Figure consistency verification method and device based on field image, and electronic equipment

Abstract

The invention relates to an image consistency verification method and device based on a field image and electronic equipment, and the method comprises the steps: obtaining a plurality of local field images, shot by an unmanned plane, of a power grid at a vertical visual angle; the field images are recognized, recognition results comprise the positions, the shapes and the sign information of the power grid devices, and the power grid devices with the same sign information are the same power grid device; splicing the different local field images based on the shapes of the same power grid equipment in the different local field images to obtain a global field image of the power grid; determining a global actual topological structure of the power grid based on the position and mark information of each power grid device in the global field image and the connection relationship between different power grid devices; and verifying the global actual topological structure of the power grid and the power grid model topological structure. The problem that an existing graph consistency verification method can only verify the connection relation between the power grid devices is solved.

Description

Technical Field

[0001] The present application relates to the field of power grids, and in particular to a method, device and electronic equipment for verifying image-reality consistency based on on-site images. Background Art

[0002] In the distribution network, the topological relationship between the grid equipment is the focus of attention, which determines the operation status of the distribution network. However, in the actual scenario, due to the scene environment factors, the actual topology of the distribution network may be different from the ideal topology, and then it is necessary to perform a consistency check, that is, to determine the difference between the actual topology of the distribution network and the designed distribution network model topology.

[0003] In a current method for verifying the consistency between images and reality, a drone is used to take a live image of the distribution network and identify the topological relationship between the power grid devices in the live image, and compare it with the topological structure of the distribution network model to determine the difference between the two. Then, in the current method for verifying the consistency between images and reality, local comparison is mainly performed based on different live images, that is, whether the topological relationship between the power grid devices in each live image is correct or not is determined respectively. However, this method for verifying the consistency between images and reality can usually only verify the connection relationship between each power grid device, but cannot verify the relative position relationship between each power grid device. For example, there are power grid devices A and B in one image, and there are power grid devices B and C in another image. At this time, only the connection relationship between power grid devices A and B and the connection relationship between power grid devices B and C can be verified, but it is impossible to verify whether the relative position relationship between power grid devices A, B and C is correct.

[0004] There is currently no effective solution to the problem that the current on-site image-based consistency verification method can usually only verify the connection relationship between various power grid devices. Summary of the invention

[0005] The present invention provides a method, device and electronic device for verifying the consistency between images and reality based on on-site images, so as to solve the problem that the current method for verifying the consistency between images and reality based on on-site images can usually only verify the connection relationship between various power grid devices.

[0006] In a first aspect, the present invention provides a method for verifying image-reality consistency based on a scene image, comprising: Acquire several local on-site images of the power grid from a vertical perspective taken by a drone; Recognize the plurality of on-site images respectively, wherein the recognition results include the position, shape and mark information of the power grid equipment, and the power grid equipment with the same mark information is the same power grid equipment; splicing different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid; Determine the actual global topology of the power grid based on the location and mark information of each power grid device in the global on-site image and the connection relationship between different power grid devices; A global actual topology of the power grid and a model topology of the power grid are verified.

[0007] In some embodiments, stitching different local scene images based on the shapes of the same power grid equipment in the different local scene images includes: Determine a reference scene image from among the plurality of local scene images, and determine other local scene images having a portion of power grid equipment identical to the power grid equipment in the reference scene image as scene images to be spliced; The scene image to be spliced ​​and the reference scene image are spliced ​​based on the shapes of the same power grid equipment in the scene image to be spliced ​​and the reference scene image to form a new reference scene image.

[0008] In some embodiments, stitching the scene image to be stitched and the reference scene image based on the shape of the same power grid equipment in the scene image to be stitched and the reference scene image comprises: When the on-site image to be spliced ​​and the reference on-site image have a pair of identical power grid devices, respectively determine shape feature points of the pair of identical power grid devices, where the shape feature points are feature points that can be uniquely determined based on the shape; Moving the positions of the on-site images to be spliced ​​so that the shape feature points of the pair of identical power grid devices overlap; Rotating and scaling the on-site images to be spliced ​​based on the coincident shape feature points so that the shapes of the pair of identical power grid devices coincide with each other; When the scene image to be spliced ​​and the reference scene image have more than two pairs of identical power grid equipment, determining shape feature points of two pairs of identical power grid equipment; The on-site images to be spliced ​​are transformed so that the shape feature points of the two pairs of identical power grid equipment coincide with each other.

