part through a B-spline curve curved surface fitting algorithm, and further combining into a three-dimensional model of the insulating blade part; and finally, projecting the three-dimensional model into a two-dimensional point cloud, calculating an edge point cloud connection line based on the two-dimensional point cloud, and further calculating a creepage distance and an overall structure distance. Based on various algorithms, the appearance parameters of the insulator can be automatically calculated.

Description

Technical Field

[0001] The present invention relates to the technical field of insulator measurement, and more specifically, to: 1. A method for detecting the appearance parameters of an insulator based on RGB point cloud data, and 2. An insulator appearance parameter detection system using this method. Background Art

[0002] Insulators are one of the important devices in the power system, used to support and protect power lines, preventing current leakage and short circuits. Refer to Figure 1 , the insulator mainly includes a metal cylinder located in the center and insulating blades wrapped around the metal cylinder. There is a small conductive section at both ends of the metal cylinder that is not wrapped by the insulating blades.

[0003] The creepage distance is an important appearance parameter of the insulator and is one of the important parameters for evaluating the performance and safety of the insulator. Refer to Figure 1 , the creepage distance is the virtual line segment. In addition, the overall structure distance is also an important appearance parameter of the insulator. Refer to Figure 1 , the overall structure distance is the distance between the conductive sections at both ends.

[0004] With the continuous development of the power system and the improvement of voltage levels, the requirements for insulators are getting higher and higher. Therefore, the measurement technology of insulator appearance parameters has become increasingly important.

[0005] In the early stage, the measurement of the creepage distance and overall structure distance of insulators mainly relied on manual measurement: the measurement personnel used tools such as tape and rulers for manual measurement to determine the size of the insulator. This method has problems such as large errors, low efficiency, and inability to measure insulators with complex shapes.

[0006] With the development of technology, optical measurement has gradually been introduced into the measurement of insulators. Using optical instruments such as microscopes and projectors, the shape and size of insulators can be measured more precisely. However, this method still has problems such as cumbersome operation, low measurement accuracy, and inability to be automated.

[0007] Currently, the inventor needs to detect the appearance parameters of a batch of insulators, but the insulating blades of this batch of insulators are made of soft materials and will undergo natural deformation, making it not suitable to use the existing optical measurement method. Summary of the Invention

[0008] Based on this, in view of the problem that the existing optical measurement method is not suitable for measuring insulators with soft blades, a method and system for detecting the appearance parameters of insulators based on RGB point cloud data are provided.

[0009] The present invention is implemented by the following technical solutions:

[0010] In a first aspect, the present invention discloses an insulator appearance parameter detection method based on RGB point cloud data for measuring the appearance parameters of a target insulator; the appearance parameters include: creepage distance and overall structure distance.

[0011] The insulator appearance parameter detection method based on RGB point cloud data includes the following steps:

[0012] Step 1: Obtain an original RGB point cloud image set of one side of the target insulator; wherein, the original RGB point cloud image set contains a complete view of this side of the target insulator.

[0013] Step 2: Denoise the original RGB point cloud image set using a bilateral filtering algorithm to obtain a denoised original RGB point cloud image set.

[0014] Calculate features for the denoised original RGB point cloud image set using the FPFH algorithm to obtain a point cloud feature group.

[0015] Perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group using the ICP algorithm to obtain a global point cloud image of one side.

[0016] Step 3: Classify the global point cloud image of one side using the PointNet algorithm and only retain the point cloud belonging to the insulating blade part.

[0017] Step 4: Perform surface fitting on the point cloud of the insulating blade part using a B-spline curve and surface fitting algorithm. After obtaining several smooth surfaces and combining them, obtain a three-dimensional model of the insulating blade part.

[0018] Step 5: Place the three-dimensional model in a three-dimensional space coordinate system and rotate the center line of the three-dimensional model to coincide with any coordinate axis of the three-dimensional space coordinate system.

[0019] Project the three-dimensional model into a two-dimensional point cloud; wherein, the projection plane of the two-dimensional point cloud contains the coordinate axis that coincides with the center line of the three-dimensional model.

[0020] Step 6: Calculate the connection line of the edge point cloud based on the two-dimensional point cloud.

[0021] Calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator.

[0022] Obtain the coordinates of the start and end points of the connection line of the edge point cloud, calculate their difference, and use it as the overall structure distance of the target insulator.

