Method, apparatus, and computer device for generating a model of an obstacle

Abstract

The present application provides a method, apparatus, and computer device for generating a model of an obstacle, belonging to the field of surveying and mapping technology. The method includes: determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, where the obstacles include: a plurality of fixed obstacles and at least one connecting obstacle connected between the plurality of fixed obstacles; obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle; generating a model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle; and performing a connection process on the vector models of each fixed obstacle and the models of each connecting obstacle to obtain a target model. The present application can achieve the effect of improving the accuracy of the generated model.

Description

Technical Field

[0001] This application relates to the field of surveying and mapping technology, and more particularly, to a method, apparatus, and computer device for generating models of obstacles. Background Art

[0002] With the development of unmanned aerial vehicle (UAV) technology, more and more fields require the use of UAVs for operations. For example, in the agricultural field, UAVs are used for plant protection operations in farmland or orchards. However, there are some obstacles in farmland or orchards that interfere with UAV operations, such as utility poles, power towers, wires, grass, and / or branches. If models of these obstacles are established, UAVs can accurately avoid these obstacles based on the established models during operations.

[0003] For example, in the related art, images of the UAV operation area are often collected, and the point cloud information of utility poles, power towers, and wires in the area is determined. Then, models are generated based on the point cloud information of the utility poles, power towers, and wires in the area, and the generated wire model is connected to the utility pole model or the power tower model to obtain the model of the obstacles in the UAV operation area.

[0004] However, because the environment of the UAV operation area is relatively complex, this method may have problems such as incomplete point cloud data of the obstacles collected or a large number of noise points in the collected data, resulting in possible information loss in the generated model, and further leading to low safety when the UAV avoids obstacles based on this model. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, and computer device for generating models of obstacles, which can solve the problem that the generated model may have information loss, and further achieve the effect of improving the accuracy of the generated model.

[0006] The embodiments of this application are implemented as follows:

[0007] On the one hand, an embodiment of this application provides a method for generating a model of an obstacle, including:

[0008] Determine the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, where the obstacles include: a plurality of fixed obstacles and at least one connecting obstacle connected between the plurality of fixed obstacles;

[0009] Obtain the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle;

[0010] Generate the model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle;

[0011] Connect the vector models of the fixed obstacles and the models of the connecting obstacles to obtain a target model.

[0012] Optionally, generating the models of the connecting obstacles according to the attribute information of the fixed obstacles and the attribute information of the connecting obstacles includes:

[0013] Generate the models of the connecting obstacles according to the first attribute information of the fixed obstacles and the second attribute information of the connecting obstacles, where the first attribute information includes: position information, and the second attribute information includes: elevation information.

[0014] Optionally, generating the models of the connecting obstacles according to the first attribute information of the fixed obstacles and the second attribute information of the connecting obstacles includes:

[0015] Perform parabolic fitting processing based on the position information of the fixed obstacles and the elevation information of the connecting obstacles to obtain the models of the connecting obstacles.

[0016] Optionally, the attribute information includes: the category of the obstacle;

[0017] Obtaining the vector models of the fixed obstacles from a preset set of vector models according to the attribute information of the fixed obstacles includes:

[0018] Obtain a vector model that matches the category of the fixed obstacle from a preset set of vector models as the vector model of the fixed obstacle.

[0019] Optionally, before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it includes:

[0020] Perform clustering processing on the original point cloud of each obstacle to obtain the clustered point cloud;

[0021] Determine the attribute information of each obstacle according to the clustered point cloud.

[0022] Optionally, before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it includes:

[0023] Perform segmentation processing on the original image of the target area to obtain the original images of the obstacles;

[0024] Determine the original point cloud of each obstacle according to the original image of each obstacle.

[0025] Optionally, segmenting the original image of the target area to obtain the original images of the obstacles includes:

[0026] Segmenting the original image of the target area based on an instance segmentation algorithm to obtain the original images of the obstacles.

[0027] Optionally, before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it further includes:

[0028] Determining the original point cloud of the target area according to the original image of the target area;

[0029] Segmenting the original point cloud of the target area to obtain the original point clouds of the obstacles in the target area.

[0030] Optionally, segmenting the original point cloud of the target area to obtain the original point clouds of the obstacles in the target area includes:

[0031] Segmenting the original point cloud of the target area based on an instance segmentation algorithm to obtain the original point clouds of the obstacles in the target area.

[0032] Optionally, before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it further includes:

[0033] Obtaining the original image of the target area;

[0034] Generating the original point clouds of the obstacles in the target area according to the original image.

[0035] In the second aspect of the embodiments of the present application, an obstacle model generation device is provided. The obstacle model generation device includes:

[0036] A determination module, configured to determine the attribute information of each obstacle according to the original point cloud of each obstacle in the target area;

[0037] An acquisition module, configured to acquire the vector models of the fixed obstacles from a preset vector model set according to the attribute information of the fixed obstacles;

[0038] A generation module, configured to generate the models of the connecting obstacles according to the attribute information of the fixed obstacles and the attribute information of the connecting obstacles;

[0039] A connection processing module, configured to perform connection processing on the vector models of the fixed obstacles and the models of the connecting obstacles to obtain a target model.

