Data processing method and device, electronic equipment and storage medium

Abstract

The application provides a data processing method and device, electronic equipment, storage medium and unmanned aerial vehicle. The method comprises: acquiring point cloud data, the point cloud data comprising point clouds of a plurality of target objects and point clouds of connecting lines between adjacent target objects in the plurality of target objects; dividing the plurality of target objects into at least one object set based on the direction of the connecting lines, wherein the direction of the connecting lines in each object set in the at least one object set is substantially the same; and performing a completion process on the point clouds of the plurality of connecting lines according to the similarity between the point clouds of the plurality of connecting lines in the object set. In the embodiments of the application, the point clouds of the plurality of connecting lines are completed according to the similarity of the point clouds of the plurality of connecting lines in the object set, so that the missing point clouds can be completed in the case of partial loss of the point clouds of the connecting lines, thereby improving the accuracy of power grid detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a data processing method and device, electronic equipment and storage medium. BACKGROUND

[0002] With the continuous expansion of the scale of the power system, the structure of the power grid is becoming more and more complex, and manual inspection cannot meet the needs of inspection. Compared with the traditional manual detection method, the unmanned aerial vehicle detection not only reduces the detection cost, but also can efficiently and safely complete the detection task. However, the unmanned aerial vehicle detection method still has the problem of insufficient detection accuracy. SUMMARY

[0003] Therefore, the embodiments of the present application aim to provide a method capable of improving the detection accuracy of the power grid to solve the problem of insufficient detection accuracy.

[0004] In a first aspect, a data processing method is provided, comprising:

[0005] obtaining point cloud data, the point cloud data comprising point clouds of a plurality of target objects and point clouds of connection lines between adjacent target objects in the plurality of target objects;

[0006] dividing the plurality of target objects into at least one object set based on the direction of the connection lines, wherein the direction of the connection lines in each object set in the at least one object set is substantially the same; and

[0007] completing the point clouds of the plurality of connection lines based on the similarity between the point clouds of the plurality of connection lines in the object set, to obtain complete point clouds of the plurality of connection lines.

[0008] In an embodiment, before the point clouds of the plurality of connection lines are completed based on the similarity between the point clouds of the plurality of connection lines in the object set, the method further comprises:

[0009] dividing the target objects in the plurality of object sets into a plurality of object groups according to the relative positions of the plurality of target objects; and

[0010] complementing the missing segments of the point clouds of the plurality of connection lines according to the similarity of the point clouds of the plurality of connection lines in each object group in the plurality of object groups.

[0011] In an embodiment, the dividing the target objects in the plurality of object sets into a plurality of object groups according to the relative positions of the target objects comprises:

[0012] If a distance between two adjacent target objects in the plurality of target objects is less than a predetermined value, and there is a segment of the point cloud of the connection line between the two adjacent target objects, the two adjacent target objects are determined as one object group.

[0013] In one embodiment, the completing the plurality of point clouds of connection lines according to the similarity between the plurality of point clouds of connection lines in the object set comprises:

[0014] According to the similarity between the plurality of point clouds of connection lines in different object groups in the object set, the missing segments of the plurality of point clouds of connection lines in each object group are mutually completed.

[0015] The plurality of point clouds of connection lines are completed according to a parabolic fitting algorithm.

[0016] In one embodiment, the completing the plurality of point clouds of connection lines according to the similarity between the plurality of point clouds of connection lines in the object set further comprises:

[0017] If there is no point cloud between two adjacent target objects of the plurality of target objects, a similar object group with a similar relative positional relationship to the two adjacent target objects is determined, and the complete point cloud between the similar object groups is matched to the two adjacent target objects without the point cloud.

[0018] In one embodiment, further comprising:

[0019] The complete point cloud is optimized according to the similarity of the curvatures of each complete point cloud in the object set.

[0020] In one embodiment, the obtaining the point cloud data comprises:

[0021] Obtaining image data, the image data comprising images of the plurality of target objects and connection lines between adjacent target objects in the plurality of target objects;

[0022] Generating the point cloud data according to the image data;

[0023] Determining point clouds of the plurality of target objects in the point cloud data, and determining point clouds of connection lines between adjacent target objects in the plurality of target objects.

[0024] In one embodiment, the determining the point clouds of the connection lines between adjacent target objects in the plurality of target objects comprises:

[0025] Hierarchically classifying the point clouds of the connection lines between adjacent target objects in the plurality of target objects by using a manifold dimension reduction and clustering algorithm, to distinguish the point clouds of the connection lines between adjacent target objects in the plurality of target objects.

[0026] In an embodiment, the image data is data collected by a drone, the target objects are line poles or line towers in the image, and the connection lines are electric wires.

[0027] In an embodiment, the multiple target objects are divided into at least one object set based on the directions of the connection lines, and specifically:

[0028] In the projection of the point cloud data on the projection plane, the multiple target objects, between which the slopes of the connection lines are similar and the connection lines of which are substantially in the same line, are divided into one object set.

[0029] In a second aspect, a data processing apparatus is provided, comprising:

[0030] An acquisition unit is configured to acquire point cloud data, the point cloud data comprising point clouds of multiple target objects and point clouds of connection lines between adjacent target objects in the multiple target objects;

[0031] A division unit is configured to divide the multiple target objects into at least one object set based on the directions of the connection lines, wherein the connection lines in each object set in the at least one object set have substantially the same direction; and

[0032] A determination unit is configured to perform a completion processing on the point clouds of the multiple connection lines according to the similarity between the point clouds of the multiple connection lines in the object set, to obtain complete point clouds of the multiple connection lines.