[0009] In some of the embodiments, the shape feature point is one of the geometric center, the center of the smallest circumscribed circle, and the center of the largest inscribed circle of the shape.

[0010] In some of these embodiments, the shape feature point is the geometric center of the shape; Rotating and scaling the on-site images to be spliced ​​based on the coincident shape feature points so that the shapes of the pair of identical power grid devices coincide with each other comprises: Rotating the on-site image to be stitched around the coincident shape feature point and continuously calculating the shape boundary distance sequence of the pair of identical power grid devices during the rotation process until the volatility of the shape boundary distance sequence is minimized, the shape boundary distance sequence including the distances of the shape boundaries of the pair of identical power grid devices in each scattering direction of the coincident shape feature point; The on-site images to be spliced ​​are scaled using the coincident shape feature points as scaling base points so that the shapes of the pair of identical power grid devices coincide.

[0011] In some embodiments, scaling the on-site images to be spliced ​​using the coincident shape feature points as scaling base points so that the shapes of the pair of identical power grid devices coincide with each other includes: The on-site image to be spliced ​​is scaled using the coincident shape feature points as scaling base points, and the shape boundary distance sequence of the pair of identical power grid devices is continuously calculated during the scaling process until the scalar average value of the shape boundary distance sequence is minimum.

[0012] In some embodiments, transforming the on-site images to be spliced ​​so that the shape feature points of the two pairs of identical power grid devices coincide with each other includes: Connecting two feature points in the scene image to be stitched to form a first line segment, and connecting two feature points in the reference scene image to form a second line segment; Moving the on-site images to be spliced ​​so that shape feature points of a pair of identical power grid devices overlap; Rotating the scene image to be spliced ​​around the coincident shape feature point so that the first line segment and the second line segment coincide with each other; The on-site images to be spliced ​​are scaled using the coincident shape feature points as scaling base points so that another pair of shape feature points of the same power grid equipment coincide with each other.

[0013] In a second aspect, the present invention provides a device for verifying image-reality consistency based on a scene image, comprising: An image acquisition module is used to acquire a number of local on-site images of the power grid taken by the UAV based on a vertical perspective; An image recognition module is used to recognize the plurality of on-site images respectively, wherein the recognition results include the position, shape and mark information of the power grid equipment, and the power grid equipment with the same mark information are the same power grid equipment; An image stitching module, used for stitching different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid; A topology generation module, used to determine the global actual topology structure of the power grid based on the location and mark information of each power grid device in the global on-site image and the connection relationship between different power grid devices; The topology verification module is used to verify the global actual topology structure of the power grid and the power grid model topology structure.

[0014] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the method for verifying image-reality consistency based on on-site images as described in the first aspect.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for verifying image-reality consistency based on on-site images described in the first aspect are implemented.

[0016] Compared with the related art, the present invention splices several local field images into a global field image. Since the position and mark information of each power grid device in the global field image and the connection relationship between different power grid devices are known, the global actual topology structure of the power grid can be constructed. Moreover, the global actual topology structure is a spatial topology structure containing the relative position information of each node. By comparing the global actual topology structure with the power grid model topology structure, it is possible to determine not only whether the connection relationship between each power grid device is correct, but also whether the relative position relationship between each power grid device is correct, which solves the problem that the current image-based consistency verification method can usually only verify the connection relationship between each power grid device.

[0017] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the method for verifying image-reality consistency based on on-site images provided in this embodiment. DETAILED DESCRIPTION

[0019] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0020] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with general skills in the technical field to which this application belongs. The words "one", "a", "the", "these" and the like in this application do not indicate a quantitative limitation, and they may be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" may mean: A exists alone, A and B exist at the same time, and B exists alone. Generally, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0021] In this embodiment, a method for verifying the consistency of an image with respect to a scene image is provided. Figure 1 is a flow chart of the method for verifying the consistency of an image based on a scene image provided in this embodiment, such as Figure 1 As shown, the process includes step S110, step S120, step S130, step S140 and step S150.