[0040] In a third aspect of the embodiments of the present application, a computer device is provided. The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the method for generating a model of an obstacle described in the first aspect above.

[0041] In a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method for generating a model of an obstacle described in the first aspect above.

[0042] The beneficial effects of the embodiments of the present application include:

[0043] A method for generating a model of an obstacle provided by the embodiments of the present application determines the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, obtains the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle, generates the model of each connected obstacle according to the attribute information of each fixed obstacle and the attribute information of each connected obstacle, and performs connection processing on the vector models of each fixed obstacle and the models of each connected obstacle to obtain a target model. Among them, obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle can ensure that the obtained vector models of each fixed obstacle are complete. In addition, generating the models of each connected obstacle according to the attribute information of each fixed obstacle and the attribute information of each connected obstacle can make the models of each connected obstacle more in line with the actual situation, thereby achieving the effect of improving the accuracy of the generated model. In this way, the problem that the generated model may have missing information is solved. Description of the Drawings

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0045] Figure 1 It is a flowchart of the first method for generating a model of an obstacle provided by the embodiments of the present application;

[0046] Figure 2 It is a schematic diagram of the shape of an obstacle provided by the embodiments of the present application;

[0047] Figure 3 It is a schematic diagram of a target model provided by the embodiments of the present application;

[0048] Figure 4 Flowchart of the second method for generating an obstacle model provided by an embodiment of the present application;

[0049] Figure 5 Flowchart of the third method for generating an obstacle model provided by an embodiment of the present application;

[0050] Figure 6 Flowchart of the fourth method for generating an obstacle model provided by an embodiment of the present application;

[0051] Figure 7 Flowchart of the fifth method for generating an obstacle model provided by an embodiment of the present application;

[0052] Figure 8 Flowchart of the sixth method for generating an obstacle model provided by an embodiment of the present application;

[0053] Figure 9 Flowchart of the seventh method for generating an obstacle model provided by an embodiment of the present application;

[0054] Figure 10 Flowchart of the eighth method for generating an obstacle model provided by an embodiment of the present application;

[0055] Figure 11 Flowchart of the ninth method for generating an obstacle model provided by an embodiment of the present application;

[0056] Figure 12 Structural schematic diagram of an obstacle model generation device provided by an embodiment of the present application;

[0057] Figure 13 Structural schematic diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some, but not all, of the embodiments of the present application. The components of the embodiments of the present application described and illustrated herein usually can be arranged and designed in various different configurations.

[0059] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application claimed, but merely represents selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the scope of protection of the present application.

[0060] It should be noted that like reference numerals and letters refer to like items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0061] In the agricultural field, drones are often used for plant protection operations in farmland or orchards. However, there are some obstacles in farmland or orchards that interfere with the operation of drones, such as utility poles, power towers, wires, grass, and / or branches. If models of these obstacles are established, the drones can accurately avoid these obstacles based on the established models during the operation. Currently, often by collecting images of the drone operation area, and determining the point cloud information of utility poles, power towers, and wires in the area according to the image, then generating a utility pole model, a power tower model, and a wire model respectively according to the point cloud information of utility poles, power towers, and wires in the area, and connecting the generated wire model with the utility pole model or the power tower model, so as to obtain the model of the obstacles in the drone operation area. However, because the environment of the drone operation area is relatively complex, there will be problems such as incomplete point cloud data of the collected obstacles or more noise points in the collected data. Then, generating a model according to the collected point cloud data may result in information loss in the generated model. For example, some thinner or smaller parts may not be present in the generated model, or some parts in the generated model may be bent. In addition, since the model may have information loss, that is, the model may be quite different from the shape, size, and / or position of the utility poles, power towers, and wires in the actual area. Then, if the drone avoids obstacles based on the model generated by this method, the safety during obstacle avoidance will be relatively low.

[0062] Therefore, the embodiment of the present application provides a method for generating a model of an obstacle. By determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, then obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle, and generating the model of each connecting obstacle according to the attribute information of each fixed obstacle and each connecting obstacle, and finally connecting the models of each connecting obstacle with the vector models of each fixed obstacle to obtain the target model. It can solve the problem that the generated model may have information loss, and further can achieve the effect of improving the accuracy of the generated model. In this way, the problem that the generated model may have information loss is solved.

[0063] The embodiment of the present application takes the method for generating a model of an obstacle for generating wires, utility poles, and / or power towers as an example for illustration. But it does not mean that the embodiment of the present application can only be used to generate wires, utility poles, and / or power towers.

[0064] The following will explain in detail the method for generating a model of an obstacle provided by the embodiment of the present application.

[0065] Figure 1 is a flowchart of a method for generating a model of an obstacle provided for this application. This method can be applied to a computer device, such as a drone, or a device set on the drone, or a remote server communicating with the drone, etc. Refer to Figure 1 An embodiment of this application provides a method for generating a model of an obstacle, including:

[0066] Step 101: Determine the attribute information of each obstacle according to the original point cloud of each obstacle in the target area.