[0033] In a third aspect, an electronic device is provided, comprising a processor and a memory, the memory storing a computer program executable by the processor, and the processor being capable of executing the computer program to implement the method described in the above embodiments.

[0034] In a fourth aspect, a computer readable storage medium is provided, on which a computer program is stored, and the computer program is executable by a processor to implement the method described in the above embodiments.

[0035] The application provides a data processing method and device, electronic equipment and a storage medium. The method comprises: obtaining point cloud data, the point cloud data comprising point clouds of a plurality of target objects and point clouds of connecting lines between adjacent target objects in the plurality of target objects; dividing the plurality of target objects into at least one object set based on the direction of the connecting lines, wherein the connecting lines in each object set in the at least one object set have substantially the same direction; and performing a completion process on the point clouds of the plurality of connecting lines according to the similarity between the point clouds of the plurality of connecting lines in the object set. In the embodiments of the application, the point clouds of the plurality of connecting lines are completed according to the similarity between the point clouds of the plurality of connecting lines in the object set, so that the missing point clouds can be completed in the case of partial missing of the point clouds of the connecting lines, thereby improving the accuracy of power grid detection. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 Fig. 1 shows a flowchart of a data processing method provided by an embodiment of the application;

[0037] Figure 2 Fig. 2 shows a schematic diagram of the projection of the point clouds of the connecting lines in a first projection plane according to an embodiment of the application;

[0038] Figure 3 Fig. 3 shows a schematic diagram of the projection of the point clouds of the connecting lines in a second projection plane according to an embodiment of the application;

[0039] Figure 4 Fig. 4 shows a schematic diagram of the projection of the point clouds of the connecting lines in a projection plane after clustering according to an embodiment of the application;

[0040] Figure 5 Fig. 5 shows a schematic diagram of the division of the object set in the data processing method according to an embodiment of the application;

[0041] Figure 6 Fig. 6 shows a flowchart of the determination of complete point clouds in the data processing method according to an embodiment of the application;

[0042] Figure 7 Fig. 7 shows a process schematic diagram of the mutual complementation of the point clouds of the connecting lines in the data processing method according to an embodiment of the application;

[0043] Figure 8 Fig. 8 shows a flowchart of a data processing method provided by another embodiment of the application;

[0044] Figure 9 Fig. 9 shows a flowchart of the acquisition of data point clouds according to another embodiment of the application;

[0045] Figure 10Fig. 2 shows a process diagram of point cloud complementing each other of connection lines according to another embodiment of the present application;

[0046] Figure 11 Fig. 3 shows a flow diagram of a data processing method according to another embodiment of the present application;

[0047] Figure 12 Fig. 4 shows a structure diagram of a data processing apparatus according to another embodiment of the present application;

[0048] Figure 13 Fig. 5 shows a structure diagram of a data processing apparatus according to another embodiment of the present application;

[0049] Figure 14 Fig. 6 shows a structure diagram of an electronic device according to another embodiment of the present application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0051] In one comparative example, a data acquisition device is carried by a drone to obtain power grid images. Specifically, point cloud is obtained by a radar in the drone, and power line point cloud is extracted by point cloud classification. The power line point cloud is segmented according to the extreme points of the power line height, and individual power line point cloud is extracted by layering to realize power line vectorization. The radar includes but is not limited to a laser radar (Light Detection and Ranging, LIDAR) and a millimeter wave radar (Millimeter-Wave Radar, MWR).

[0052] However, the method in the comparative example has high requirements for the precision and completeness of the power line point cloud, and cannot accurately segment or layer the power line when the power line point cloud is incomplete, or the power line fitting precision is affected when there is much noise.

[0053] To solve the above technical problems, an embodiment of the present application provides a data processing method to improve the detection precision of the power grid. In the technical solution of the present application, a plurality of target objects are divided into different object sets according to the power line direction from the original point cloud data, and the missing point cloud in the point cloud data is complemented according to the similarity of the power line shape in the object set to obtain complete power line vectors.

[0054] It should be understood that in the following embodiments, the point cloud data is exemplarily acquired by a data acquisition device in a UAV. However, in other optional implementations, the point cloud data can also be acquired by different types of devices, such as mobile devices (e.g., an airplane and a car).

[0055] Firstly, the following describes the explanations of the terms in the embodiments of the present application.

[0056] The point cloud data refers to a set of vectors in a three-dimensional coordinate system. In addition to representing geometric position information, the point cloud data can also represent the RGB (Red-Green-Blue) color, the gray value, the depth, the segmentation result, and the like of a point. The form of the point cloud data is equivalent to a 3D sample data. The point cloud data in the present application is described by taking the data acquired by an RGB camera as an example. Each point contains the coordinates of a three-dimensional point and the attribute information of the corresponding three-dimensional point, such as the red-green-blue (RGB) color information. The point cloud data can be a set of 6-component points (X, Y, Z, R, G, B) or equivalently (X, Y, Z, Y, U, V), where (X, Y, Z) defines the spatial position of the point in a three-dimensional space, and (R, G, B) or (Y, U, V) defines the color of the point.

[0057] The target object refers to a tower or a pole or the like in a power grid.

[0058] The connection line refers to a wire in the power grid.