[0023] This insulator appearance parameter detection method based on RGB point cloud data implements the method or process according to the embodiments of the present disclosure.

[0024] In a second aspect, the present invention discloses an insulator appearance parameter detection system based on RGB point cloud data, which uses the insulator appearance parameter detection method based on RGB point cloud data disclosed in the first aspect.

[0025] The insulator appearance parameter detection system based on RGB point cloud data includes: an RGB image acquisition module, an image denoising module, a feature acquisition module, a point cloud registration module, a point cloud classification module, a three-dimensional reconstruction module, a model projection module, and an appearance parameter calculation module.

[0026] The RGB image acquisition module is used to acquire an original RGB point cloud image set of one side of the target insulator; wherein, the original RGB point cloud image set contains a complete view of this side of the target insulator. The image denoising module is used to denoise the original RGB point cloud image set by using a bilateral filtering algorithm to obtain a denoised original RGB point cloud image set. The feature acquisition module is used to calculate features for the denoised original RGB point cloud image set by using the FPFH algorithm to obtain a point cloud feature group. The point cloud registration module is used to perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group by using the ICP algorithm to obtain a global point cloud image of one side. The point cloud classification module is used to perform classification processing on the global point cloud image of one side by using the PointNet algorithm and only retain the point cloud belonging to the insulating blade part. The three-dimensional reconstruction module is used to perform surface fitting on the point cloud of the insulating blade part by using a B-spline curve and surface fitting algorithm, and after obtaining several smooth surfaces and combining them, obtain a three-dimensional model of the insulating blade part. The model projection module is used to place the three-dimensional model in a three-dimensional space coordinate system and rotate the center line of the three-dimensional model to coincide with any coordinate axis of the three-dimensional space coordinate system; project the three-dimensional model into a two-dimensional point cloud; wherein, the projection plane of the two-dimensional point cloud contains the coordinate axis coinciding with the center line of the three-dimensional model. The appearance parameter calculation module is used to calculate the connection line of the edge point cloud based on the two-dimensional point cloud; calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator; obtain the coordinates of the head and tail points of the connection line of the edge point cloud, calculate their difference and use it as the overall structure distance of the target insulator.

[0027] This insulator appearance parameter detection system based on RGB point cloud data implements the method or process according to the embodiments of the present disclosure.

[0028] In a third aspect, the present invention discloses a readable storage medium. Computer program instructions are stored in the readable storage medium, and when the computer program instructions are read and run by a processor, the steps of the insulator appearance parameter detection method based on RGB point cloud data disclosed in the first aspect are executed.

[0029] The present invention discloses an electronic device. The electronic device includes a memory and a processor. The memory is used for storing computer-executable instructions, and the processor is connected to the memory through a bus; when the computer runs, the processor executes the computer-executable instructions stored in the memory, so that the computer executes the insulator appearance parameter detection method based on RGB point cloud data disclosed in the first aspect.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. Based on the original RGB point cloud image set of one side of the target insulator, the present invention first performs denoising through a bilateral filtering algorithm, then extracts features through an FPFH algorithm, and then performs registration processing through an ICP algorithm, so as to obtain a high-quality global point cloud image of the side, thereby improving the accuracy of subsequent classification.

[0032] 2. The present invention classifies the global point cloud image of the side through a PointNet algorithm, thereby extracting the point cloud belonging to the insulating blade part as the basis for calculating appearance parameters later.

[0033] 3. The present invention performs surface reconstruction on the point cloud of the insulating blade part through a B-spline curve and surface fitting algorithm, and then generates a three-dimensional model of the insulating blade part, which can adapt to target insulators of different shapes.

[0034] 4. The present invention projects the three-dimensional model into a two-dimensional point cloud, and calculates the connection of the edge point clouds through an alpha shapes algorithm and a nearest neighbor search algorithm, and then calculates the creepage distance and the overall structure distance.

