WO2020233724A1

WO2020233724A1 - Visual slam-based grid operating environment map construction method and system

Info

Publication number: WO2020233724A1
Application number: PCT/CN2020/092144
Authority: WO
Inventors: 彭林; 于海; 王鹤; 钱堃; 徐敏; 侯战胜; 王刚; 鲍兴川; 韩海韵; 何志敏; 朱亮
Original assignee: 全球能源互联网研究院有限公司; 国家电网有限公司; 国网江苏省电力有限公司
Priority date: 2019-05-23
Filing date: 2020-05-25
Publication date: 2020-11-26
Also published as: CN110276826A

Abstract

The present invention provides a visual SLAM-based grid operating environment map construction method and system. The method comprises: obtaining grid operating environment sub-maps created in stages, and key frame images corresponding to the sub-maps; respectively performing closed-loop detection on the key frame images corresponding to the sub-maps to generate superimposed images among the sub-maps; and respectively calculating, according to the superimposed images, attitudes corresponding to the sub-maps in a global coordinate system, and connecting the sub-maps according to the attitudes corresponding to the sub-maps to generate a grid operating environment global map. By means of the implementation of the present invention, the process of creating the global map is converted into the creation of local sub-maps in stages so as to avoid the problem of large mapping error even mapping failure of the existing visual SLAM method in the indoor substation operating environment having complex environment and large mapping scale, improve the accuracy of mapping in the grid operating environment, greatly shorten the mapping time, and improve the mapping efficiency.

Description

Method and system for constructing power grid operating environment map based on visual SLAM

Technical field

The invention relates to the technical field of three-dimensional map construction, in particular to a method and system for constructing a power grid operating environment map based on visual SLAM.

Background technique

With the continuous development of industrial automation and intelligence, the internal structure of the substation is highly integrated, and the logical relationship is more complicated, which leads to the continuous increase of the difficulty of the operation of the operators. If a three-dimensional map of the substation can be established, the operation content and experience knowledge can be integrated into the three-dimensional map to provide intuitive and convenient operation guidance for the operators. It is beneficial to improve the work quality, efficiency and safety level of on-site operation of the power grid. Among them, using the wearable equipment of inspection workers to simultaneously establish a three-dimensional model of the substation operating environment in daily inspection work is an effective way to build maps.

Visual SLAM is a technology that uses a camera as the only external sensor to locate itself in an unknown environment and create an environment map at the same time. Due to the complex environment and large mapping scale in the indoor substation operation environment, traditional visual SLAM methods often have large mapping errors and even mapping failures in this environment.

Summary of the invention

In view of this, the embodiments of the present invention provide a method and system for constructing a power grid operating environment map based on visual SLAM to overcome the traditional SLAM method in the prior art that often has large mapping errors or even large errors in the indoor substation operating environment. The problem of failed map creation.

The embodiment of the present invention provides a method for constructing a power grid operating environment map based on visual SLAM, including: acquiring each sub-map of the power grid operating environment established in stages and the key frame image corresponding to each sub-map; Perform closed-loop detection on the key frame images corresponding to the sub-maps to generate overlapping images between the sub-maps; calculate the corresponding poses of the sub-maps in the global coordinate system according to the overlapping images, and Sub-map splicing is performed on the sub-maps according to the poses corresponding to the sub-maps to generate a global map of the power grid operating environment.

Optionally, the closed-loop detection of the key frame images corresponding to the sub-maps to generate overlapping images between the sub-maps includes: using the DBoW2 algorithm to perform the key frame images corresponding to the sub-maps Perform closed-loop detection to generate pairs of closed-loop matching feature points; screen each pair of closed-loop matching feature points by using an anti-polar geometric constraint method, and eliminate each of the closed-loop matching feature points that do not meet the requirements of the anti-polar geometric constraint method Yes; according to each of the selected closed-loop matching feature point pairs, the overlapping images are generated.

Optionally, the respectively calculating the corresponding poses of the sub-maps in the global coordinate system according to the overlapping images includes: generating, according to the overlapping images, a three-dimensional image of the overlapping area between the sub-maps Point cloud; according to the three-dimensional point cloud of the overlapping area between the sub-maps, the point cloud registration algorithm is used to calculate the transformation relationships between the sub-maps; the sub-maps are calculated separately according to the transformation relationships Corresponding poses in the global coordinate system, and sub-map splicing the sub-maps according to the poses corresponding to the sub-maps to generate the global map.