[0059] The power grid detection accuracy refers to the deviation between the detected coordinates of the tower, the pole, and the wire and the actual coordinates of the tower, the pole, and the wire in the power grid. The smaller the deviation, the higher the detection accuracy, and the more accurately the detected value reflects the actual state of the power grid.

[0060] The technical solutions of the embodiments of the present application will be described below in combination with specific embodiments.

[0061] Figure 1 Fig. 1 shows a flowchart of a data processing method provided by an embodiment of the present application. The data processing method is executed by a data processing device to process image data collected by a camera of a UAV. The data processing device can be a chip or a processor including a control circuit, and the like. The chip or the processor can be arranged in a UAV, a mobile terminal, or a server. It should be understood that the data processing device can also be a server or a mobile terminal. Figure 1 The method shown includes steps S110, S120, and S130, but these steps or operations are only examples, and other operations or steps can also be performed by the embodiments of the present application. Figure 1modifications of the methods described herein, and that the steps described with respect to one method can be performed with respect to another method. Not all steps are required, or performed, in all implementations, and steps can be performed in an order other than the order described.

[0062] As shown in Figure 1 The data processing method comprises the following steps:

[0063] Step S110: Obtain point cloud data. The point cloud data comprises point clouds of a plurality of target objects and point clouds of a plurality of connection lines.

[0064] Step S120: Divide the plurality of target objects into at least one object set based on the directions of the connection lines. The directions of the connection lines in each object set in the at least one object set are substantially the same.

[0065] Step S130: Perform completion processing on the point clouds of the plurality of connection lines based on the similarity between the point clouds of the plurality of connection lines in the object set. In step S110, the point cloud data is obtained. The point cloud data comprises point clouds of a plurality of target objects and point clouds of connection lines between adjacent target objects in the plurality of target objects.

[0066] In one implementation, the data processing apparatus is in communication connection with the unmanned aerial vehicle to obtain point cloud data collected by a data collection apparatus in the unmanned aerial vehicle. The unmanned aerial vehicle processes the original point cloud data obtained by the data collection apparatus. The communication apparatus in the unmanned aerial vehicle sends the processed point cloud data to the data processing apparatus.

[0067] In another implementation, the unmanned aerial vehicle sends the original point cloud data collected by the data collection apparatus to a server. The original point cloud data is processed in the server. The server is in communication connection with the data processing apparatus to send data comprising point clouds of target objects and point clouds of connection lines to the data processing apparatus.

[0068] In this embodiment, the point cloud data is obtained by aerial photography of the unmanned aerial vehicle. The data collection apparatus is a depth camera, which can be a red-green-blue (RGB) camera. The RGB camera is a three-dimensional depth camera, i.e., a camera that can measure the distance between the target and the camera. The point cloud data is an environmental image represented by intensity values of R (red), G (green), and B (blue) colors. Figure 2 FIG. 3 shows a schematic diagram of projection of point clouds of connection lines in a first projection plane according to an embodiment of the present application. Figure 3 FIG. 4 shows a schematic diagram of projection of point clouds of connection lines in a second projection plane according to an embodiment of the present application. Figure 4 FIG. 5 shows a schematic diagram of projection of clustered point clouds of connection lines in a second projection plane according to an embodiment of the present application. Figure 2 and Figure 3As shown, the initial point cloud data is processed by a dimension reduction algorithm to obtain the projection of the point cloud data in the first projection plane (XOY plane) and the second projection plane (YOZ plane) described in this embodiment. The first projection plane is a plane substantially perpendicular to the plane in which the connecting line is located. In this embodiment, the first projection plane is a horizontal plane, and the second projection plane is perpendicular to the horizontal plane and the plane in which the connecting line is located. Then, the k-means clustering algorithm is used to cluster the projected points to divide the power line point cloud into layers, and the point cloud of the plurality of connecting lines is obtained Figure 4 The point cloud of the plurality of connecting lines is shown. Since the power line itself is relatively thin and the background is complex, the point cloud data of the connecting line is relatively incomplete, and the point cloud of the connecting line needs to be completed.

[0069] In this embodiment, the data acquisition device is a depth camera. Compared with the laser radar used in the comparative example to obtain point cloud data, the depth camera has low cost and relatively mature technology, which can reduce the cost.

[0070] In other optional implementations, the data acquisition device can include at least one of an RGB camera, a monocular camera, a binocular camera, a TOF (Time-of-Flight) camera, and the like.

[0071] In step S120, based on the direction of the connecting line, the plurality of target objects are divided into at least one object set. Each object set in the at least one object set has connecting lines with substantially the same direction.

[0072] The substantially same direction of the connecting line can be that the projection of the connecting line in the horizontal first projection plane (XOY plane) and the coordinate axis X in the first projection plane have substantially the same angle. That is, the absolute difference between the angle between the projection of each two adjacent target objects in the first projection plane and the coordinate axis X and the angle between the coordinate axis X and the coordinate axis X is less than a predetermined angle. For example, the predetermined angle is 3°, when the angle between the projection of one connecting line and the coordinate axis X is 30°, and the angle between the projection of another connecting line and the coordinate axis X is 32°, the absolute difference between the angle between the two connecting lines and the coordinate axis X is 2°, which is less than the predetermined angle, and it can be considered that the directions of the two connecting lines are substantially the same. It should be understood that the predetermined angle can be adjusted as needed.