[0035] 5. The present invention can also rotate the target insulator at different angles and repeat to obtain multiple calculation results, thereby improving the accuracy and reliability of appearance parameter measurement. Description of the Drawings

[0036] Figure 1 It is a structural diagram of the insulator mentioned in the background art;

[0037] Figure 2 It is a flowchart of the insulator appearance parameter detection method based on RGB point cloud data proposed in Embodiment 1 of the present invention;

[0038] Figure 3 For Figure 1 It is a schematic diagram of obtaining the original RGB point cloud image set in

[0039] Figure 4 For Figure 1 It is a structural diagram of the PointNet network model in Detailed Embodiments

[0040] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0041] It should be noted that when a component is referred to as being "installed on" another component, it can be directly on the other component or there can also be an intermediate component. When a component is considered to be "disposed on" another component, it can be directly disposed on the other component or there may be an intermediate component at the same time. When a component is considered to be "fixed to" another component, it can be directly fixed to the other component or there may be an intermediate component at the same time.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "or / and" used herein includes any and all combinations of one or more of the related listed items.

[0043] Embodiment 1

[0044] Please refer to Figure 2 , Figure 2 which is a brief flowchart of the insulator appearance parameter detection method based on RGB point cloud data in Embodiment 1, and includes the following steps:

[0045] Step 1: Obtain the original RGB point cloud image set of one side of the target insulator. Among them, the original RGB point cloud image set contains the complete view of this side of the target insulator.

[0046] Generally, the original RGB point cloud image of the side of the target insulator facing the RGB depth camera is captured by the RGB depth camera. The target insulator can be horizontally installed on the detection table, and the RGB depth camera is set above the detection table so as to capture the target insulator located on the detection table. At this time, the side of the target insulator facing the RGB depth camera is the target side to be captured. Of course, the detection table is equipped with a slide rail, and the slide rail is parallel to the axial direction of the target insulator. The RGB depth camera is connected to the carriage of the slide rail. In this way, by moving the carriage along the slide rail, the RGB depth camera can complete the capture of the target side.

[0047] However, as mentioned in the background art, since the insulating blades of the target insulator are made of soft materials, natural deformation will occur, which will block the shooting. Therefore, the shooting angle of the RGB depth camera is very important and it is necessary to ensure that the complete view of the side of the target can be captured.

[0048] Referring to Figure 3 , specifically, one end of the metal column of the target insulator is designated as A and the other end is designated as B;

[0049] First, adjust the shooting angle of the RGB depth camera so that it is biased towards the side facing B and move from B to A to complete the first round of shooting and obtain the first original RGB point cloud image, as Figure 3 shown.

[0050] Then, adjust the shooting angle of the RGB depth camera so that it is biased towards the side facing A and move from A to B to complete the second round of shooting and obtain the second original RGB point cloud image;

[0051] The first original RGB point cloud image and the second original RGB point cloud image form the original RGB point cloud image set.

[0052] In this way, by adjusting the shooting angle of the RGB depth camera, the side of the target is photographed from two angles to ensure that the complete view of the side of the target can be captured.

[0053] Step Two:

[0054] S2.1, denoise the original RGB point cloud image set using the bilateral filtering algorithm to obtain the denoised original RGB point cloud image set.

[0055] Specifically, denoise the first original RGB point cloud image using the bilateral filtering algorithm to obtain the first denoised RGB point cloud image;

[0056] Denoise the second original RGB point cloud image using the bilateral filtering algorithm to obtain the second denoised RGB point cloud image;

[0057] The first denoised RGB point cloud image and the second denoised RGB point cloud image form the denoised original RGB point cloud image set.

[0058] The bilateral filtering algorithm used in this step acts on the original RGB point cloud image. In addition to considering the Euclidean distance, the filtering conditions of the bilateral filtering algorithm also consider the color intensity information: in the flat area of the image, the pixel values change very little, and the corresponding pixel range domain weight is close to 1; in the edge area of the image, the pixel values change greatly, and the corresponding pixel range domain weight becomes larger; in this way, the edge information can be retained.

[0059] S2.2. Calculate the features of the denoised original RGB point cloud image set using the FPFH algorithm to obtain a point cloud feature group.

[0060] Specifically,

[0061] Calculate the features of the first denoised RGB point cloud image using the FPFH algorithm to obtain the first group of point cloud features;

[0062] Calculate the features of the second denoised RGB point cloud image using the FPFH algorithm to obtain the second group of point cloud features;

[0063] The first group of point cloud features and the second group of point cloud features form a point cloud feature group.