Optionally, the generating a three-dimensional point cloud of the overlapping area between the sub-maps according to the overlapping image includes: acquiring each of the key frame images including the overlapping image, and acquiring each of the key frames The pose information corresponding to the image; the three-dimensional point cloud of a single frame image is generated according to each key frame image; the three-dimensional point cloud corresponding to each key frame image is calculated according to the pose information corresponding to each key frame image The clouds are spliced to generate a three-dimensional point cloud of the overlapping area.

Optionally, the respectively calculating the corresponding poses of the sub-maps in the global coordinate system according to the transformation relations includes: constructing the directed weighted graphs of the sub-maps according to the transformation relations. ; Depth-first traversal is performed on each of the directed weighted maps respectively, and the pose corresponding to each sub-map in the global coordinate system is calculated.

Optionally, the method for constructing a power grid operating environment map based on visual SLAM further includes: acquiring operating information of the power grid operating environment; adding the operating information to the global map to generate a complete global map .

Optionally, the method for constructing a map of the power grid operating environment based on visual SLAM further includes: comparing the complete global map, each sub-map, each key frame image, each overlapping image, and each sub-map corresponding pose Store it.

The embodiment of the present invention also provides a system for constructing a grid operating environment map based on visual SLAM, including: an information acquisition module for acquiring sub-maps of the grid operating environment established in stages, and corresponding sub-maps Key frame image; overlapping image generation module, used to respectively perform closed-loop detection on the key frame image corresponding to each sub-map, and generate each overlapping image between each sub-map; global map construction module, used to generate each sub-map according to the Each overlapping image calculates the corresponding pose of each sub-map in the global coordinate system, and stitches each sub-map into sub-maps according to the pose corresponding to each sub-map to generate the global operating environment of the power grid. map.

An embodiment of the present invention also provides an electronic device, including: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory is stored with computer instructions, and the processor executes the The computer instructions are used to execute the method for constructing a visual SLAM-based power grid operating environment map provided by the embodiment of the present invention.

The embodiment of the present invention also provides a computer-readable storage medium that stores computer instructions, and the computer instructions are used to make the computer execute the visual SLAM-based Construction method of power grid operating environment map.

The technical scheme of the present invention has the following advantages:

The embodiment of the present invention provides a method for constructing a power grid operating environment map based on visual SLAM. By acquiring each sub-map of the power grid operating environment established in stages and the key frame image corresponding to each sub-map, each key frame image Perform closed-loop detection to obtain overlapping images between each sub-map, and then calculate the pose of each sub-map in the global coordinate system based on the overlapping image, and stitch each sub-map according to the pose to generate a global map of the power grid operating environment, thereby By converting the process of establishing a global map to the establishment of local sub-maps in stages, and then through the image overlap relationship between each sub-map, the pose of each sub-map in the global coordinate system can be calculated, and then each local The sub-maps are spliced together to form a complete global map, thereby avoiding the problem of large mapping errors or even failure of mapping in the indoor substation operation environment with complex mapping and large scale in the existing visual SLAM method, and improving the power grid operation environment The accuracy of map construction has greatly shortened the construction time and improved the efficiency of map construction.

Description of the drawings

In order to explain the specific embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings that need to be used in the specific embodiments or the description of the prior art. Obviously, the appendix in the following description The drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

FIG. 1 is a flowchart of a method for constructing a power grid operating environment map based on visual SLAM in an embodiment of the present invention;

2A is a schematic diagram of the topological relationship between different sub-maps in an embodiment of the present invention;

2B is another schematic diagram of the topological relationship between different submaps in an embodiment of the present invention;

3 is a schematic structural diagram of a system for constructing a power grid operating environment map based on visual SLAM in an embodiment of the present invention;

Fig. 4 is a schematic structural diagram of an electronic device in an embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of the embodiments of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the protection scope of the present invention.

The technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

The embodiment of the present invention provides a method for constructing a power grid operating environment map based on visual SLAM. As shown in FIG. 1, the method for constructing a power grid operating environment map based on visual SLAM specifically includes:

Step S1: Obtain each sub-map of the power grid operating environment established in stages, and the key frame image corresponding to each sub-map.

Specifically, in practical applications, workers in the power grid operating environment can scan the power grid operating environment in sections by equipping RGB-D sensors to obtain environmental images at different locations and stages in the power grid operating environment, and create an environment based on these acquired images The various sub-maps of different locations. While creating each sub-map of the environment, save relevant map data of the sub-map for positioning, such as: equipment information and operation information in the power grid operating environment, and the RGB-D image of the key frame in each sub-map. It should be noted that, in the embodiment of the present invention, the creation process of each sub-map can be implemented by using a map construction method suitable for small-scale in the prior art to obtain a more accurate sub-map. The selected specific implementation The method, the present invention is not limited to this.