[0073] The multiple target objects are divided into at least one object set according to the directions of the connection lines, specifically, in the projection of the point cloud data on the projection plane, the slopes of the connection lines between adjacent target objects are similar, and the multiple target objects in the same straight line are divided into one object set. That is, the multiple target objects in the same object set are substantially in the same straight line. In an implementation, the multiple target objects can also be divided into multiple object sets according to the similarity of the slopes of the connection lines between adjacent target objects, and the absolute value of the difference between the slopes of the connection lines between two adjacent target objects in each object set is less than a predetermined threshold, for example, the predetermined threshold is 0.1. That is, the included angle between the connection lines between two adjacent target objects is less than a predetermined angle, for example, the predetermined angle is 5°.

[0074] Figure 5 Fig. 1 shows a schematic diagram of dividing an object set in a data processing method provided by an embodiment of the present application. Figure 5 Fig. 1 shows a schematic diagram of dividing an object set in a data processing method provided by an embodiment of the present application.

[0075] In step S130, the point clouds of the multiple connection lines are complemented according to the similarity between the point clouds of the multiple connection lines in the object set, so as to obtain complete point clouds of the multiple connection lines. The complete line point clouds are the connection lines connecting different target objects.

[0076] Figure 6 Fig. 2 shows a flowchart of determining complete point clouds in a data processing method provided by an embodiment of the present application. Figure 6 As shown in Fig. 2, determining the multiple complete line point clouds according to the similarity of the point clouds of the connection lines in the object set includes the following steps:

[0077] Step S131: According to the similarity between the point clouds of the multiple connection lines in different object groups in the object set, the missing segments of the point clouds of the multiple connection lines in each object group are complemented.

[0078] Step S132: Completing the point cloud of the plurality of connection lines according to a parabola fitting algorithm.

[0079] In step S131, according to the similarity between the point clouds of the plurality of connection lines in different object groups in the object set, the missing segments of the point clouds of the plurality of connection lines in each object group are complemented.

[0080] The target objects in the same object set are similar in distance, and the power lines between the target objects are hung on the power poles or towers with similar structures, so the shapes of the power lines in the same object set are similar. According to the similarity, the point clouds of each power connection line in each group are complemented with each other.

[0081] Figure 7 The process of complementing the point clouds of the connection lines with each other in an embodiment of the present application is shown. As shown in the figure, Figure 7 There are point clouds cd of the connection lines between target objects b2 and b3. There is no complete line point cloud connection between target objects b2 and b3. There are point clouds ef of the connection lines between target objects b3 and b4. There is no complete line point cloud connection between target objects b3 and b4. Since the point clouds of each power connection line in each group are similar, the missing point clouds of the connection lines between target objects b2 and b3 can be completed according to the point clouds gf of the connection lines similar to the missing part, to form a complete line point cloud connecting target objects b2 and b3. Specifically, the point clouds gf of the connection lines can be moved to between target objects b2 and b3, and the end point d of cd and the end point g of gf are connected. The line point cloud between target objects b3 and b4 is completed according to a similar method.

[0082] It should be understood that the present application only takes the point clouds of a connection line between target objects as an example for illustration, and the point clouds of other connection lines can be completed by the above method.

[0083] It should be understood that the method of complementing the power line point cloud in the embodiments of the present application is illustrated by taking the connection of two power lines between two adjacent target objects as an example. In other embodiments, one, three, and multiple power lines can be connected between two adjacent target objects, which is not limited herein.

[0084] In step S132, the point clouds of the plurality of connection lines are completed according to a parabola fitting algorithm.

[0085] In one implementation, parabola fitting is performed on all power line point cloud data to generate power line vector data. The parabola fitting algorithm is shown in the following formula:

[0086] z=ax 2 +bx+c.

[0087] Wherein, z represents an axial vector perpendicular to the horizontal direction and towards the sky, and x represents a horizontal extension direction of the electric wire.

[0088] In the embodiment, the plurality of complete line point clouds are determined according to the similarity of the target objects in the object set, the missing point cloud data can be filled in the case of partial missing of the point cloud data, and thus the accuracy of the power grid detection can be improved.

[0089] Another embodiment of the present application provides a data processing method. Figure 8 As shown in the figure, the processing method comprises the following steps: Figure 8 As shown in the figure, the processing method comprises the following steps:

[0090] Step S810: acquiring point cloud data. The point cloud data comprises point clouds of a plurality of target objects and point clouds of connecting lines between adjacent target objects in the plurality of target objects.

[0091] Step S820: dividing the plurality of target objects into at least one object set based on the direction of the connecting lines. Wherein, the connecting lines in each object set in the at least one object set have substantially the same direction.

[0092] Step S830: dividing the target objects in the plurality of object sets into a plurality of object groups according to the relative positions of the target objects.

[0093] Step S840: mutually supplementing missing segments of the point clouds of the plurality of connecting lines according to the similarity of the point clouds of the plurality of connecting lines in each object group.

[0094] Step S850: performing a completion processing on the point clouds of the plurality of connecting lines according to the similarity between the point clouds of the plurality of connecting lines in the object set. Complete point clouds of the plurality of connecting lines are obtained.

[0095] In step S810, point cloud data is acquired. The point cloud data comprises point clouds of a plurality of target objects and point clouds of connecting lines between adjacent target objects in the plurality of target objects.

[0096] Figure 9 As shown in the figure, the processing method comprises the following steps: Figure 9 As shown in the figure, the processing method comprises the following steps:

[0097] Step S811: acquiring image data. The image data comprises images of target objects and lines.