[0022] Step S110, obtaining a plurality of local on-site images of the power grid at a vertical viewing angle taken by the drone.

[0023] Before executing the method for verifying the consistency between the image and the reality in this embodiment, it is necessary to use a drone to photograph the power grid site. Since the method for verifying the consistency between the image and the reality in this embodiment needs to obtain images under a vertical perspective, the drone is required to photograph the power grid site under a vertical perspective. Specifically, during the horizontal flight of the drone, the camera perspective is always vertically downward to avoid perspective deformation of the image as much as possible. For slight perspective deformation of the image caused by the shaking of the drone, the image can be transformed and corrected by perspective correction technology.

[0024] Step S120, identifying a plurality of on-site images respectively, the identification results including the position, shape and mark information of the power grid equipment, and the power grid equipment with the same mark information are the same power grid equipment.

[0025] Grid equipment includes poles, switches, transformers, generators and other equipment. Since the shapes of different grid equipment are different and specific, the grid equipment in the image can be identified through image recognition technology and the shape outline and position (position in the image coordinate system) of each grid equipment can be given. At the same time, the grid equipment is marked with logo information, such as equipment ID and equipment nameplate information, and the logo information of each grid equipment can be identified through image recognition technology to distinguish different grid equipment. In summary, the shape and logo information of grid equipment can be obtained through the currently mature image recognition technology.

[0026] In this embodiment, a neural network model is used to perform the image recognition task. Among them, the shape recognition of the power grid equipment can be achieved through the target detection task and the instance segmentation task, that is, the positioning and shape detection of the power grid equipment are completed by combining the target detection model and the instance segmentation model. For example, the target detection model can adopt YOLOv8, which outputs the bounding box and category label of the power grid equipment. The instance segmentation model can adopt Mask R-CNN, which generates a pixel-level mask based on the target detection and accurately extracts the contour shape of the power grid equipment. The sign information of the power grid equipment can be achieved through the text recognition task, that is, the text recognition model is used to complete the extraction of the sign information. For example, the text recognition model can adopt CRNN, and its output is structured text such as the ID and nameplate information of each power grid equipment. A multi-task model can be obtained by combining multiple models, through which the power grid equipment in the image can be located and the shape and sign information can be identified.

[0027] It should be noted that the image recognition in this step can be realized based on the current mature image recognition technology, which is not the focus of the present invention and will not be further explained.

[0028] Step S130 , splicing different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid.

[0029] In this step, different local scene images are stitched together to form a global scene image. Image stitching is mainly achieved through the same parts in the two images. Specifically, since the shape and logo information of the power grid equipment in each local scene image have been identified in the previous step, it can be judged based on the logo information whether there is the same power grid equipment in any two local scene images. If so, there must be extremely similar or even identical shape graphics (corresponding to the same power grid equipment) in the corresponding two local scene images, and then stitching can be performed based on this pair of shape graphics.

[0030] In this embodiment, a new image stitching method suitable for power grid image stitching is also provided. Wherein, step S130 specifically includes step S131 and step S132.

[0031] Step S131 : determining a reference scene image from among the plurality of local scene images, and determining other local scene images having a portion of power grid equipment identical to the power grid equipment in the reference scene image as scene images to be spliced.

[0032] Step S132 : splicing the scene image to be spliced ​​and the reference scene image based on the shapes of the same power grid equipment in the scene image to be spliced ​​and the reference scene image to form a new reference scene image.

[0033] In this embodiment, a local scene image is randomly used as a reference scene image, and then a scene image to be spliced ​​that can be spliced ​​with the reference scene image is determined, and finally the spliced ​​image is used as a new reference scene image, until there is no local scene image that meets the splicing conditions, and the final reference scene image is used as the global scene image. Among them, the splicing conditions include: there is at least one pair of identical power grid equipment between the scene image to be spliced ​​and the reference scene image (the basis for image splicing), and only part of the power grid equipment in the scene image to be spliced ​​is the same as the power grid equipment in the reference scene image. If all the power grid equipment in the scene image exists in the reference scene image, it means that the scene image does not provide additional power grid topology information, and thus does not need to participate in splicing, so as to improve the efficiency of image splicing.