Step S2: Perform closed-loop detection on the key frame images corresponding to each sub-map, and generate overlapping images between each sub-map.

Specifically, in an embodiment, in the above step S2, performing closed-loop detection on the key frame image corresponding to each sub-map to generate each overlapping image between each sub-map specifically includes the following steps:

Step S21: Use the DBoW2 algorithm to perform closed-loop detection on the key frame image corresponding to each sub-map, and generate each closed-loop matching feature point pair. In practical applications, the specific process of using the DBoW2 algorithm to achieve closed-loop detection between sub-maps is: first extract features from a set of training images, which are composed of key frame images between two different sub-maps, and then extract The features of is clustered by k-means++ algorithm, and the description subspace is divided into k categories. Then in each divided subspace, continue to use the k-means++ algorithm to do clustering, until the descriptor is generated into a tree structure with the number of layers L.

Then, there are two key frames I and I'to be looped back. Perform feature detection and feature description on the image. Pass each feature through a tree structured dictionary to obtain the BoW2 vector V ₁ =[v ₁ ,…,v _n ] And V ₂ =[v′ ₁ ,…,v′ _n ], according to the BoW2 vector, the image similarity formula between I and I′ can be obtained, see the following formula:

_{_{Wherein, L 1 -score s (V 1}} , V 2) represents a measure of similarity between the vectors V ₁ and V ₂ are the L1 norm, v _i represents the value of the vector V ₁ of a dimension, v _'i represents the vector Values of the same dimension in V ₂ . It should be noted that the above-mentioned specific process of implementing the closed-loop detection between sub-maps by using the DBoW2 algorithm is only a simple description. For the specific detailed process, refer to the solution of using the DBoW2 algorithm for closed-loop detection in the prior art, which will not be repeated here.

Step S22: Use the antipolar geometric constraint method to screen each closed-loop matching feature point pair, and eliminate each closed-loop matching feature point pair that does not meet the requirements of the antipolar geometric constraint method. In practical applications, in the indoor substation environment, due to the existence of a large number of similar features, and the DBoW2 algorithm itself does not use the spatial proximity relationship of the feature points, it is easy to produce false loops in this environment. In the embodiment of the present invention, erroneous loops are eliminated through epipolar geometric constraints to ensure that all closed-loop matching feature point pairs are accurate. Since the epipolar geometric constraints require that the matching feature points in the two images need to meet a certain transformation relationship, the wrong closed loop cannot satisfy this transformation relationship and is therefore excluded. This eliminates the error loop and improves the subsequent global construction. The accuracy of the graph.

Step S23: Generate overlapping images according to the selected closed-loop matching feature point pairs. In practical applications, the image area contained in the filtered closed-loop matching feature point pair in the two key frame images is the overlapping image between the two sub-maps corresponding to the two key frame images.

Step S3: Calculate the corresponding poses of each sub-map in the global coordinate system according to the overlapping images, and stitch the sub-maps according to the poses corresponding to the sub-maps to generate a global map of the power grid operating environment.

Specifically, in an embodiment, in the above step S3, the corresponding poses of each sub-map in the global coordinate system are calculated according to each overlapping image, which specifically includes the following steps:

Step S31: Generate a three-dimensional point cloud of the overlapping area between the sub-maps according to the overlapping images. In practical applications, the specific generation principle of the 3D point cloud is as follows: According to the projection model of the camera, assume that the 3D coordinates of any point p in the space in the camera coordinate system are p _c =[x,y,z] ^T , which is in the image The coordinates of the imaging point in is p=[u,v,d] ^T (where u,v represents the coordinate position of the pixel in the image coordinate system, and d represents the depth value of the pixel), then the corresponding relationship between them As shown in the following formula:

d=z·s

Among them, f _x and f _y represent the focal length of the camera on the x and y axes, c _x and c _y are the coordinates of the camera's aperture center, and s represents the scale factor of the depth map.

Therefore, the three-dimensional space coordinates of each point in the image can be restored according to the color and depth images, and then the space coordinates can be projected into the global coordinates to obtain the three-dimensional points of all the pixels contained in the overlapping area between the sub-maps cloud.