[0098] In an implementation, the data processing apparatus obtains the image data through a communication connection with the unmanned aerial vehicle. The image data is obtained by a data acquisition apparatus in the unmanned aerial vehicle and sent to the data processing apparatus through a communication apparatus in the unmanned aerial vehicle.

[0099] Step S812: generating point cloud data according to the image data.

[0100] In an implementation, the data processing apparatus generates regional dense point cloud data after processing the image data. The point cloud data can be a plurality of data points extracted from the image data by a predetermined extraction model. The extraction model can be a view geometry model, such as the NVS algorithm, etc.

[0101] Step S813: determining the point cloud of the plurality of target objects in the point cloud data, and determining the point cloud of the connection lines between adjacent target objects in the plurality of target objects.

[0102] In an implementation, the determination of the point cloud of the connection lines between adjacent target objects in the plurality of target objects includes hierarchical clustering of the point cloud of the connection lines between adjacent target objects in the plurality of target objects by using a manifold dimension reduction and clustering algorithm, so as to distinguish the point cloud of the connection lines between adjacent target objects in the plurality of target objects.

[0103] Specifically, the data processing apparatus determines the point cloud of the plurality of target objects and the point cloud of the plurality of connection lines in the point cloud data by using an instance segmentation algorithm. The projection distribution of the point cloud of the power line on the first projection plane and the second projection plane is obtained according to a dimension reduction algorithm, which can be an isomap manifold dimension reduction algorithm. The projection distribution of the point cloud of the connection lines on the first projection plane and the second projection plane can be as shown in Figure 2 and Figure 3 Then, the point cloud of the connection lines is hierarchically clustered according to a clustering algorithm, which can be a k-means clustering algorithm.

[0104] In step S820, the plurality of target objects are divided into at least one object set based on the direction of the connection lines. The direction of the connection lines in each object set in the at least one object set is substantially the same.

[0105] In an implementation, the specific method can refer to step S120.

[0106] In step S830, the target objects in the plurality of object sets are divided into a plurality of object groups according to the relative positions of the plurality of target objects.

[0107] In one implementation, dividing the target objects in the multiple object set into multiple object groups according to the relative positions of the multiple target objects includes: if the distance between two adjacent target objects is less than a predetermined value, and there is a point cloud of connecting lines between the adjacent target objects, the two adjacent target objects are determined as an object group.

[0108] For example, to search for multiple nearest neighbor objects around a target object, a predetermined distance of 200 meters can be set. This means searching for other target objects within 200 meters of the target object. Then, it's determined whether there are point clouds of connecting lines between the target object and its neighbors. If there are, the target object is grouped with the other objects. Several wires will be attached to two target objects in a group, and the shape similarity of the wires attached to two target objects in the same group is higher than the shape similarity of wires in different groups.

[0109] In another implementation, when the point cloud of the target objects is incomplete, the target objects in each object set are divided into multiple object groups based on the relative positions of the target objects and the extreme points of the wire height. That is, the point cloud of the target objects is first completed based on the extreme points of the wire height, and then the target objects are grouped according to their relative positions.

[0110] In step S840, the missing segments of the point cloud of the multiple connecting lines are mutually supplemented based on the similarity of the point clouds between the multiple connecting lines in each of the multiple object groups.

[0111] Figure 10 The diagram shown illustrates the process of point cloud complementarity of connecting lines according to another embodiment of this application. Figure 10 As shown, the object group includes the point clouds cd and ij of the connecting line, which are basically parallel. The method for supplementing the point clouds of the connecting line specifically includes: first, projecting the point clouds of the connecting line onto the XOY plane; then, determining the projection point h of the endpoint j of the point cloud ij onto the point cloud cd; then calculating the vector hj between points h and j in 3D space; and based on the vector hj, supplementing the point cloud hd of the connecting line to the endpoint j of the point cloud ij.

[0112] In this embodiment, before determining the complete line point cloud based on the similarity of the point clouds of each connecting line in the object set, the complete line point cloud is determined by the similarity of the point clouds of each connecting line in the same group. This is because the similarity of the point clouds of connecting lines in the same group is higher than that of the point clouds of connecting lines in the same object set. Therefore, in this embodiment, the accuracy of the point cloud of connecting lines can be improved by first completing the point clouds of connecting lines in the same group.

[0113] In step S850, the point clouds of the multiple connecting lines are completed based on the similarity between the point clouds of the multiple connecting lines in the object set, so as to obtain the complete point clouds of the multiple connecting lines.

[0114] In one implementation, step S130 can be referred to.

[0115] In another implementation, if there is no point cloud between two adjacent target objects of the plurality of target objects, a similar object group with a relative positional relationship similar to the two adjacent target objects is determined, and the complete point cloud between the similar object group is matched between the two adjacent target objects that do not have a point cloud.

[0116] Specifically, we can determine the slope and distance of the projections of two adjacent target objects in a point cloud without connecting lines onto the first projection plane. We then search for the other two adjacent target objects in the already determined complete line point cloud that have the highest similarity to this slope and distance. Finally, we match the point cloud of the connecting line between these other two adjacent target objects to the point cloud without connecting lines between the two adjacent target objects. Alternatively, we can directly substitute the 3D coordinates of two adjacent target objects into the parabolic expression of the already fitted connecting line in the object set, thus matching the complete line point cloud that can match the adjacent target objects to the point cloud without connecting lines between the two adjacent target objects.