[0067] Optionally, the target area can be an area that the drone will pass through during operation, or other possible areas. The embodiments of this application do not limit this.

[0068] Optionally, the original point cloud of each obstacle can include the point cloud data of each obstacle. This point cloud data refers to a set of vectors in a three-dimensional coordinate system. These vectors are usually represented in the form of three-dimensional coordinates of X, Y, and Z, and generally mainly used to represent the outer surface shape of an object and / or the geometric position information of an object.

[0069] Optionally, each obstacle can include: a plurality of fixed obstacles and at least one connecting obstacle connected between the plurality of fixed obstacles.

[0070] Optionally, the plurality of fixed obstacles can be utility poles and / or power towers, and the plurality of connecting obstacles can be wires and / or optical fibers.

[0071] Optionally, the attribute information of each obstacle can include the shape, size, attitude, position, quantity, and / or elevation information of each obstacle. Specifically, the attribute information of the fixed obstacle can include the shape, size, position, and / or attitude of the fixed obstacle, and the attribute information of the connecting obstacle can include the quantity, position, and / or elevation information of the connecting obstacle.

[0072] Exemplarily, refer to Figure 2 , Figure 2 shows the shapes of two types of utility poles and two types of power towers. For example, Figure 2 (a) in shows a utility pole with a triangular shape, Figure 2 (b) in shows a utility pole with a shape of an upside-down character, Figure 2 (c) in shows a power tower with a shape of an owl, Figure 2 (d) in shows a power tower with a shape of a T. Naturally, the shapes of the utility poles and / or power towers can be as shown in Figure 2The shape shown may also be other shapes that meet relevant standards or regulations. The embodiments of the present application do not limit this.

[0073] Optionally, the attitude of each obstacle may be parallel to the ground or at a certain angle to the ground. The embodiments of the present application do not limit this.

[0074] Optionally, the elevation information may include the distance from any point on each connected obstacle along the vertical line direction to the absolute reference plane. The absolute reference plane may be the ground or any arbitrarily set reference plane. The embodiments of the present application do not limit this.

[0075] It should be noted that since the point cloud information includes the three-dimensional coordinate information of each point of each obstacle, by using the point cloud information of each obstacle to determine the attribute information of each obstacle, the accuracy of the attribute information can be improved.

[0076] Step 1002: Obtain the vector models of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle.

[0077] Optionally, the preset vector model set may be a set including vector models of all utility poles and / or transmission towers that meet relevant standards and / or regulations, or a set including vector models of some utility poles and / or transmission towers that meet relevant standards and / or regulations. Moreover, the vector models in the preset vector model set can be adjusted according to actual needs. The embodiments of the present application do not limit this.

[0078] It should be noted that by obtaining the vector models of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle, the obtained vector models of each fixed obstacle can be ensured to be complete. In this way, the problem that the generated model may have missing information can be solved, and the accuracy of the generated model can be improved.

[0079] Step 1003: Generate the models of each connected obstacle according to the attribute information of each fixed obstacle and the attribute information of each connected obstacle.

[0080] Optionally, the models of each connected obstacle can be generated according to the positions of each fixed obstacle and the elevation information of each connected obstacle.

[0081] Optionally, the models of each connected obstacle may be vector models generated according to the actual positions of each fixed obstacle and the actual positions of each connected obstacle. The models of each connected obstacle may also be vector models generated according to the actual positions of each fixed obstacle and preset parabola parameters. The embodiments of the present application do not limit this.

[0082] Since each connecting obstacle is connected between the multiple fixed obstacles, based on the attribute information of each fixed obstacle and the attribute information of each connecting obstacle, a model of each connecting obstacle is generated. In this way, the models of the connecting obstacles are more in line with the actual situation, and the accuracy of the generated models of the connecting obstacles can be improved.

[0083] Step 1004: Connect the vector models of the fixed obstacles and the models of the connecting obstacles to obtain a target model.

[0084] Optionally, the actual positional relationship and / or the actual connection relationship between the fixed obstacles and the connecting obstacles can be determined according to the original point clouds of the fixed obstacles and the connecting obstacles, and the vector models of the fixed obstacles and the models of the connecting obstacles are connected based on the actual positional relationship and / or the actual connection relationship.

[0085] Exemplarily, Figure 3 shows a target model obtained after connecting three connecting obstacles and two fixed obstacles. Refer to Figure 3 , where the two fixed obstacles are the triangular power pole T1 and the triangular power pole T2, and the three connecting obstacles are the wire L1, the wire L2, and the wire L3. Among them, the ground G is used as the absolute base surface for determining elevation information. The wire L1 is connected between the connection point P1 on the triangular power pole T1 and the connection point P2 on the triangular power pole T2. The wire L2 is connected between the connection point P2 on the triangular power pole T1 and the connection point P3 on the triangular power pole T2. The wire L3 is connected between the connection point P3 on the triangular power pole T1 and the connection point P6 on the triangular power pole T2.

[0086] It should be noted that since the vector models of the fixed obstacles and the models of the connecting obstacles are more complete and have higher accuracy, connecting the vector models of the fixed obstacles and the models of the connecting obstacles to obtain a target model can achieve the effect of improving the accuracy of the obtained target model.