[0064] The object of the FPFH algorithm used in this step is the denoised RGB point cloud image: The FPFH (Fast Point Feature Histogram) algorithm is improved based on the PFH algorithm. The two have the same principle, but the former is a simplification of the latter, aiming to improve the calculation efficiency.

[0065] Specifically, the FPFH algorithm is to perform statistical analysis on any denoised RGB point cloud image: For the neighborhood of any point in the denoised RGB point cloud image, a certain geometric feature (such as normal, distance, angle, normal vector information of the point, etc.) is statistically analyzed, and then a statistical feature is obtained. In this way, by traversing various geometric features, multiple statistical features are obtained and form a group of point cloud features.

[0066] S2.3. Perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group using the ICP algorithm to obtain a side global point cloud image.

[0067] Specifically, based on the first denoised RGB point cloud image, the second denoised RGB point cloud image, the first group of point cloud features, and the second group of point cloud features, perform point cloud registration using the ICP algorithm to obtain a side global point cloud image.

[0068] In this step, the ICP algorithm is used for iterative calculation based on the least squares method: Set the first denoised RGB point cloud image as the target point cloud, and the second denoised RGB point cloud as the point cloud to be registered. The ultimate goal is to register and align the point cloud to be registered with the target point cloud. Specifically, select an initial transformation parameter (such as translation, rotation) to preliminarily align the point cloud to be registered, then use the nearest point search algorithm to find the nearest points between the target point cloud and the point cloud to be registered, and calculate the optimal transformation parameter between them based on the feature relationship between the target point cloud features (i.e., the first group of point cloud features) and the point cloud features to be registered (i.e., the second group of point cloud features); then transform the point cloud to be registered according to the calculated optimal transformation parameter, and use the least squares method to calculate the error between the transformed point cloud to be registered and the untransformed point cloud to be registered, and continuously repeat the whole process until the minimum error requirement is met or the maximum number of iterations is reached.

[0069] Generally speaking, the side global point cloud image obtained based on Step 2 is a relatively complete side view point cloud map.

[0070] Step 3: Classify the side global point cloud image using the PointNet algorithm and only retain the point cloud belonging to the insulating blade part.

[0071] Specifically, first classify the side global point cloud image using the PointNet algorithm to obtain the classification results; among them, the classification results include: the point cloud belonging to the insulating blade part, the point cloud belonging to the conductive section, and the point cloud belonging to the interference part (such as the detection tabletop). Then, based on the classification results, only retain the point cloud belonging to the insulating blade part.

[0072] Among them, the PointNet algorithm used in this step is based on the PointNet network model, and its structure is shown in Figure 4 :

[0073] Specifically, input the side global point cloud image into the PointNet network model. This input data is a point cloud set, represented as an n*3 2d tensor; where n represents the number of point clouds, and 3 corresponds to the x, y, and z coordinates;

[0074] The input data is first multiplied by a transformation matrix (learned by a T-Net) for alignment, and then a T-Net is used to align the feature of each point cloud data extracted by the mlp multi-layer perceptron, and a maxpooling operation is performed on each dimension of the feature to obtain the final global feature; finally, the global feature is passed through the mlp multi-layer perceptron to predict the final classification category.

[0075] Among them, there are a total of 3 classification categories: the point cloud belonging to the insulating blade part, the point cloud belonging to the conductive segment, and the point cloud belonging to the interference part. These 3 categories respectively correspond to different category score intervals. Then, a certain point cloud in the input data can output a category score through the MLP multi-layer perceptron, and the corresponding category indexed based on the category score is the predicted classification category.

[0076] That is to say, through the above classification algorithm, any point cloud belongs to one of the three categories of "the point cloud of the insulating blade part, the point cloud of the conductive segment, and the point cloud of the interference part", thus completing the tagging operation and classification. Since all point cloud information is still stored in the same structured data at this time, the structured data is traversed and output to obtain the structured data corresponding to the three point cloud categories. At this time, the unnecessary "point cloud belonging to the conductive segment" and "point cloud belonging to the interference part" are removed to obtain the "point cloud of the insulating blade part".

[0077] Step four, use the B-spline curve and surface fitting algorithm to fit the point cloud of the insulating blade part to obtain several smooth surfaces and combine them to obtain the three-dimensional model of the insulating blade part.

[0078] In this step, the B-spline curve and surface fitting algorithm used constructs several smooth surfaces by linearly combining the basis functions to fit the point cloud of the insulating blade part.

