Image video scene recognition method based on distance prediction information

Abstract

The invention discloses an image video scene recognition method based on distance prediction information, and the method comprises the steps: marking and dividing a distribution region of power grid equipment through a satellite remote sensing image, determining the distribution position information of the power grid equipment in a preset region, and carrying out the inspection of the interior of the preset region through an unmanned plane, the whole preset area is directly inspected, the whole inspection image video data is divided according to the distribution area of the power grid equipment, and the pre-constructed anomaly recognition models are matched according to the corresponding inspection image video data so as to realize recognition of hidden danger anomaly scenes, environment anomaly scenes and the like of the power grid equipment. And therefore, the hidden danger abnormity of the power grid equipment can be inspected. According to the invention, automatic inspection can be realized, the power grid equipment does not need to be distinguished or the power grid equipment does not need to be positioned during inspection, and the inspection is more convenient and higher in efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of image and video scene recognition, and in particular to an image and video scene recognition method based on distance prediction information. Background Art

[0002] At present, the inspection of equipment in the power grid is generally carried out by manual on-site inspection and observation of equipment abnormalities, and automated inspection and identification cannot be achieved. Even if there is currently an inspection based on drones, each grid equipment to be inspected needs to be located before the drone inspection, and then the corresponding inspection route is planned for each grid equipment in turn, and then inspected in turn. If the positioning is inaccurate, inspection errors are likely to occur, resulting in the inability of drones to inspect normally, and then unable to determine what abnormal scene the grid equipment is in, resulting in untimely abnormal judgment, and even affecting the normal operation of the power grid. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides an image and video scene recognition method based on distance prediction information, which can solve the problems of cumbersome inspection operation and poor efficiency in existing manual inspections, as well as the problem that existing drone inspections require each power grid device to be positioned in sequence, which is prone to inaccurate positioning and leads to inspection errors. As a result, it is impossible to determine what abnormal scene the power grid equipment is in, resulting in untimely abnormal judgment and even affecting the normal operation of the power grid.

[0005] In order to solve the above technical problems, the present invention provides the following technical solution: an image video scene recognition method based on distance prediction information, comprising:

[0006] Determine the satellite image of the power grid in the preset area, and mark it on the satellite image of the power grid according to the distribution location of the power grid in the preset area, so as to obtain the distribution area of ​​each power grid equipment in the preset area;

[0007] By controlling the inspection equipment to inspect the preset area, the inspection image video data in the preset area can be obtained. At the same time, during the inspection process, the inspection drone will obtain the positioning information of the inspection drone in real time;

[0008] The inspection image video data is segmented according to the positioning information obtained by the inspection drone and the minimum distribution area of ​​each power grid device in the preset area to generate first inspection image video data and second inspection image video data;

[0009] The first inspection image video data of each power grid device is converted into multiple inspection images after key frame image extraction, and the multiple inspection images are matched with the hidden danger scene recognition model of the power grid device constructed in the system to obtain the hidden danger scene of the power grid device;

[0010] The acquired second inspection image video data is subjected to key frame image extraction to match it with the abnormal environment scene recognition model constructed in the system to determine the type of abnormal environment scene in the preset area.

[0011] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, wherein: the marking in the power grid satellite image map according to the distribution position of the power grid in the preset area includes identifying the distribution area of ​​each power grid device in the power grid satellite image map according to the distribution position of the power grid equipment in the power grid, and obtaining the distribution area of ​​each power grid device by marking the square area, and the distribution area includes the positioning information of the power grid equipment.

[0012] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, wherein: the distribution area includes, determined by the coordinate values ​​of the four corners of the square area, wherein the square area is set according to actual needs, including the minimum square area of ​​the power grid equipment, or the minimum circular area.

[0013] As a preferred solution of the image and video scene recognition method based on distance prediction information described in the present invention, wherein: the acquisition of positioning information passed by the inspection drone includes the need to give the inspection drone a specific inspection area range, that is, the coordinate range of the preset area. The inspection drone inspects the preset area according to the established route to obtain inspection image and video data within the preset area. The inspection image and video data include image and video data of power grid equipment, and also include the surrounding environment without power grid equipment. At the same time, when the drone is shooting, a corresponding preset shooting distance is also set.