[0087] In an embodiment of the present application, by determining the attribute information of each obstacle in the target area according to the original point cloud of each obstacle in the target area, obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle, generating the model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle, and performing connection processing on the vector models of each fixed obstacle and the models of each connecting obstacle, a target model is obtained. Among them, obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle can ensure that the obtained vector models of each fixed obstacle are complete. In addition, generating the model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle can make the models of each connecting obstacle more in line with the actual situation, and further can achieve the effect of improving the accuracy of the generated model. In this way, the problem that the generated model may have missing information is solved.

[0088] In a possible implementation manner, referring to Figure 4 , before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it further includes:

[0089] Step 1005: Obtain the original image of the target area.

[0090] Optionally, the original image may include at least one obstacle in the target area.

[0091] Optionally, the original image may be multiple photos obtained by a photographic device, or multiple consecutive video stream images obtained by a video camera device.

[0092] For example, a surveying and mapping unmanned aerial vehicle (UAV) can be used to obtain the original image of the target area. When the surveying and mapping UAV is flying in the target area, the sampling device of the surveying and mapping UAV can take pictures or continuously photograph at a certain interval.

[0093] In addition, when the surveying and mapping UAV is flying in the target area, it can fly back and forth along multiple straight lines while obtaining the original image of the target area so that the acquisition range of the surveying and mapping UAV covers the target area, or it can also fly in the target area along other regular flight routes and obtain the original image of the target area. The embodiments of the present application do not limit this.

[0094] Step 1006: Generate the original point cloud of each obstacle in the target area according to the original image.

[0095] Optionally, the original point cloud of each obstacle may include the point cloud data of any point in each obstacle in the target area.

[0096] Generate the original point cloud of each obstacle from the original image, which facilitates subsequent operations, such as determining the attribute information of each obstacle.

[0097] In a possible implementation, refer to Figure 5 , before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it includes:

[0098] Step 1007: Perform segmentation processing on the original image of the target area to obtain the original images of each obstacle.

[0099] Optionally, the segmentation processing may be to separate the images corresponding to each object in the original image.

[0100] For example, the utility poles and / or electric towers and / or wires in the target area can be regarded as obstacles, and each utility pole and / or each electric tower and / or each wire can be separated one by one to obtain the original images of each utility pole and / or each electric tower and / or each wire.

[0101] Step 1008: Determine the original point cloud of each obstacle according to the original images of each obstacle.

[0102] Optionally, the original point cloud of each obstacle can be determined based on the same coordinate system.

[0103] In this way, only the original point cloud of each obstacle needs to be determined, rather than the point cloud information of other objects in the target area except for each obstacle. It can achieve the effect of reducing the computing pressure and improving the generation efficiency.

[0104] Furthermore, the performing segmentation processing on the original image of the target area to obtain the original images of each obstacle includes:

[0105] Perform segmentation processing on the original image of the target area based on the instance segmentation algorithm to obtain the original images of each obstacle.

[0106] Optionally, through the instance segmentation algorithm, the semantic segmentation result of each point in the original image of the target area, the positions of each obstacle, the postures of each obstacle, and the size information of each obstacle can be obtained. Through the segmentation result, the original images of each obstacle can be further obtained. The embodiments of the present application do not limit this.

[0107] It should be noted that the instance segmentation algorithm is an algorithm that can automatically frame different instances from an image using the object detection method and then perform per-pixel marking in different instance areas using the semantic segmentation method. Performing segmentation processing on the original image of the target area based on the instance segmentation algorithm can accurately determine the point cloud information of each pixel point in the image, and thus can improve the accuracy of determining the point cloud information of each obstacle.

[0108] In a possible implementation, refer to Figure 6 , before determining the attribute information of each obstacle based on the original point cloud of each obstacle in the target area, it further includes:

[0109] Step 1009: Determine the original point cloud of the target area according to the original image of the target area.

[0110] Generating the original point cloud of each obstacle through the original image facilitates subsequent operations, such as determining the attribute information of each obstacle.

[0111] Step 1010: Perform segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area.

[0112] Optionally, the segmentation processing may be to separate the original point clouds corresponding to each object in the target area.

[0113] For example, the utility poles and / or electric towers and / or wires in the target area can be used as obstacles. After obtaining the original point cloud of the target area in step 1009, the original point clouds belonging to each utility pole and / or each electric tower and / or each wire are separated respectively, and the original point clouds of each utility pole and / or each electric tower and / or each wire can be obtained.

[0114] In this way, the point cloud information of each pixel point in the image can be accurately determined, and thus the accuracy of determining the point cloud information of each obstacle can be improved.

[0115] Furthermore, the performing segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area includes:

[0116] Based on an instance segmentation algorithm, perform segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area.

[0117] In a possible implementation, refer to Figure 7 , the determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area includes:

[0118] Step 1011: Perform clustering processing on the original point cloud of each obstacle to obtain the clustered point cloud.