[0117] In this embodiment, before supplementing the wires using the point cloud of connecting lines within the object set, the missing parts are supplemented by the point cloud of connecting lines within the object group. The similarity between two wires within the same object group is higher than the similarity between wires in the same object set. Therefore, supplementing the wires by the point clouds of two parallel connecting lines within the same group can improve the accuracy of power grid detection.

[0118] Another embodiment of this application provides a data processing method. Figure 11 The diagram shown is a flowchart illustrating a data processing method according to another embodiment of this application. The method includes the following steps:

[0119] Step S1110: Acquire point cloud data. The point cloud data includes point clouds of multiple target objects and point clouds of connecting lines between adjacent target objects.

[0120] Step S1120: Divide the plurality of target objects into at least one object set. Wherein, in the projection of the point cloud data onto the projection plane, the slopes of the lines connecting adjacent target objects in each of the at least one object set are similar.

[0121] Step S1130: Divide the target objects in the multiple object set into multiple object groups according to the relative positions of the multiple target objects.

[0122] Step S1140: According to the similarity of the point clouds between the plurality of connection lines in each object group in the plurality of object groups, complementing the point cloud missing segments of the plurality of connection lines.

[0123] Step S1150: According to the similarity of the point clouds between the plurality of connection lines in the object set, completing the point clouds of the plurality of connection lines to obtain complete point clouds of the plurality of connection lines.

[0124] Step S1160: Optimizing the complete point clouds according to the similarity of the curvatures of the complete point clouds in the object set.

[0125] Wherein, the steps S1110-S1150 can refer to the data processing method described in the above embodiments, which will not be repeated here.

[0126] The difference between the present embodiment and the previous embodiment is step S1160.

[0127] In step S1160, the complete line point clouds are optimized according to the similarity of the curvatures of the complete line point clouds in the object set.

[0128] In one implementation, the wire vector is globally optimized according to the similarity principle of the curvatures of the complete line point clouds in the object set. Specifically, the curvatures of the complete line point clouds in the same object set are twice fitted according to the least square method. The curvatures of the wires between the target objects in the same object set are set to the same curvatures after fitting. In actual power grid, the curvatures of the wires between the poles or towers with the same direction are basically the same. Therefore, in the present embodiment, the wire vector data that is more similar to the actual wire curvature can be fitted through twice fitting.

[0129] In the present embodiment, the accuracy of the wire vector data can be improved through twice fitting, and further, the accuracy of the power grid detection can be further improved.

[0130] Another embodiment of the present application provides a data processing device. The data processing device can be arranged in a UAV, and can also be arranged in a terminal and a server. Figure 12 The structure schematic diagram of the data processing device provided by another embodiment of the present application is shown. As shown in the figure, Figure 12 The data processing device provided by the present application includes an acquisition unit 1210, a division unit 1220 and a determination unit 1230.

[0131] The acquisition unit 1210 is used for acquiring point cloud data, and the point cloud data includes point clouds of a plurality of target objects and point clouds of connection lines between adjacent target objects in the plurality of target objects.

[0132] The partitioning unit 1220 is used to divide the plurality of target objects into at least one object set based on the direction of the connecting lines. The connecting lines in each of the at least one object set have substantially the same direction.

[0133] The determining unit 1230 is used to perform completion processing on the point clouds of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set, so as to obtain the complete point cloud of the multiple connecting lines.

[0134] It should be understood that the operations and functions of acquiring unit 1210, dividing unit 1220, and determining unit 1230 can be referred to the above. Figures 1 to 11 The data processing methods provided will not be elaborated here to avoid repetition.

[0135] Another embodiment of this application provides a data processing apparatus. The data processing apparatus can be installed in a drone, or in a terminal or server. Figure 13 The diagram shown is a structural schematic of a data processing apparatus provided in another embodiment of this application. Figure 13 As shown, the data processing apparatus provided in this application includes: an acquisition unit 1310, a division unit 1320, a grouping unit 1330, a completion unit 1340, a determination unit 1350, and an optimization unit 1360.

[0136] The acquisition unit 1310 is used to acquire point cloud data, which includes point clouds of multiple target objects and point clouds of multiple connecting lines.

[0137] The acquisition unit 1310 includes: an acquisition module 1311, a generation module 1312, and a first determination module 1313.

[0138] The acquisition module 1311 is used to acquire image data, which includes images of the target object and lines.

[0139] The generation module 1312 is used to generate point cloud data based on the image data.

[0140] The determining module 1313 is used to determine the point cloud of multiple target objects in the point cloud data, and to determine the point cloud of the connecting lines between adjacent target objects among the multiple target objects.

[0141] The point cloud method for determining the connection lines between adjacent target objects among the plurality of target objects includes:

[0142] The point cloud of the connection lines between adjacent target objects in the plurality of target objects is hierarchically layered using manifold dimensionality reduction and clustering algorithms to distinguish the point cloud of the connection lines between adjacent target objects in the plurality of target objects.

[0143] The dividing unit 1320 is configured to divide the plurality of target objects into at least one object set, wherein in the projection of the point cloud data on the projection plane, the slopes of the lines connecting adjacent target objects in each of the at least one object set are similar.

[0144] The grouping unit 1330 is configured to group the target objects in the plurality of object sets into a plurality of object groups according to the relative positions of the target objects.

[0145] The grouping unit 1330 comprises:

[0146] The grouping module 1331 is configured to determine two adjacent target objects as one object group if the distance between the two adjacent target objects is less than a predetermined value and there is a point cloud of a connecting line between the two adjacent target objects.