[0079] Specifically, let the knot vector U = u0, u1, …, u m ; where, u i is a knot; u i ≤ u i+1 , i = 0, 1, … m - 1. That is to say, U is a non-decreasing instance sequence.

[0080] Use N i,p (u) to represent the i-th p-th (p + 1)-order B-spline basis function, and its definition is:

[0081]

[0082]

[0083] Among them,

[0084] 1, N i,0 (u) is a step function that is zero outside the semi-open interval [ui, ui+1];

[0085] 2, when p > 0, N i,p (u) is a linear combination of two p - 1-order basis functions;

[0086] 3, When calculating a set of basis functions, it is necessary to specify the knot vector U and the degree p in advance.

[0087] It should be noted that since the smooth surface is derived from the point cloud of the insulating blade part, and the point cloud of the insulating blade part is derived from the classification of the relatively complete side global point cloud image, several constructed smooth surfaces can be spliced to form a three-dimensional model reflecting the side of the target insulator.

[0088] Step Five

[0089] S5.1, Place the three-dimensional model in a three-dimensional space coordinate system and rotate the center line of the three-dimensional model to coincide with any coordinate axis of the three-dimensional space coordinate system.

[0090] Specifically, if the center line of the three-dimensional model rotates to coincide with the x-axis, the distance from the reference origin of the three-dimensional space coordinate system along the center line direction is the x coordinate, and the distance from the x-axis longitudinally is the distance within the cross-section perpendicular to the center line of the three-dimensional model.

[0091] Since the target insulator has the characteristic that the blade is approximately perpendicular to the metal cylinder, fix the y-axis. As the x coordinate increases, the coordinate of the z-axis will continuously change.

[0092] If the center line of the target insulator rotates to coincide with the y-axis or the z-axis, the coordinates will be adjusted accordingly, which will not be elaborated here.

[0093] S5.2, Project the three-dimensional model into a two-dimensional point cloud; among them, the projection plane of the two-dimensional point cloud contains the coordinate axis that coincides with the center line of the three-dimensional model.

[0094] Generally, if the center line of the three-dimensional model rotates to coincide with the x-axis, project the three-dimensional model onto the x-y plane or the x-z plane to form a two-dimensional point cloud;

[0095] If the center line of the three-dimensional model rotates to coincide with the y-axis, project the three-dimensional model onto the y-x plane or the y-z plane to form a two-dimensional point cloud;

[0096] If the center line of the three-dimensional model rotates to coincide with the z-axis, project the three-dimensional model onto the z-x plane or the z-y plane to form a two-dimensional point cloud.

[0097] Step Six

[0098] S6.1, Calculate the connection line of the edge point cloud based on the two-dimensional point cloud.

[0099] Specifically, first use the alpha shapes algorithm to obtain the edge point cloud of the two-dimensional point cloud, and then use the nearest neighbor search algorithm to calculate the connection line of the edge point cloud.

[0100] Among them, the alpha shapes algorithm is a method for obtaining the contour points of a 2D point cloud. Taking the example of rotating the center line of a 3D model to coincide with the x-axis and projecting the 3D model onto the x-y plane or x-z plane to form a 2D point cloud, the alpha shapes algorithm will perform segmentation processing on the edge of the 2D point cloud to find the edge point cloud of the 2D point cloud. It should be noted that, at this time, the edge point cloud is still a number of isolated points.

[0101] The nearest neighbor search algorithm is a method for obtaining the connection lines of the contour of a 2D point cloud. The nearest neighbor search algorithm uses a KD tree to perform nearest neighbor search on each point of the edge point cloud, finds several points with the closest distance to it, filters to obtain the optimal points, and connects these points in series to form the connection line of the edge point cloud.

[0102] S6.2, calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator.

[0103] S6.3, obtain the coordinates of the start and end points of the connection line of the edge point cloud, calculate their difference and use it as the overall structural distance of the target insulator.

[0104] Based on the above steps, the present invention can automatically calculate the appearance parameters.

[0105] Of course, the above process only processes the projection of one perspective of the 3D model. To improve the accuracy and reliability of the measurement, the target insulator can be rotated to another side and the measurement can be repeated. If the difference between multiple measurement results is less than the preset error threshold, it indicates that the obtained appearance parameters are correct. Finally, the mean value of multiple measurement results can be taken as the final appearance parameter.