[0119] Optionally, the clustering processing may be an operation of classifying data according to specific rules.

[0120] Optionally, performing clustering processing on the original point cloud of each obstacle may specifically be performing density clustering processing on the original point cloud of each obstacle.

[0121] Optionally, the original point clouds of each obstacle can be clustered by the Density-Based Spatial Clustering of Application with Noise (DBSCAN) algorithm with noise application, or the original point clouds of each obstacle can be clustered by the K-means Clustering Algorithm (K-means).

[0122] It should be noted that by clustering the original point clouds of each obstacle, the original point clouds generated by noise can be removed, making the clustered point clouds closer to the actual situation of each obstacle. Furthermore, the accuracy of the generated model can be improved, thus solving the problem that the generated model may have missing information.

[0123] Step 1012: Determine the attribute information of each obstacle according to the clustered point clouds.

[0124] Exemplarily, referring to Figure 8 , such as Figure 8 in (a) shows the distribution of the original point clouds of each obstacle. As can be seen from Figure 8 in (a), there are still some scattered point clouds distributed around the two main point cloud clusters. These scattered point clouds may be the point clouds generated by noise or the original point clouds of each obstacle that deviate from these two point cloud clusters due to other interferences.

[0125] After clustering the original point clouds of each obstacle, the distribution of the clustered point clouds is as shown in Figure 8 in (b). As can be seen from Figure 8 in (b), after clustering, there are only two point cloud clusters left. After clustering, the original point clouds of each obstacle that deviate from these two point cloud clusters due to other interferences will be added to these two point cloud clusters, and the point clouds generated by noise will be removed.

[0126] It should be noted that determining the attribute information of each obstacle according to the clustered point clouds can make the determined attribute information of each obstacle closer to the actual situation of each obstacle, and further achieve the effect of improving the accuracy of the generated model. Thus, the problem that the generated model may have missing information is solved.

[0127] In a possible implementation, the attribute information includes: the category of the obstacle.

[0128] Optionally, the category of the obstacle can be classified based on the shape, size, and / or posture of each obstacle, etc.

[0129] Further, referring to Figure 9 , obtaining the vector models of the fixed obstacles from a preset set of vector models according to the attribute information of each fixed obstacle includes:

[0130] Step 1013: Obtain a vector model that matches the category of the fixed obstacle from the preset set of vector models as the vector model of the fixed obstacle.

[0131] Exemplarily, the vector models of the fixed obstacles can be obtained from the preset set of vector models according to the shapes, sizes, and postures of the obstacles.

[0132] For example, the obstacle can be a utility pole. The shape of a utility pole is triangular, the length is four meters, and the posture is perpendicular to the ground. Traverse all the vector models in the preset set of vector models. First, filter out the first vector models of all utility poles with a triangular shape, then filter out the second vector models with a length of four meters from all the first vector models, and finally filter out the third vector models perpendicular to the ground from all the second vector models.

[0133] By obtaining a vector model that matches the category of the fixed obstacle from the preset set of vector models as the vector model of the fixed obstacle, in this way, it can be ensured that the obtained vector models of the fixed obstacles are complete, and further, the accuracy of the generated model can be improved. Thus, the problem that the generated model may have missing information is solved.

[0134] In a possible implementation manner, generating the models of the connecting obstacles according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle includes:

[0135] Generating the models of the connecting obstacles according to the first attribute information of each fixed obstacle and the second attribute information of each connecting obstacle.

[0136] Optionally, the first attribute information includes: position information.

[0137] Optionally, the second attribute information includes: elevation information.

[0138] Exemplarily, referring to Figure 11 , generating the models of the connecting obstacles according to the first attribute information of each fixed obstacle and the second attribute information of each connecting obstacle includes:

[0139] Step 1014: Perform parabolic fitting processing according to the position information of the fixed obstacles and the elevation information of each connecting obstacle to obtain the models of the connecting obstacles.

[0140] Optionally, the position information of the fixed obstacles can be used to determine the distances between the fixed obstacles, and can also be used to determine the positions where the connecting obstacles are connected to the fixed obstacles.

[0141] Optionally, the specific manner of the parabolic fitting process can be:

[0142] Determine the connection points of any two fixed obstacles.

[0143] Optionally, the any two fixed obstacles can be two adjacent fixed obstacles, can also be two non - adjacent fixed obstacles, or can also be two fixed obstacles selected by those skilled in the art. The embodiments of the present application do not limit this.

[0144] Optionally, the connection point can be the point at the position on each fixed obstacle for connecting each connecting obstacle. For example, if each fixed obstacle is a utility pole or an electric tower, then each connecting obstacle can be a wire, and the connection point can be the point indicating the position where the wire is connected to the utility pole or the electric tower. The embodiments of the present application do not limit this.

[0145] If there is no point cloud data of the connecting obstacle between the any two fixed obstacles, a vector model of each connecting obstacle can be generated according to the preset parabolic curvature.

[0146] Optionally, the preset parabolic curvature can be set in advance. Specifically, the preset parabolic curvature can be set according to the types of the any two fixed obstacles and the distance between the any two fixed obstacles to obtain the model of each connecting obstacle. The embodiments of the present application do not limit this.