[0014] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, wherein: the generation of the first inspection image video data and the second inspection image video data includes, when the inspection drone is inspecting, recording the inspection image video data, obtaining the positioning information, and segmenting the inspection image video data according to the acquisition time of the positioning information and the coordinate information of the positioning information to obtain two types of inspection image video data;

[0015] The first inspection image video data is inspection image video data including power grid equipment, and the second inspection image video data refers to inspection image video data other than the first inspection image video data;

[0016] When there are more than two power grid devices, each power grid device corresponds to one first inspection image video data while passing the first inspection image video data;

[0017] When dividing the image and video data, it is necessary to divide it based on the preset shooting distance of the drone. When dividing the image and video data, it is necessary to consider the preset shooting distance of the drone. When the positioning information of the drone appears outside the area of ​​the power grid equipment, the image and video data taken are still in the distribution area of ​​the power grid equipment. Therefore, the image and video data are still divided into the first inspection image and video data;

[0018] At the same time, during the division, images of drones at different shooting angles corresponding to each power grid device are constructed in the system, and the corresponding features are extracted to construct an image feature set of the power grid equipment. Based on the image feature set, auxiliary identification of the power grid equipment in the image video data is achieved.

[0019] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, wherein: the key frame image extraction of the acquired second inspection image video data includes, in the extraction scene, the types of environmental abnormal scenes include abnormal fire scenes, abnormal mudslide scenes, and abnormal vegetation overgrowth scenes.

[0020] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, wherein: the judgment of the type of abnormal environmental scene in the preset area includes timely judging the scene type of the environment around the power grid equipment in the preset area, and judging the potential hidden danger scene where the power grid equipment may be located. When a mudslide occurs in the preset area, the pole tower equipment in the power grid will have a pole tower tilt problem. By judging the environmental abnormalities in the preset area, data support is provided for estimating the existence of hidden dangers and abnormalities in the power grid equipment.

[0021] As a preferred solution of the image video scene recognition method based on distance prediction information described in the present invention, the constructed environmental abnormal scene recognition model includes updating according to the newly acquired inspection image video data, outputting the abnormal recognition type judgment result after each inspection image video data is acquired, classifying the inspection image video data according to the judgment result, and then extracting the key frame image, adding it to the data set of the corresponding abnormal recognition model, and re-training the abnormal recognition model according to the updated data set.

[0022] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, steps of an image video scene recognition method based on distance prediction information are implemented.

[0023] A computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of an image or video scene recognition method based on distance prediction information are implemented.

[0024] Beneficial effects of the present invention: The present invention firstly marks and divides the distribution area of ​​the power grid equipment through satellite remote sensing images, so as to determine the distribution location information of the power grid equipment in the preset area, and then inspects the preset area through drones, without distinguishing the power grid equipment, but directly inspects the entire preset area, and then divides the entire inspection image video data according to the distribution area of ​​the power grid equipment to distinguish the inspection image video data of the power grid equipment and the inspection image video data that does not include the power grid equipment, and then matches the pre-constructed abnormal recognition model according to the corresponding inspection image video data to realize the recognition of the hidden danger abnormal scene of the power grid equipment and the abnormal environment scene, so as to realize the inspection of the hidden danger abnormality of the power grid equipment. The present invention can realize automatic inspection, and at the same time, there is no need to distinguish the power grid equipment or to inspect each power grid equipment separately after locating the power grid equipment during the inspection, so the inspection is more convenient and more efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0026] Figure 1 A schematic flow chart of an image or video scene recognition method based on distance prediction information is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0028] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0030] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0031] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0032] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0033] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides an image video scene recognition method based on distance prediction information, comprising:

[0034] Step S1, determine a satellite image of a power grid in a preset area, and mark the satellite image of the power grid according to the distribution location of the power grid in the preset area, so as to obtain the distribution area of ​​each power grid device in the preset area.

[0035] The types of power grid equipment include tower equipment, substations, and transmission lines. Generally speaking, there are many towers in an area, so the power grid equipment in the preset area includes several tower equipment, several substations, several transmission lines, etc.

[0036] The distribution area of ​​each power grid device can be obtained by identifying the distribution location of the power grid devices in the power grid satellite image map. Specifically, the distribution area of ​​each power grid device can be obtained by marking a square area, and the distribution area includes the location information of the power grid device, such as longitude and latitude information.

[0037] The distribution area can be determined by the coordinate values ​​of the four corners of the square area. The size of the square area can be set according to actual needs, and can include the smallest square area of ​​the power grid equipment, or the smallest circular area, etc.

[0038] Step S2: Control the inspection equipment to inspect the preset area to obtain inspection image video data in the preset area. At the same time, the inspection drone will obtain the positioning information of the inspection drone in real time during the inspection process.