[0034] Through the above stitching method, local scene images can be stitched into global scene images in an orderly and efficient manner, which has a higher stitching efficiency than other random and disordered stitching methods.

[0035] Further, in this embodiment, step S132 includes step S132a, step S132b, step S132c, step S132d and step S132e.

[0036] Step S132a, when there is a pair of identical power grid devices in the site image to be spliced ​​and the reference site image, shape feature points of the pair of identical power grid devices are determined respectively, and the shape feature points are feature points that can be uniquely determined based on the shape. In this step, the shape feature point is the geometric center of the shape.

[0037] Step S132b, moving the position of the on-site images to be spliced ​​so that a pair of shape feature points of the same power grid equipment overlap.

[0038] Step S132c: rotating and scaling the on-site images to be spliced ​​based on the overlapping shape feature points so that the shapes of a pair of identical power grid equipment overlap.

[0039] Steps S132a, S132b and S132c are mainly used to achieve image stitching when there is only one pair of identical power grid devices in two images. For example, if there is only one power grid device A in the scene image to be stitched and one power grid device B in the reference scene image is the same power grid device, then the power grid device A and the power grid device B constitute a pair of identical power grid devices, which are essentially the same power grid devices but exist in different images.

[0040] Specifically, in the absence of perspective differences, completing the stitching of two images requires determining the translation relationship and rotation relationship between the two images. In this embodiment, the shape feature points of the same power grid equipment are innovatively used as reference points for moving and rotating the on-site images to be stitched. First, the shape feature points of the same power grid equipment in the two images are determined, and the translation operation of the image to be stitched is completed by moving the image to be stitched so that the two corresponding graphic feature points coincide, thereby eliminating the translation error between the two images. Furthermore, it is also necessary to eliminate the rotation angle error and the scaling error (this error is caused by the different shooting heights of the drone or caused by perspective correction) between the image to be stitched and the reference on-site image, so that the shapes of the two identical power grid equipment completely coincide, thereby aligning the two images.

[0041] Furthermore, in the present embodiment, step S132c specifically includes: rotating the scene image to be stitched around the overlapping shape feature points and continuously calculating the shape boundary distance sequence of a pair of identical power grid equipment during the rotation process until the volatility of the shape boundary distance sequence is minimized, the shape boundary distance sequence including the distances of the shape boundaries of a pair of identical power grid equipment in each scattering direction of the overlapping shape feature points; scaling the scene image to be stitched using the overlapping shape feature points as the scaling base point so that the shapes of a pair of identical power grid equipment overlap.

[0042] Since the scaling ratio between the image to be spliced ​​and the reference field image may be different, it is impossible to determine whether the rotation angle error is eliminated by judging whether the shapes of two identical power grid devices are completely overlapped. In this embodiment, considering that for two graphics of the same shape but different sizes, when there is no rotation error between the two graphics (the corresponding point connection line in the two graphics points to the geometric center), the volatility of the shape boundary between the two should be the smallest, such as common regular polygons, matrices, circles and ellipses all meet the above rules, and these graphics are all common power grid equipment shapes. Exemplarily, for two squares of different sizes, when the geometric centers of the two coincide and there is no rotation error, the minimum shape boundary distance between the two is the distance between the shape boundaries of the two in the vertical or horizontal direction, and the maximum shape boundary distance between the two is the distance between the shape boundaries of the two in the 45° oblique direction. At this time, the difference between the maximum and minimum shape boundary distances is the smallest, and the shape boundary distance changes in different directions continuously. When there is any rotation error between the two squares, the difference between the maximum and minimum shape boundary distances will become larger, and the shape boundary distance changes in different directions continuously will be more obvious. Therefore, in this embodiment, the rotation error between the scene image to be spliced ​​and the reference scene image can be effectively eliminated by stopping the rotation of the scene image to be spliced ​​when the volatility of the shape boundary distance sequence is the smallest. For example, the volatility can be characterized by indicators such as the standard deviation, variance, and range of the shape boundary distance sequence.

[0043] After eliminating the rotation error between the scene image to be spliced ​​and the reference scene image, the scaling error between the two can be eliminated. Since the rotation error has been eliminated, scaling the scene image to be spliced ​​with the coincident shape feature points as the scaling base point can make the shapes of the same power grid equipment coincide.