[0147] If there is point cloud data of the connecting obstacle between the any two fixed obstacles, the parabolic equation can be solved by the least - squares method or the Gaussian elimination method to obtain the model of each connecting obstacle.

[0148] Optionally, the parabolic equation can be an equation for determining the shape of each connecting obstacle. Specifically, the position information of each fixed obstacle and the elevation information of each connecting obstacle can be input into the parabolic equation to obtain a parabola, and then the model of each connecting obstacle can be generated according to the parabola. The embodiments of the present application do not limit this.

[0149] It should be noted that performing parabolic fitting processing based on the position information of the fixed obstacles and the elevation information of each connecting obstacle can better simulate the natural state of each connecting obstacle under gravity. In this way, the model of each connecting obstacle obtained is closer to the actual situation of each connecting obstacle, and further, the effect of improving the accuracy of the generated model can be achieved. Thus, the problem that the generated model may have missing information is solved.

[0150] Exemplarily, models of each connecting obstacle can be generated according to the position information of each fixed obstacle and the elevation information of each connecting obstacle.

[0151] Next, continue to refer to Figure 3 , and take the step of connecting the wire L1 between the connection point P1 on the triangular utility pole T1 and the connection point P2 on the triangular utility pole T2 as an example for explanation.

[0152] After steps 1001 and 1002, the original point clouds and attribute information of the wire L1, the triangular utility pole T1, and the triangular utility pole T2 have been obtained. Then, the elevation information of the wire L1, as well as the position information of the triangular utility pole T1 and the triangular utility pole T2, have also been obtained. Since the connection point P1 and the connection point P2 are respectively located on the triangular utility pole T1 and the triangular utility pole T2, the position information of the connection point P1 and the connection point P2 can also be determined, such as the straight-line distance between the connection point P1 and the connection point P2.

[0153] Then, according to the straight-line distance between the connection point P1 and the connection point P2 and the elevation information of each point on the wire L1, parabolic fitting processing is performed. In this way, models of each connecting obstacle can be obtained.

[0154] It should be noted that performing parabolic fitting processing according to the position information of the utility pole T1 and the utility pole T2 and the elevation information of the wire L1 can better simulate the natural state of the wire L1 under gravity. In this way, the model of the wire L1 obtained is closer to the actual situation of the wire L1 under the influence of gravity, and further, the effect of improving the accuracy of the generated model can be achieved. Thus, the problem that the generated model may have missing information is solved.

[0155] Next, through Figure 12 an example of the method for generating a model of an obstacle applied to simulate a model of a wire and a utility pole or an electric tower will be given for detailed explanation.

[0156] Exemplarily, Figure 12 a flowchart of a method for generating a model of an obstacle is provided. Refer to Figure 12 , and this method includes:

[0157] Step 2001: Obtain the original image of the target area.

[0158] Optionally, the original image may include at least one obstacle in the target area.

[0159] Optionally, the original image may be multiple photos obtained by a photographic device, or multiple consecutive video stream images obtained by a video camera device.

[0160] Step 2002: Generate the original point cloud of each obstacle in the target area based on the original image.

[0161] Optionally, the original point cloud of each obstacle may include the point cloud data of any point in each obstacle in the target area.

[0162] Generating the original point cloud of each obstacle through the original image facilitates subsequent operations, such as determining the attribute information of each obstacle.

[0163] Step 2003: Determine the original point cloud of the target area based on the original image of the target area.

[0164] Generating the original point cloud of each obstacle through the original image facilitates subsequent operations, such as determining the attribute information of each obstacle.

[0165] Step 2004: Perform segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area.

[0166] Optionally, the segmentation processing may be to separate the original point clouds corresponding to the respective objects in the target area.

[0167] For example, power poles and / or electric towers and / or wires in the target area may be used as obstacles. After obtaining the original point cloud of the target area in step 1009, the original point clouds belonging to each power pole and / or each electric tower and / or each wire are separated respectively, and the original point clouds of each power pole and / or each electric tower and / or each wire can be obtained.

[0168] Optionally, based on an instance segmentation algorithm, the original point cloud of the target area may be segmented to obtain the original point cloud of each obstacle in the target area.

[0169] In this way, the point cloud information of each pixel point in the image can be accurately determined, and further, the accuracy of determining the point cloud information of each obstacle can be improved.

[0170] Step 2005: Determine the attribute information of each obstacle based on the original point cloud of each obstacle in the target area.

[0171] Optionally, the original point cloud of each obstacle may include the point cloud data of each obstacle.

[0172] It should be noted that since the point cloud information includes the three-dimensional coordinate information of each point of each obstacle, determining the attribute information of each obstacle through the point cloud information of each obstacle can achieve the effect of improving the accuracy of the attribute information.

[0173] Step 2006: Perform clustering processing on the original point cloud of each obstacle to obtain the clustered point cloud.

[0174] Optionally, the clustering process may be an operation of classifying data according to specific rules.

[0175] Optionally, density clustering processing may be performed on the original point cloud of each obstacle. Specifically, density clustering processing may be performed on the original point cloud of each obstacle.