[0106] Embodiment 2

[0107] This Embodiment 2 discloses an insulator appearance parameter detection system based on RGB point cloud data, which uses the insulator appearance parameter detection method based on RGB point cloud data in Embodiment 1.

[0108] The insulator appearance parameter detection system based on RGB point cloud data includes: an RGB image acquisition module, an image denoising module, a feature acquisition module, a point cloud registration module, a point cloud classification module, a 3D reconstruction module, and an appearance parameter calculation module.

[0109] The RGB image acquisition module is configured to acquire an original RGB point cloud image set of one side of the target insulator; among them, the original RGB point cloud image set contains a complete perspective of this side of the target insulator.

[0110] The image denoising module is configured to denoise the original RGB point cloud image set using a bilateral filtering algorithm to obtain a denoised original RGB point cloud image set.

[0111] The feature acquisition module is configured to calculate features for the denoised original RGB point cloud image set using the FPFH algorithm to obtain a point cloud feature group.

[0112] The point cloud registration module is configured to perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group using the ICP algorithm to obtain a side global point cloud image.

[0113] The point cloud classification module is configured to classify the side global point cloud image using the PointNet algorithm and only retain the point cloud belonging to the insulating blade part.

[0114] The 3D reconstruction module is configured to perform surface fitting on the point cloud of the insulating blade part using the B-spline curve and surface fitting algorithm. After obtaining several smooth surfaces and combining them, a 3D model of the insulating blade part is obtained.

[0115] The model projection module is configured to place the 3D model in a 3D space coordinate system and rotate the center line of the 3D model to coincide with any coordinate axis of the 3D space coordinate system; project the 3D model into a 2D point cloud; wherein, the projection plane of the 2D point cloud contains the coordinate axis that coincides with the center line of the 3D model.

[0116] The appearance parameter calculation module is configured to calculate the connection line of the edge point cloud based on the 2D point cloud; calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator; obtain the coordinates of the head and tail points of the connection line of the edge point cloud, calculate their difference and use it as the overall structure distance of the creepage distance of the target insulator. Among them, the appearance parameter calculation module includes: an edge point cloud calculation sub-module, an edge point cloud connection line calculation module, a creepage distance calculation sub-module, and an overall structure distance calculation sub-module. The edge point cloud calculation sub-module is configured to obtain the edge point cloud of the 2D point cloud using the alpha shapes algorithm. The edge point cloud connection line calculation module is configured to calculate the connection line of the edge point cloud using the nearest neighbor search algorithm. The creepage distance calculation sub-module is configured to calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator. The overall structure distance calculation sub-module is configured to obtain the coordinates of the head and tail points of the connection line of the edge point cloud, calculate their difference and use it as the overall structure distance of the target insulator.

[0117] Embodiment 3

[0118] This Embodiment 3 discloses a readable storage medium in which computer program instructions are stored. When the computer program instructions are read and run by a processor, the steps of the insulator appearance parameter detection method based on RGB point cloud data disclosed in Embodiment 1 are executed.

[0119] When the method of Embodiment 1 is applied, it can be applied in the form of software, such as being designed as an independently running program on a computer-readable storage medium. The computer-readable storage medium can be a USB flash drive, designed as a USB key, and through the USB flash drive, a program for starting the entire method by external triggering is designed.

[0120] When the method of Embodiment 1 is applied, it can be applied in the form of software, such as being designed as an independently running program on a computer-readable storage medium. The computer-readable storage medium can be a USB flash drive, designed as a USB key, and through the USB flash drive, a program for starting the entire method by external triggering is designed.

[0121] Embodiment 3 also discloses an electronic device, including a memory and a processor. The memory is used to store computer-executable instructions, and the processor is connected to the memory through a bus; when the computer runs, the processor executes the computer-executable instructions stored in the memory, so that the computer executes the insulator appearance parameter detection method based on RGB point cloud data disclosed in Embodiment 1.

[0122] The electronic device can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack server, a blade server, a tower server or a cabinet server (including an independent server, or a server cluster composed of multiple servers) that can execute programs, etc. The electronic device of this embodiment at least includes but is not limited to: a memory and a processor that can communicate with each other through a system bus.