Step S32: According to the three-dimensional point cloud of the overlapping area between the sub-maps, a point cloud registration algorithm is used to calculate each transformation relationship between the sub-maps. In practical applications, the specific calculation process of the transformation relationship is as follows: Assume that the coordinates of the two sub-maps M ₁ and M ₂ in the world coordinate system are W ₁ and W _{2 respectively} , and the transformation relationship between the two sub-maps is

Since there is an overlapping area between the two sub-maps, assuming that a point P can be observed in both sub-maps,

They are the coordinates of the point P in the W ₁ , W ₂ coordinate system, and the coordinates of the point P are known

In the case of ICP (Iterative Closest Point, an algorithm used for point cloud registration) to calculate the transformation relationship between the two sub-maps

Make it satisfy

In the specific application process, it may be that the sub-maps M ₁ and M ₂ are often far apart, which makes it impossible to accurately calculate

Then selecting the map selected in the sub M ₁ to the point P can be observed key frame image K _1, and the key frame image in the sub-select the map M ₂ can be observed at point P K _2, the following equation can be obtained:

among them,

Represents the coordinates of the coordinate system W ₁ in the coordinate system K ₁ ,

Represents the coordinates of the coordinate system K ₂ in the coordinate system K ₁ , that is, the transformation relationship between the key frames K ₂ and K ₁ ,

Indicates the coordinates of the coordinate system W ₂ in the coordinate system K ₂ .

Therefore, we can first obtain

That is, the transformation relationship between two adjacent key frame images K ₁ and K ₂ in the global coordinate system, and then the transformation relationship between the two sub-maps M ₁ and M ₂ can be obtained

Step S33: Calculate the corresponding pose of each sub-map in the global coordinate system according to each transformation relationship, and perform sub-map splicing on each sub-map according to the pose corresponding to each sub-map to generate a global map.

Specifically, in one embodiment, in the above step S33, calculating the corresponding poses of each sub-map in the global coordinate system according to each transformation relationship includes the following steps:

Step S331: Construct a directed weighted graph of each sub-map according to each transformation relationship.

Step S332: Depth-first traversal is performed on each directed weighted map respectively, and the corresponding pose of each sub-map in the global coordinate system is calculated.

In practical applications, according to the transformation relationship calculated in the above step S32, the topological relationship between different sub-maps can be constructed. Assume that M ₁ , M ₂ , M ₃ ... are three established sub-maps,

They are represented as the key frame images inside the sub-maps M ₁ , M ₂ , and M _3. The topological relationship constructed is shown in Figure 2A. The solid lines in the figure represent the key frame images that belong to the same sub-map, and the dashed lines represent detected Loops between sub-maps. Calculate the conversion relationship between sub-maps based on the results of loopback detection

As the global coordinate system, the transformation matrix between the submaps M _i and M _j, topological map can be simplified as shown in FIG. 2B weighted graph has the form of FIG. By traversing this graph, the poses of all sub-maps in global coordinates can be calculated. It should be noted that, in the embodiment of the present invention, the traversal of the weighted directed graph is implemented through a depth-first traversal algorithm. In practical applications, other algorithms may also be used to implement it, and the present invention is not limited thereto.

Specifically, in an embodiment, the above-mentioned method for constructing a visual SLAM-based power grid operating environment map further includes:

Step S4: Obtain operation information of the grid operation environment. In practical applications, the job information includes information about the device and the operating environment. The device information specifically includes the ID, name, location, status, and historical data of the device, and the operating environment information includes the type information of the area. Some precautions related to the area, such as whether it is a dangerous area or not.

Step S5: Add the job information to the global map to generate a complete global map. In practical applications, in order to facilitate the user's use, in addition to the location and location coordinates of the map, the created global map should also contain the information of the devices existing in the map and some operation information of the map environment.

Step S6: Store the complete global map, each sub-map, each key frame image, each overlapping image, and the pose corresponding to each sub-map. In practical applications, a hierarchical map data structure can be used to store map-related data. The stored map data is divided into two parts: the global map and each sub-map. The sub-map is created and saved when the map is created. The map is calculated and generated after the mapping is completed. Among them, each sub-map also contains map data, job information, and key frame image and position relationship information. For example, the data of the map may include the following content: (1) MapPoints, specifically including the number of MapPoints and their spatial locations. (2) KeyFrames. Specifically, it includes the number of KeyFrame, the pose of each frame of image, the feature description of all the feature points in each frame, and the corresponding relationship between these feature points and MapPoint is saved. (3) The relationship between different frame images, that is, the parent key frame of each key frame and the key frame connected to it are saved. By saving the above information, a complete map can be recovered. Among them, the operation information includes the information of the equipment and the information of the operating environment. The information of the equipment includes the ID, name, location, status, and historical data of the equipment. The operating environment information includes the type information of the area and the information of the area. Some precautions, such as whether it is a hazardous area, etc.