[0044] Further, considering that in actual situations, rotation errors are difficult to be completely eliminated, and there may also be errors in perspective correction, so that the shapes of the same power grid equipment may not completely overlap. Therefore, scaling the on-site image to be spliced ​​with the overlapping shape feature points as the scaling base point so that the shapes of a pair of identical power grid equipment overlap includes: scaling the on-site image to be spliced ​​with the overlapping shape feature points as the scaling base point, and continuously calculating the shape boundary distance sequence of a pair of identical power grid equipment during the scaling process until the scalar average value of the shape boundary distance sequence is the smallest. In this embodiment, the shape overlap of the two identical power grid equipment is measured by the scalar average value of the shape boundary distance sequence of the two identical power grid equipment. It can be understood that when the shapes of the two identical power grid equipment completely overlap, the scalar average value of the shape boundary distance sequence of the two identical power grid equipment is 0, so the scaling is stopped when the scalar average value of the shape boundary distance sequence of the two identical power grid equipment is the smallest, and the image splicing is completed.

[0045] Step S132d: when there are two or more pairs of identical power grid equipment in the scene image to be spliced ​​and the reference scene image, determine the shape feature points of two pairs of identical power grid equipment. In this step, the shape feature point is one of the geometric center of the shape, the center of the smallest circumscribed circle, and the center of the largest inscribed circle.

[0046] Step S132e, transforming the on-site images to be spliced ​​so that the shape feature points of two pairs of identical power grid equipment correspond to each other and overlap.

[0047] Step S132d and step S132e are mainly used to realize image stitching when there are at least two pairs of identical power grid devices in two images. Compared with step S132a, step S132b and step S132c, image stitching can be realized more efficiently and quickly. It only needs to transform the scene image to be stitched so that the shape feature points of two pairs of identical power grid devices coincide with each other. For example, the power grid devices A1 and A2 in the scene image to be stitched are the same devices as the power grid devices B1 and B2 in the reference scene image, respectively. Then, it is only necessary to make the shape feature points of the power grid device A1 coincide with the shape feature points of the power grid device B1 and the shape feature points of the power grid device A2 coincide with the shape feature points of the power grid device B2.

[0048] Among them, step S132e specifically includes: connecting two feature points in the scene image to be spliced ​​to form a first line segment, and connecting two feature points in the reference scene image to form a second line segment; moving the scene image to be spliced ​​so that one pair of shape feature points of the same power grid equipment overlaps; rotating the scene image to be spliced ​​around the overlapping shape feature points so that the first line segment and the second line segment overlap; scaling the scene image to be spliced ​​using the overlapping shape feature points as the scaling base point so that another pair of shape feature points of the same power grid equipment overlaps.

[0049] Exemplarily, it is assumed that the power grid devices A1 and A2 in the scene image to be stitched are the same devices as the power grid devices B1 and B2 in the reference scene image, respectively. Then the line connecting the shape feature points of the power grid devices A1 and B1 is the first line segment, and the line connecting the shape feature points of the power grid devices A2 and B2 is the second line segment. First, the scene image to be stitched can be moved so that the shape feature points of the power grid devices A1 and B1 coincide with each other, so as to eliminate the translation error between the two images; then, the scene image to be stitched is rotated around the shape feature points of the power grid devices A1 and B1 so that the first line segment and the second line segment coincide with each other, so as to eliminate the rotation error between the two images; finally, the scene image to be stitched is scaled with the shape feature points of the power grid devices A1 and B1 as the scaling base point so that the shape feature points of the power grid devices A2 and B2 coincide with each other, so as to eliminate the scaling error between the two images.

[0050] It can be seen from the above description that when there are at least two pairs of identical power grid devices in the on-site image to be stitched and the reference on-site image, image stitching can be completed in a more efficient and rapid manner.

[0051] The above is a specific description of step S130. Through step S130, a global scene image can be quickly and efficiently constructed based on a number of local scene images.

[0052] Step S140, determining the global actual topology of the power grid based on the location and mark information of each power grid device in the global on-site image and the connection relationship between different power grid devices.