[0123] In this embodiment, the memory (i.e., the computer-readable storage medium) includes flash memory, a hard disk, a multimedia card, a card-type memory (such as an SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory can be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the memory can also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. Of course, the memory can also include both the internal storage unit and the external storage device of the computer device. In this embodiment, the memory is usually used to store the operating system and various application software installed on the computer device. In addition, the memory can also be used to temporarily store various data that have been output or will be output.

[0124] The processor may be a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor is generally used to control the overall operation of the computer device. In this embodiment, the processor is used to run the program code stored in the memory or process data to implement the process of the insulator appearance parameter detection method based on RGB point cloud data in the foregoing Embodiment 1.

[0125] The technical features of the above-described embodiments may be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0126] The above-described embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

Claims

1. An insulator appearance parameter detection method based on RGB point cloud data, which is used to measure the appearance parameters of the target insulator; The appearance parameters include: Creepage distance and overall structure distance; characterized in that The method for detecting the appearance parameters of an insulator based on RGB point cloud data includes the following steps: Step 1: Obtain the original RGB point cloud image set of one side of the target insulator; wherein, the original RGB point cloud image set contains the complete view of this side of the target insulator; Step 2: Denoise the original RGB point cloud image set using the bilateral filtering algorithm to obtain the denoised original RGB point cloud image set; Calculate the features of the denoised original RGB point cloud image set using the FPFH algorithm to obtain the point cloud feature group; Perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group using the ICP algorithm to obtain a global point cloud image of one side; Step 3: Classify the global point cloud image of one side using the PointNet algorithm and only retain the point cloud belonging to the insulating blade part; Step 4: Fit the point cloud of the insulating blade part using the B-spline curve and surface fitting algorithm to obtain several smooth surfaces and then combine them to obtain the three-dimensional model of the insulating blade part; Step 5: Place the three-dimensional model in the three-dimensional space coordinate system and rotate the center line of the three-dimensional model to coincide with any coordinate axis of the three-dimensional space coordinate system; Project the three-dimensional model into a two-dimensional point cloud; wherein, the projection plane of the two-dimensional point cloud contains the coordinate axis that coincides with the center line of the three-dimensional model; Step 6: Calculate the connection line of the edge point cloud based on the two-dimensional point cloud; Calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator; Obtain the coordinates of the head and tail points of the connection line of the edge point cloud, calculate their difference and use it as the overall structure distance of the target insulator.

2. The insulator appearance parameter detection method based on RGB point cloud data according to claim 1, characterized in that In Step 1: Horizontally install the target insulator on the detection table and set the RGB depth camera above the detection table; One end of the metal cylinder of the target insulator is set as A and the other end is set as B; First, adjust the shooting angle of the RGB depth camera so that it is biased towards the side facing B and move from B to A to complete the first round of shooting and obtain the first original RGB point cloud image; Then, adjust the shooting angle of the RGB depth camera so that it is biased towards the side facing A and move from A to B to complete the second round of shooting and obtain the second original RGB point cloud image; The first original RGB point cloud image and the second original RGB point cloud image constitute the original RGB point cloud image set.

3. The insulator appearance parameter detection method based on RGB point cloud data according to claim 2, characterized in that Step 2 includes: Denoise the first original RGB point cloud image using the bilateral filtering algorithm to obtain the first denoised RGB point cloud image; Denoise the second original RGB point cloud image using the bilateral filtering algorithm to obtain the second denoised RGB point cloud image; The first denoised RGB point cloud image and the second denoised RGB point cloud image constitute the denoised original RGB point cloud image set; Calculate the features of the first denoised RGB point cloud image using the FPFH algorithm to obtain the first group of point cloud features; Calculate the features of the second denoised RGB point cloud image using the FPFH algorithm to obtain the second group of point cloud features; The first group of point cloud features and the second group of point cloud features constitute the point cloud feature group; Based on the first denoised RGB point cloud image, the second denoised RGB point cloud image, the first set of point cloud features, and the second set of point cloud features, the ICP algorithm is used for point cloud registration to obtain a side global point cloud image.

4. The insulator appearance parameter detection method based on RGB point cloud data according to claim 1 or 3, characterized in that Step three includes: First, the side global point cloud image is classified using the PointNet algorithm to obtain a classification result; among them, the classification result includes: point clouds belonging to the insulating blade part, point clouds belonging to the conductive section, and point clouds belonging to the interference part; Then, based on the classification result, only the point clouds belonging to the insulating blade part are retained.