Through the above steps S1 to S6, the method for constructing a power grid operating environment map based on visual SLAM provided by the embodiment of the present invention obtains the sub-maps of the power grid operating environment established in stages and the key frame images corresponding to each sub-map. Perform closed-loop detection on each key frame image to obtain the overlapping image between each sub-map, and then calculate the pose of each sub-map in the global coordinate system based on the overlapping image, and stitch each sub-map according to the pose to generate a power grid job A global map of the environment, by converting the process of establishing a global map to the establishment of a staged local sub-map, and then calculating the pose of each sub-map in the global coordinate system through the image overlap relationship between the sub-maps. According to the pose, the local sub-maps are spliced together to form a complete global map, thereby avoiding the problem of large mapping errors or even failure of the existing visual SLAM method in the indoor power substation operation environment with complex mapping and large scale. It improves the accuracy of mapping the power grid operating environment, greatly shortens the mapping time, and improves the mapping efficiency.

The embodiment of the present invention also provides a system for constructing a grid operation environment map based on visual SLAM. As shown in FIG. 3, the system for constructing a grid operation environment map based on visual SLAM includes:

The information acquisition module 1 is used to acquire the sub-maps of the power grid operating environment established in stages and the key frame images corresponding to each sub-map. For details, please refer to the related description of step S1 in the above method embodiment, which will not be repeated here.

The overlapping image generating module 2 is used to perform closed-loop detection on the key frame images corresponding to each sub-map, and generate each overlapping image between each sub-map. For details, please refer to the relevant description of step S2 in the above method embodiment, which will not be repeated here.

The global map building module 3 is used to calculate the corresponding poses of each sub-map in the global coordinate system according to the overlapping images, and to splice each sub-map according to the poses corresponding to each sub-map to generate the grid operating environment Global map. For details, please refer to the relevant description of step S3 in the above method embodiment, which will not be repeated here.

Through the collaboration of the above-mentioned various components, the visual SLAM-based power grid operating environment map construction system provided by the embodiment of the present invention obtains each sub-map of the power grid operating environment established in stages and the key frame image corresponding to each sub-map , Perform closed-loop detection on each key frame image to obtain the overlapping image between each sub-map, and then calculate the pose of each sub-map in the global coordinate system based on the overlapping image, and stitch each sub-map according to the pose to generate a power grid The global map of the working environment, by converting the process of establishing the global map to the establishment of phased local sub-maps, and then calculating the position and pose of each sub-map in the global coordinate system through the image overlap relationship between each sub-map, and then The local sub-maps can be spliced together to form a complete global map according to the pose, thereby avoiding the problem of large mapping errors or even failure of the existing visual SLAM method in the indoor substation operation environment with complex mapping and large scale. , Improve the accuracy of mapping the operating environment of the power grid, and greatly shorten the mapping time and improve the efficiency of mapping.

According to an embodiment of the present invention, an electronic device is also provided. As shown in FIG. 4, the electronic device may include a processor 901 and a memory 902, where the processor 901 and the memory 902 may be connected by a bus or in other ways. Take the bus connection as an example.

The processor 901 may be a central processing unit (Central Processing Unit, CPU). The processor 901 may also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA), or Chips such as other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, or a combination of the above types of chips.

The memory 902, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as the construction of a visual SLAM-based power grid operating environment map in the embodiment of the present invention The program instruction/module corresponding to the method. The processor 901 executes various functional applications and data processing of the processor by running the non-transient software programs, instructions, and modules stored in the memory 902, that is, realizes the visual SLAM-based power grid operating environment map in the foregoing method embodiment The construction method.

The memory 902 may include a program storage area and a data storage area. The program storage area may store an operating system and an application program required by at least one function; the data storage area may store data created by the processor 901 and the like. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory 902 may optionally include memories remotely provided with respect to the processor 901, and these remote memories may be connected to the processor 901 via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

One or more modules are stored in the memory 902, and when executed by the processor 901, the method for constructing a visual SLAM-based power grid operating environment map in the foregoing method embodiment is executed.