[0053] Since the positions of each power grid device (the positions of each power grid device in the local scene image have been identified, so the positions of each power grid device in the spliced ​​global scene image are also known) and the sign information have been identified in the previous step, in the topology recognition, it is only necessary to continue to identify the connection relationship between each power grid device in the global scene image. Power grid devices are mainly connected by cables. Therefore, by detecting the cables and judging the positions of the two ends of the cables, the power grid devices connected to the cables can be determined, that is, the connection relationship between different power grid devices can be determined. Cable recognition can be achieved through instance segmentation tasks. For example, the U-Net with Attention model is used to identify the global scene image, which outputs the pixel-level segmentation mask of the cable, so that the positions of the two ends of the cable in the global scene image can be determined, and then the power grid devices connected to each cable can be determined.

[0054] Step S150, verifying the global actual topology structure of the power grid and the power grid model topology structure.

[0055] Since the location and mark information of each power grid device in the global field image and the connection relationship between different power grid devices are known, the global actual topology of the power grid can be constructed, and the global actual topology is a spatial topology containing the relative position information of each node. By comparing the global actual topology with the power grid model topology, it is possible to determine not only whether the connection relationship between each power grid device is correct, but also whether the relative position relationship between each power grid device is correct, which solves the problem that the current image-based consistency verification method can usually only verify the connection relationship between each power grid device. Among them, the power grid model topology can be established in advance. During the establishment process, the connection relationship between each power grid device can be established based on the power grid CIM model, and the relative position relationship between each power grid device can be established in combination with the map system.

[0056] In this embodiment, a device for verifying the consistency of an image based on a scene image is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. The terms "module", "unit", "subunit", etc. used below can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0057] The image-reality consistency verification device based on on-site images includes: an image acquisition module, an image recognition module, an image splicing module, a topology generation module and a topology verification module.

[0058] The image acquisition module is used to acquire a number of local on-site images of the power grid taken by the UAV based on a vertical perspective.

[0059] The image recognition module is used to recognize several on-site images respectively. The recognition results include the position, shape and mark information of the power grid equipment. The power grid equipment with the same mark information is the same power grid equipment.

[0060] The image stitching module is used to stitch different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid.

[0061] The topology generation module is used to determine the actual global topology of the power grid based on the location and mark information of each power grid device in the global field image and the connection relationship between different power grid devices.

[0062] The topology verification module is used to verify the global actual topology structure of the power grid and the power grid model topology structure.

[0063] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0064] In this embodiment, an electronic device is also provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the method for verifying image-reality consistency based on on-site images in this embodiment.

[0065] In this embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for verifying the consistency between images and reality based on on-site images in this embodiment are implemented.

[0066] It should be understood that the specific embodiments described herein are only used to explain the application, rather than to limit it. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of this application.

[0067] Obviously, the drawings are only some examples or embodiments of the present application. For ordinary technicians in the field, the present application can also be applied to other similar situations based on these drawings without creative work. In addition, it is understandable that although the work done in this development process may be complicated and lengthy, for ordinary technicians in the field, certain changes in design, manufacturing or production based on the technical content disclosed in this application are only conventional technical means and should not be regarded as insufficient content disclosed in this application.

Claims

1. A method for verifying image-reality consistency based on on-site images, characterized in that: include: Acquire several local on-site images of the power grid from a vertical perspective taken by a drone; Recognize the plurality of on-site images respectively, wherein the recognition results include the position, shape and mark information of the power grid equipment, and the power grid equipment with the same mark information is the same power grid equipment; splicing different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid; Determine the actual global topology of the power grid based on the location and mark information of each power grid device in the global on-site image and the connection relationship between different power grid devices; A global actual topology of the power grid and a model topology of the power grid are verified.

2. The method for verifying the consistency between image and reality based on on-site images according to claim 1, characterized in that: The step of stitching different local scene images based on the shapes of the same power grid equipment in the different local scene images comprises: Determine a reference scene image from among the plurality of local scene images, and determine other local scene images having a portion of power grid equipment identical to the power grid equipment in the reference scene image as scene images to be spliced; The scene image to be spliced ​​and the reference scene image are spliced ​​based on the shapes of the same power grid equipment in the scene image to be spliced ​​and the reference scene image to form a new reference scene image.