[0147] The complementing unit 1340 is configured to complement the missing segments of the point clouds of the lines in each object group according to the similarity of the point clouds of the lines in each object group.

[0148] The determining unit 1350 is configured to complement the point clouds of the lines in the object set according to the similarity between the point clouds of the lines in the object set, to obtain complete point clouds of the lines. In one implementation, the determining unit 1350 comprises a complementing module 1351, a fitting module 1352 and a matching module 1353.

[0149] The complementing module 1351 is configured to complement the missing segments of the point clouds of the lines in each object group according to the similarity of the point clouds of the lines in different object groups in the object set.

[0150] The fitting module 1352 is configured to determine the complete point clouds of the lines according to a parabolic fitting algorithm.

[0151] The matching module 1353 is configured to, if there is no point cloud of a connecting line between two adjacent target objects, determine a similar object group with a similar relative position relationship and the two adjacent target objects, and match the complete point clouds of the lines between the similar object groups to the two adjacent target objects without the point cloud of the connecting line.

[0152] The optimizing unit 1360 is configured to optimize the complete point clouds of the lines in each object set according to the similarity of the curvatures of the complete point clouds of the lines.

[0153] It should be understood that the operations and functions of the acquiring unit 1310, the dividing unit 1320, the grouping unit 1330, the complementing unit 1340, the determining unit 1350 and the optimizing unit 1360 can refer to the above-mentioned data processing method, and will not be repeated here. Figures 1 to 11 The data processing method is provided. In order to avoid repetition, it will not be repeated here.

[0154] Figure 14 Fig. 1 shows a schematic diagram of an electronic device according to an embodiment of the present application. Figure 14 Fig. 2 shows a schematic diagram of an electronic device according to another embodiment of the present application.

[0155] As shown in Fig. 1, the electronic device 100 comprises one or more processors 101 and a memory 102. Figure 14

[0156] The processor 101 can be a central processing unit (CPU) or other form of processor that has data processing capabilities and / or instruction execution capabilities. The processor 101 can control other components in the electronic device 100 to perform desired functions.

[0157] The memory 102 can comprise one or more computer program products, which can comprise various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory, for example, can include random access memory (RAM), cache memory, and / or the like. The non-volatile memory, for example, can include read-only memory (ROM), hard disk, flash memory, and / or the like. The computer-readable storage media can store one or more computer program instructions, which can be executed by the processor 101 to realize the data processing method of the embodiments of the present application described above or other desired functions.

[0158] In one embodiment, the electronic device 100 can further comprise an input device 103 and an output device 104, which are interconnected through a bus system and / or other forms of connection mechanism (not shown).

[0159] The input device 103 can comprise, for example, a keyboard, a mouse, and / or the like.

[0160] The output device 104 can output various information to the outside, including the determined motion data and the like. The output device 104 can comprise, for example, a display, a communication network and a remote output device connected thereto, and / or the like.

[0161] Of course, for the sake of simplicity, Figure 14 only some of the components in the electronic device 100 related to the present application are shown in Fig. 1, and components such as bus, input / output interface, and the like are omitted. In addition, the electronic device 100 can comprise any other appropriate components according to specific application cases.

[0162] ​In addition to the methods and devices described above, embodiments of the present application can also be a computer program product that includes computer program instructions that, when run by a processor, cause the processor to perform steps of the alignment method according to various embodiments of the present application described in the specification.

[0163] The computer program product can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++, etc., and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server.

[0164] In addition, embodiments of the present application can also be a computer readable storage medium having stored thereon computer program instructions that, when run by a processor, cause the processor to perform steps of the alignment method according to various embodiments of the present application described in the specification.

[0165] The computer readable storage medium can be any combination of one or more non-transitory media. The non-transitory medium can be a non-transitory signal medium or a non-transitory storage medium. The non-transitory storage medium can include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the above. More specific examples (a non-exhaustive list) of the non-transitory storage medium include an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0166] The above describes the basic principles of the present application in combination with specific embodiments, but it should be noted that the advantages, advantages, effects, etc. mentioned in the present application are only embodiments and not limitations, and these advantages, advantages, effects, etc. cannot be considered as the must-have of each embodiment of the present application. In addition, the above-mentioned specific details are only for the role of the embodiment and the role of understanding, and are not limited to the present application, and the above-mentioned specific details are not limited to the present application.

[0167] The block diagrams of the devices, apparatuses, equipment, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connection, arrangement, configuration must be as shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, systems can be connected, arranged, configured in any manner. Words such as "include", "contain", "have", and the like are open-ended words, mean "including but not limited to", and can be used interchangeably with each other. The words "or" and "and" used herein mean the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used herein means the phrase "such as but not limited to", and can be used interchangeably with each other. It is also pointed out that in the devices, apparatuses, and methods of this application, each of the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions of this application.

[0168] The above description of disclosed aspects is provided to enable any person skilled in the art to make or use the application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0169] The above description is merely illustrative of the application, and is not to be taken in a limiting sense. Any modification of the application which comes within the spirit and scope of the application will be considered equivalent to the application as defined in the claims.