[0176] Optionally, the original point cloud of each obstacle may be clustered by the Density-Based Spatial Clustering of Application with Noise (DBSCAN) algorithm with noise application, or the original point cloud of each obstacle may be clustered by the K-means Clustering Algorithm (K-means).

[0177] It should be noted that by clustering the original point cloud of each obstacle, the original point cloud generated by noise can be removed, making the clustered point cloud closer to the actual situation of each obstacle. Furthermore, the accuracy of the generated model can be improved, thus solving the problem of possible information loss in the generated model.

[0178] Step 2007: Determine the attribute information of each obstacle according to the clustered point cloud.

[0179] It should be noted that by determining the attribute information of each obstacle according to the clustered point cloud, the determined attribute information of each obstacle can be closer to the actual situation of each obstacle. Furthermore, the accuracy of the generated model can be improved, thus solving the problem of possible information loss in the generated model.

[0180] Step 2008: Obtain a vector model that matches the category of the fixed obstacle from the preset vector model set as the vector model of the fixed obstacle.

[0181] By obtaining a vector model that matches the category of the fixed obstacle from the preset vector model set as the vector model of the fixed obstacle, the vector models of each fixed obstacle obtained can be ensured to be complete. Furthermore, the accuracy of the generated model can be improved, thus solving the problem of possible information loss in the generated model.

[0182] Step 2009: Perform parabolic fitting processing according to the position information of the fixed obstacle and the elevation information of each connecting obstacle to obtain the models of each connecting obstacle.

[0183] Optionally, the specific method of parabolic fitting processing may be:

[0184] Determine the connection points of any two fixed obstacles.

[0185] If there is point cloud data of the connecting obstacles between any two fixed obstacles, the parabolic equation can be solved according to the least squares method or the Gaussian elimination method to obtain the models of the connecting obstacles.

[0186] If there is no point cloud data of the connecting obstacles between any two fixed obstacles, the vector models of the connecting obstacles can be generated according to the preset parabolic curvature.

[0187] It should be noted that, by performing parabolic fitting according to the position information of the fixed obstacles and the elevation information of the connecting obstacles, the natural state of the connecting obstacles under gravity can be better simulated. In this way, the models of the connecting obstacles obtained are closer to the actual situation of the connecting obstacles, and thus the accuracy of the generated models can be improved, and the problem of possible information loss in the generated models can be solved.

[0188] Step 2010: Connect the vector models of the fixed obstacles and the models of the connecting obstacles to obtain the target model.

[0189] Optionally, the actual position relationship and / or actual connection relationship of the fixed obstacles and the connecting obstacles can be determined according to the original point clouds of the fixed obstacles and the connecting obstacles, and the vector models of the fixed obstacles and the models of the connecting obstacles can be connected based on the actual position relationship and / or actual connection relationship.

[0190] It should be noted that since the vector models of the fixed obstacles and the models of the connecting obstacles are more complete and have higher accuracy, connecting the vector models of the fixed obstacles and the models of the connecting obstacles to obtain the target model can improve the accuracy of the obtained target model.

[0191] The following describes the device, equipment, computer-readable storage medium, etc. for implementing the obstacle model generation method provided by the present application. For the specific implementation process and technical effects, please refer to the above, and will not be repeated below.

[0192] Figure 13 It is a schematic structural diagram of an obstacle model generation device provided by an embodiment of the present application. Refer to Figure 13 This device includes:

[0193] A determination module 301, configured to determine the attribute information of each obstacle according to the original point clouds of the obstacles in the target area;

[0194] An acquisition module 302, configured to acquire the vector models of the fixed obstacles from a preset vector model set according to the attribute information of the fixed obstacles;

[0195] A generation module 303, configured to generate models of the respective connecting obstacles according to the attribute information of the respective fixed obstacles and the attribute information of the respective connecting obstacles;

[0196] A connection processing module 304, configured to perform connection processing on the vector models of the respective fixed obstacles and the models of the respective connecting obstacles to obtain a target model.

[0197] Optionally, the generation module 303 may also be configured to generate models of the respective connecting obstacles according to the first attribute information of the respective fixed obstacles and the second attribute information of the respective connecting obstacles.

[0198] Optionally, the generation module 303 may also be configured to perform parabolic fitting processing according to the position information of the fixed obstacles and the elevation information of the respective connecting obstacles to obtain models of the respective connecting obstacles.

[0199] Optionally, the acquisition module 302 may also be configured to acquire a vector model matching the category of the fixed obstacle from a preset set of vector models as the vector model of the fixed obstacle.

[0200] Optionally, the determination module 301 may also be configured to perform clustering processing on the original point cloud of each obstacle to obtain a clustered point cloud.

[0201] According to the clustered point cloud, determine the attribute information of each obstacle.

[0202] Optionally, the determination module 301 may also be configured to determine the original point cloud of the target area according to the original image of the target area.

[0203] Perform segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area.

[0204] Optionally, the acquisition module 302 may also be configured to acquire the original image of the target area.

[0205] Generate the original point cloud of each obstacle in the target area according to the original image.