The specific details of the above electronic device can be understood by referring to the corresponding related description and effect in the above method embodiment, and will not be repeated here.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing relevant hardware through a computer program. The program can be stored in a computer readable storage medium. At this time, it may include the procedures of the above-mentioned method embodiments. Among them, the storage medium can be magnetic disk, optical disk, read-only memory (Read-Only Memory, ROM), random access memory (RAM), flash memory (Flash Memory), hard disk (Hard Disk Drive) , Abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the foregoing types of memories.

Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations fall into the appended claims. Within the limited range.

Claims

A method for constructing a power grid operating environment map based on visual SLAM, which is characterized in that it includes:

Acquiring each sub-map of the power grid operating environment established in stages and the key frame image corresponding to each sub-map;

Performing closed-loop detection on the key frame images corresponding to the sub-maps to generate overlapping images between the sub-maps;

Calculate the corresponding poses of the sub-maps in the global coordinate system according to the overlapping images, and perform sub-map splicing on the sub-maps according to the poses corresponding to the sub-maps to generate the power grid operation Global map of the environment.
The method for constructing a power grid operating environment map based on visual SLAM according to claim 1, wherein the closed-loop detection of the key frame images corresponding to the sub-maps is performed to generate each sub-map. Overlapping images, including:

Using the DBoW2 algorithm to perform closed-loop detection on the key frame images corresponding to each sub-map, and generate each closed-loop matching feature point pair;

Screening the closed-loop matching feature point pairs by adopting an antipolar geometric constraint method, and rejecting each closed-loop matching feature point pair that does not meet the requirements of the antipolar geometric constraint method;

According to the filtered pairs of closed-loop matching feature points, the overlapping images are generated.
The method for constructing a power grid operating environment map based on visual SLAM according to claim 1, wherein said calculating the respective poses of said sub-maps in a global coordinate system according to said overlapping images respectively comprises:

Generating a three-dimensional point cloud of the overlapping area between the sub-maps according to the overlapping images;

Using a point cloud registration algorithm to calculate the transformation relations between the sub-maps according to the three-dimensional point cloud of the overlapping area between the sub-maps;

Calculate the corresponding poses of the sub-maps in the global coordinate system according to the transformation relations, and perform sub-map splicing on the sub-maps according to the poses corresponding to the sub-maps to generate the global map .
The method for constructing a power grid operating environment map based on visual SLAM according to claim 3, wherein the generating a three-dimensional point cloud of the overlapping area between the sub-maps according to the overlapping image comprises:

Acquiring each of the key frame images including the overlapping images, and acquiring the pose information corresponding to each of the key frame images;

Generating a three-dimensional point cloud of a single frame image according to each of the key frame images;

According to the pose information corresponding to each of the key frame images, the three-dimensional point clouds corresponding to each of the key frame images are spliced to generate a three-dimensional point cloud of the overlapping area.
The method for constructing a power grid operating environment map based on visual SLAM according to claim 3, wherein the calculation of the corresponding poses of the sub-maps in the global coordinate system according to the transformation relations includes:

Constructing the directed weighted graph of each sub-map according to the transformation relations;

Depth-first traversal is performed on each of the directed weighted maps respectively, and the pose corresponding to each sub-map in the global coordinate system is calculated.
The method for constructing a power grid operating environment map based on visual SLAM according to any one of claims 1 to 5, further comprising:

Obtaining operation information of the grid operation environment;

The job information is added to the global map to generate a complete global map.
The method for constructing a power grid operating environment map based on visual SLAM according to claim 6, characterized in that it further comprises:

Store the complete global map, each sub-map, each key frame image, each overlapping image, and the pose corresponding to each sub-map.
A visual SLAM-based power grid operating environment map construction system is characterized in that it includes:

An information acquisition module for acquiring each sub-map of the power grid operating environment established in stages, and the key frame image corresponding to each of the sub-maps;

The overlapping image generating module is configured to perform closed-loop detection on the key frame images corresponding to the sub-maps to generate overlapping images between the sub-maps;

The global map construction module is used to calculate the corresponding poses of the sub-maps in the global coordinate system according to the overlapping images, and perform sub-maps on the sub-maps according to the poses corresponding to the sub-maps Splicing to generate a global map of the grid operating environment.
An electronic device, characterized in that it comprises:

A memory and a processor, the memory and the processor are communicatively connected to each other, the memory is stored with computer instructions, and the processor executes the computer instructions by executing any one of claims 1-7 The construction method of power grid operating environment map based on visual SLAM is described.
A computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the vision-based SLAM according to any one of claims 1-7. Method of constructing a map of the power grid operating environment.