5. The insulator appearance parameter detection method based on RGB point cloud data according to claim 1, wherein, In step five, If the center line of the three-dimensional model rotates to coincide with the x-axis, the three-dimensional model is projected onto the x-y plane or the x-z plane to form a two-dimensional point cloud; If the center line of the three-dimensional model rotates to coincide with the y-axis, the three-dimensional model is projected onto the y-x plane or the y-z plane to form a two-dimensional point cloud; If the center line of the three-dimensional model rotates to coincide with the z-axis, the three-dimensional model is projected onto the z-x plane or the z-y plane to form a two-dimensional point cloud.

6. The insulator appearance parameter detection method based on RGB point cloud data according to claim 5, wherein, In step six, The calculation method of the edge point cloud connection line includes: First, the alpha shapes algorithm is used to obtain the edge point cloud of the two-dimensional point cloud, and then the nearest neighbor search algorithm is used to calculate the edge point cloud connection line.

7. An insulator appearance parameter detection system based on RGB point cloud data, characterized in that, It uses the insulator appearance parameter detection method based on RGB point cloud data described in any one of claims 1-6; The insulator appearance parameter detection system based on RGB point cloud data includes: An RGB image acquisition module, which is used to acquire an original RGB point cloud image set of one side of the target insulator; among them, the original RGB point cloud image set contains the complete view of this side of the target insulator; An image denoising module, which is used to denoise the original RGB point cloud image set using the bilateral filtering algorithm to obtain a denoised original RGB point cloud image set; A feature acquisition module, which is used to calculate features of the denoised original RGB point cloud image set using the FPFH algorithm to obtain a point cloud feature group; A point cloud registration module, which is used to perform point cloud registration on the denoised original RGB point cloud image set and the point cloud feature group using the ICP algorithm to obtain a side global point cloud image; A point cloud classification module, which is used to classify the side global point cloud image using the PointNet algorithm and only retain the point clouds belonging to the insulating blade part; A three-dimensional reconstruction module, which is used to perform surface fitting on the point cloud of the insulating blade part using the B-spline curve and surface fitting algorithm, and after obtaining several smooth surfaces and combining them, obtain a three-dimensional model of the insulating blade part; A model projection module, which is used to place the three-dimensional model in a three-dimensional space coordinate system and rotate the center line of the three-dimensional model to coincide with any coordinate axis of the three-dimensional space coordinate system; project the three-dimensional model into a two-dimensional point cloud; among them, the projection plane of the two-dimensional point cloud contains the coordinate axis with which the center line of the three-dimensional model coincides; And An appearance parameter calculation module, which is used to calculate the edge point cloud connection line based on the two-dimensional point cloud; calculate the length of the edge point cloud connection line and use it as the creepage distance of the target insulator; obtain the coordinates of the start and end points of the edge point cloud connection line, calculate their difference and use it as the overall structure distance of the target insulator.

8. The insulator appearance parameter detection system based on RGB point cloud data according to claim 7, characterized in that, The appearance parameter calculation module includes: An edge point cloud calculation sub-module, which is used to obtain the edge point cloud of the two-dimensional point cloud by using the alpha shapes algorithm; An edge point cloud connection line calculation module, which is used to calculate the connection line of the edge point cloud by using the nearest neighbor search algorithm; A creepage distance calculation sub-module, which is used to calculate the length of the connection line of the edge point cloud and use it as the creepage distance of the target insulator; And An overall structure distance calculation sub-module, which is used to obtain the coordinates of the start and end points of the connection line of the edge point cloud, calculate their difference, and use it as the overall structure distance of the target insulator.

9. A readable storage medium, characterized in that, The readable storage medium stores computer program instructions. When the computer program instructions are read and run by a processor, the steps of the insulator appearance parameter detection method based on RGB point cloud data according to any one of claims 1-6 are executed.

10. An electronic device, characterized in that, It includes a memory and a processor; the memory is used to store computer execution instructions, and the processor is connected to the memory through a bus; when the computer runs, the processor executes the computer execution instructions stored in the memory, so that the computer executes the insulator appearance parameter detection method based on RGB point cloud data according to any one of claims 1-6.

Images