[0206] The above device is used to execute the method provided in the foregoing embodiment, and its implementation principle and technical effect are similar, and will not be described in detail here.

[0207] The above modules may be one or more integrated circuits configured to implement the above methods. For example: one or more Application Specific Integrated Circuits (ASICs), or, one or more microprocessors, or, one or more Field Programmable Gate Arrays (FPGAs), etc. Again, when a certain module above is implemented in the form of a processing element scheduler code, the processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processors that can call program code. Again, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0208] Figure 13 is a schematic structural diagram of a computer device provided by an embodiment of the present application. Refer to Figure 10 , the computer device 500 includes: a memory 501 and a processor 502. A computer program that can run on the processor 502 is stored in the memory 501. When the processor 502 executes the computer program, the steps in any of the above method embodiments are implemented.

[0209] An embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments can be implemented.

[0210] Optionally, the present application also provides a program product, such as a computer-readable storage medium,

[0211] including a program that is used to execute any of the above interface image rendering method embodiments when executed by a processor.

[0212] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical, mechanical or other forms.

[0213] The unit described as a separation component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0214] In addition, each functional unit in various embodiments of the present invention may be integrated in a processing unit, may exist separately as individual physical units, or two or more units may be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of hardware plus software functional units.

[0215] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit stored in a storage medium includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to execute some steps of the methods in various embodiments of the present invention. The foregoing storage medium includes: USB flash drives, mobile hard disks, read-only memories (English: Read-Only Memory, abbreviated as: ROM), random access memories (English: Random Access Memory, abbreviated as: RAM), magnetic disks or optical disks and other various media that can store program codes.

[0216] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0217] The above is only the preferred embodiment of this application and is not used to limit this application. For those skilled in the art, this application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included in the protection scope of this application.

Claims

1. A method for generating a model of obstacles, characterized in that, it includes: Determine the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, where the obstacles include: a plurality of fixed obstacles and at least one connecting obstacle connected between the plurality of fixed obstacles; Obtain the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle; Generate the model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle; Perform connection processing on the vector models of each fixed obstacle and the models of each connecting obstacle to obtain a target model; The step of generating the model of each connecting obstacle according to the attribute information of each fixed obstacle and the attribute information of each connecting obstacle includes: Perform parabolic fitting processing according to the position information of each fixed obstacle and the elevation information of each connecting obstacle to generate the model of each connecting obstacle.

2. The model generation method according to claim 1, characterized in that, the attribute information includes: the category of the obstacle; The step of obtaining the vector model of each fixed obstacle from a preset vector model set according to the attribute information of each fixed obstacle includes: Obtain a vector model matching the category of the fixed obstacle from a preset vector model set as the vector model of the fixed obstacle.

3. The model generation method according to claim 1, characterized in that, The step of determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area includes: Perform clustering processing on the original point cloud of each obstacle to obtain the clustered point cloud; Determine the attribute information of each obstacle according to the clustered point cloud.

4. The model generation method according to any one of claims 1-3, characterized in that, Before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it includes: Perform segmentation processing on the original image of the target area to obtain the original images of each obstacle; Determine the original point cloud of each obstacle respectively according to the original images of each obstacle.

5. The model generation method according to claim 4, characterized in that, The step of performing segmentation processing on the original image of the target area to obtain the original images of each obstacle includes: Based on an instance segmentation algorithm, perform segmentation processing on the original image of the target area to obtain the original images of each obstacle.

6. The model generation method according to any one of claims 1-3, characterized in that, Before determining the attribute information of each obstacle according to the original point cloud of each obstacle in the target area, it further includes: Determine the original point cloud of the target area according to the original image of the target area; Perform segmentation processing on the original point cloud of the target area to obtain the original point cloud of each obstacle in the target area.

7. The model generation method according to claim 6, characterized in that, Performing segmentation processing on the original point cloud of the target area to obtain the original point clouds of the obstacles in the target area includes: Performing segmentation processing on the original point cloud of the target area based on an instance segmentation algorithm to obtain the original point clouds of the obstacles in the target area.

8. The model generation method according to any one of claims 1-3, characterized in that, Before determining the attribute information of each obstacle according to the original point clouds of the obstacles in the target area, it further includes: Obtaining the original image of the target area; Generating the original point clouds of the obstacles in the target area according to the original image.

9. An obstacle model generation device, characterized in that, The device includes: A determination module for determining the attribute information of each obstacle according to the original point clouds of the obstacles in the target area; An acquisition module for acquiring the vector models of the fixed obstacles according to the attribute information of the fixed obstacles from a preset vector model set; A generation module for generating the models of the connecting obstacles according to the attribute information of the fixed obstacles and the attribute information of the connecting obstacles; A connection processing module for performing connection processing on the vector models of the fixed obstacles and the models of the connecting obstacles to obtain a target model; The generation module is specifically configured to perform parabolic fitting processing according to the position information of the fixed obstacles and the elevation information of the connecting obstacles to generate the models of the connecting obstacles.

10. A computer device, characterized in that, It includes: A memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8 above.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images