3. The method for verifying the consistency between image and reality based on on-site images according to claim 2, characterized in that: The step of stitching the scene image to be stitched and the reference scene image based on the shape of the same power grid equipment in the scene image to be stitched and the reference scene image comprises: When the on-site image to be spliced ​​and the reference on-site image have a pair of identical power grid devices, respectively determine shape feature points of the pair of identical power grid devices, where the shape feature points are feature points that can be uniquely determined based on the shape; Moving the positions of the on-site images to be spliced ​​so that the shape feature points of the pair of identical power grid devices overlap; Rotating and scaling the on-site images to be spliced ​​based on the coincident shape feature points so that the shapes of the pair of identical power grid devices coincide with each other; When the scene image to be spliced ​​and the reference scene image have more than two pairs of identical power grid equipment, determining shape feature points of two pairs of identical power grid equipment; The on-site images to be spliced ​​are transformed so that the shape feature points of the two pairs of identical power grid equipment coincide with each other.

4. The method for verifying image-reality consistency based on on-site images according to claim 3, characterized in that: The shape feature point is one of the geometric center of the shape, the center of the smallest circumscribed circle, and the center of the largest inscribed circle.

5. The method for verifying image-reality consistency based on on-site images according to claim 3, characterized in that: The shape feature point is the geometric center of the shape; Rotating and scaling the on-site images to be spliced ​​based on the coincident shape feature points so that the shapes of the pair of identical power grid devices coincide with each other comprises: Rotating the on-site image to be stitched around the coincident shape feature point and continuously calculating the shape boundary distance sequence of the pair of identical power grid devices during the rotation process until the volatility of the shape boundary distance sequence is minimized, the shape boundary distance sequence including the distances of the shape boundaries of the pair of identical power grid devices in each scattering direction of the coincident shape feature point; The on-site images to be spliced ​​are scaled using the coincident shape feature points as scaling base points so that the shapes of the pair of identical power grid devices coincide.

6. The method for verifying the consistency between image and reality based on on-site images according to claim 5, characterized in that: Scaling the on-site images to be spliced ​​using the coincident shape feature points as scaling base points so that the shapes of the pair of identical power grid devices coincide with each other comprises: The on-site image to be spliced ​​is scaled using the coincident shape feature points as scaling base points, and the shape boundary distance sequence of the pair of identical power grid devices is continuously calculated during the scaling process until the scalar average value of the shape boundary distance sequence is minimum.

7. The method for verifying image-reality consistency based on on-site images according to claim 3, characterized in that: Transforming the on-site images to be spliced ​​so that the shape feature points of the two pairs of identical power grid equipment coincide with each other comprises: Connecting two feature points in the scene image to be stitched to form a first line segment, and connecting two feature points in the reference scene image to form a second line segment; Moving the on-site images to be spliced ​​so that shape feature points of a pair of identical power grid devices overlap; Rotating the scene image to be spliced ​​around the coincident shape feature point so that the first line segment and the second line segment coincide with each other; The on-site images to be spliced ​​are scaled using the coincident shape feature points as scaling base points so that another pair of shape feature points of the same power grid equipment coincide with each other.

8. A device for verifying the consistency between images and reality based on on-site images, characterized in that: include: An image acquisition module is used to acquire a number of local on-site images of the power grid taken by the UAV based on a vertical perspective; An image recognition module is used to recognize the plurality of on-site images respectively, wherein the recognition results include the position, shape and mark information of the power grid equipment, and the power grid equipment with the same mark information are the same power grid equipment; An image stitching module, used for stitching different local scene images based on the shapes of the same power grid equipment in the different local scene images to obtain a global scene image of the power grid; A topology generation module, used to determine the global actual topology structure of the power grid based on the location and mark information of each power grid device in the global on-site image and the connection relationship between different power grid devices; The topology verification module is used to verify the global actual topology structure of the power grid and the power grid model topology structure.

9. An electronic device, comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to execute the method for verifying image-reality consistency based on on-site images according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for verifying image-reality consistency based on on-site images according to any one of claims 1 to 7 are implemented.

Images