Claims

1. A data processing method, characterized in that, include: Acquire point cloud data, which includes point clouds of multiple target objects and point clouds of connecting lines between adjacent target objects, wherein there are missing segments in the point cloud of connecting lines; Based on the direction of the connecting lines, the plurality of target objects are divided into at least one object set, wherein the absolute difference between the angles between the projections of the connecting lines between any two adjacent target objects in each of the at least one object set and the coordinate axis rays on the first projection plane is less than a predetermined angle; and The point clouds of multiple connecting lines in the object set are completed based on the similarity between them to obtain the complete point clouds of the multiple connecting lines. Before performing completion processing on the point clouds of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set, the method further includes: Based on the relative positions of the multiple target objects, the target objects in the multiple object set are divided into multiple object groups; and Based on the similarity of the point clouds of multiple connecting lines within each of the multiple object groups, the missing segments of the point clouds of the multiple connecting lines are mutually supplemented. The step of supplementing missing segments of the point clouds of multiple connecting lines based on the similarity of the point clouds within each of the multiple object groups includes: Project the point cloud of the plurality of connecting lines onto the target plane; Determine the projection point of the first endpoint of the point cloud of the first connecting line onto the point cloud of the second connecting line; within the target plane, the first connecting line sequentially includes a first endpoint and a second endpoint in a first direction, and the second connecting line sequentially includes a third endpoint and a fourth endpoint in the first direction. Calculate the vector between the projection point and the first endpoint in three-dimensional space, and based on the vector, complete the point cloud corresponding to the projection point on the second connecting line to the third endpoint and then to the first endpoint of the point cloud of the first connecting line.

2. The data processing method according to claim 1, characterized in that, The step of dividing the target objects in the multiple object sets into multiple object groups according to the relative positions of the target objects includes: If the distance between two adjacent target objects among the plurality of target objects is less than a predetermined value, and there is a point cloud segment with a connecting line between the two adjacent target objects, the two adjacent target objects are determined as an object group.

3. The data processing method according to claim 1 or 2, characterized in that, The step of completing the point cloud of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set includes: Based on the similarity between the point clouds of multiple connecting lines in different object groups within the object set, missing segments of the point clouds of multiple connecting lines in each object group are mutually supplemented; and The point cloud of the multiple connecting lines is completed using a parabolic fitting algorithm.

4. The data processing method according to claim 3, characterized in that, The step of completing the point cloud of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set also includes: If there is no point cloud between two adjacent target objects of the plurality of target objects, a similar object group with a relative positional relationship similar to the two adjacent target objects is determined, and the complete point cloud between the similar object group is matched between the two adjacent target objects that do not have a point cloud.

5. The data processing method according to claim 1, characterized in that, Also includes: The complete point cloud is optimized based on the similarity of the curvature of each complete point cloud within the object set.

6. The data processing method according to claim 1, characterized in that, The acquisition of point cloud data includes: Acquire image data, the image data including images of the plurality of target objects and connecting lines between adjacent target objects among the plurality of target objects; The point cloud data is generated based on the image data; Determine the point cloud of the plurality of target objects in the point cloud data, and determine the point cloud of the connecting lines between adjacent target objects among the plurality of target objects.

7. The data processing method according to claim 6, characterized in that, The point cloud method for determining the connection lines between adjacent target objects among the plurality of target objects includes: The point cloud of the connection lines between adjacent target objects in the plurality of target objects is hierarchically layered using manifold dimensionality reduction and clustering algorithms to distinguish the point cloud of the connection lines between adjacent target objects in the plurality of target objects.

8. The data processing method according to claim 6, characterized in that, The image data is collected by a drone, the target object is a utility pole or tower in the image, and the connecting line is an electric wire.

9. The data processing method according to claim 1, characterized in that, Based on the direction of the connecting lines, the plurality of target objects are divided into at least one object set, specifically as follows: In the projection of the point cloud data onto the projection plane, multiple target objects with similar slopes of the connecting lines between adjacent target objects and whose connecting lines are basically in the same straight line are grouped into a set of objects.

10. A data processing apparatus, characterized in that, include: An acquisition unit is used to acquire point cloud data, which includes point clouds of multiple target objects and point clouds of connecting lines between adjacent target objects, wherein there are missing segments in the point cloud of the connecting lines. A partitioning unit is used to divide the plurality of target objects into at least one object set based on the direction of the connecting lines, wherein the absolute difference between the angles between the projections of the connecting lines between any two adjacent target objects in each of the at least one object set and the coordinate axis rays on the first projection plane is less than a predetermined angle; and The determining unit is used to complete the point clouds of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set, so as to obtain the complete point clouds of the multiple connecting lines. Before performing completion processing on the point clouds of multiple connecting lines based on the similarity between the point clouds of multiple connecting lines in the object set, the method further includes: Based on the relative positions of the multiple target objects, the target objects in the multiple object set are divided into multiple object groups; and Based on the similarity of the point clouds of multiple connecting lines within each of the multiple object groups, the missing segments of the point clouds of the multiple connecting lines are mutually supplemented. The step of supplementing missing segments of the point clouds of multiple connecting lines based on the similarity of the point clouds within each of the multiple object groups includes: Project the point cloud of the plurality of connecting lines onto the target plane; Determine the projection point of the first endpoint of the point cloud of the first connecting line onto the point cloud of the second connecting line; within the target plane, the first connecting line sequentially includes a first endpoint and a second endpoint in a first direction, and the second connecting line sequentially includes a third endpoint and a fourth endpoint in the first direction. Calculate the vector between the projection point and the first endpoint in three-dimensional space, and based on the vector, complete the point cloud corresponding to the projection point on the second connecting line to the third endpoint and then to the first endpoint of the point cloud of the first connecting line.

11. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program executable by the processor, the processor being able to execute the computer program to implement the data processing method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the data processing method according to any one of claims 1 to 9.

Images