[0039] The current inspection of power grid equipment generally requires drones to inspect specific power grid equipment, and then return after each inspection to inspect the next power grid equipment again. At the same time, during the inspection, it is also necessary to obtain the specific location of the power grid equipment. Once an error occurs in the positioning, the inspection route will be offset or the normal inspection cannot be completed. In this embodiment, there is no need to locate the specific location of the power grid equipment in advance. It only needs to give the inspection drone a specific inspection area range, that is, the coordinate range of the above-mentioned preset area. The inspection drone can inspect the preset area according to the established route to obtain the inspection image and video data within the preset area. The inspection image and video data include not only the image and video data of the power grid equipment, but also the surrounding environment without the power grid equipment.

[0040] In addition, the drone also has a preset shooting distance when shooting. For example, when shooting power grid equipment, it is generally shot at a preset shooting distance.

[0041] Step S3: segment the inspection image video data according to the positioning information obtained by the inspection drone and the minimum distribution area of ​​each power grid device in the preset area to generate first inspection image video data and second inspection image video data.

[0042] Normally, when the inspection drone is inspecting, in addition to recording the inspection image and video data, it also obtains positioning information. The inspection image and video data can be segmented according to the acquisition time of the positioning information and the coordinate information of the positioning information to obtain two types of inspection image and video data.

[0043] The first inspection image video data refers to inspection image video data including power grid equipment, and the second inspection image video data refers to inspection image video data other than the first inspection image video data.

[0044] When there are multiple power grid devices, there are also multiple first inspection image video data, and each power grid device corresponds to one first inspection image video data.

[0045] When dividing the image and video data, it is also necessary to divide it based on the preset shooting distance of the drone. Normally, the location of the picture taken by the drone may be a certain distance away from the positioning information of the drone. Therefore, when dividing the image and video data, it is also necessary to consider the preset shooting distance of the drone. For example, when the positioning information of the drone appears outside the area of ​​the power grid equipment, due to the problem of shooting distance, the image and video data it shoots is still in the distribution area of ​​the power grid equipment. Therefore, the image and video data is still divided into the first inspection image and video data.

[0046] At the same time, when dividing, it is also possible to construct images of different shooting angles of drones corresponding to each power grid device in the system and extract corresponding features to construct an image feature set of the power grid device. According to the above image feature set, auxiliary recognition of the power grid equipment in the image video data is realized to achieve the division accuracy of the first inspection image video data and the second inspection image video data.

[0047] In order to ensure the accuracy of the division of inspection image video data, other means can also be used, such as type identification and feature identification of power grid equipment. Specifically, corresponding technical means can be used according to actual needs, and are not limited to the technical means described in this embodiment.

[0048] Step S4: extract key frame images from the first inspection image video data of each power grid device and convert them into multiple inspection images, and match the multiple inspection images with the hidden danger scene recognition model of the power grid equipment constructed in the system to obtain the hidden danger scene of the power grid equipment.

[0049] Among them, the hidden danger scenario recognition model of power grid equipment is obtained by collecting historical inspection images of power grid equipment to construct a data set, and combining it with big data models, such as neural network models and other machine learning models for training.

[0050] At the same time, during model training, the inspection images of power grid equipment collected historically can be preprocessed to distinguish inspection images of power grid equipment in various abnormal states, such as inspection images in rain and snow scenes, inspection images in rainstorm scenes, inspection images in strong wind scenes, etc., and construct a data set, and then divide the data set into a training set and a validation set. The constructed model is trained with the data of the training set, and then the trained model is validated with the data of the validation set to determine whether the trained model meets the requirements. If so, the trained model is recorded as the hidden danger scene recognition model of the power grid equipment.

[0051] At the same time, the normal inspection images of power grid equipment collected in history are also recorded, and the feature vectors of the normal inspection images are collected to construct a set of feature vectors of the inspection images of the power grid equipment in a normal state. This set is used to assist in the identification of the hidden danger scenes of the power grid equipment. Specifically, for example, when each inspection image in the power grid equipment has a high degree of similarity with the inspection image in the system, the power grid equipment is very likely to be in a normal scene. That is, by constructing a hidden danger scene recognition model for the power grid equipment and a set of feature vectors of the inspection images of the power grid equipment in a normal state, the hidden danger scenes of the power grid equipment are identified together, thereby improving the accuracy of hidden danger identification.

[0052] Step S5: extract key frame images from the acquired second inspection image video data to match them with the abnormal environment scene recognition model constructed in the system to determine the type of abnormal environment scene in the preset area.

[0053] The types of abnormal environment scenes include abnormal fire scenes, abnormal debris flow scenes, abnormal vegetation overgrowth scenes, etc. The types of abnormal environment scenes here can be set according to actual needs, and this embodiment is just an example.

[0054] By timely determining the scene type of the environment around the power grid equipment in the preset area, the potential hidden danger scene of the power grid equipment can be determined. For example, when a mudslide occurs in the preset area, the poles and towers in the power grid are likely to tilt. That is, by determining the abnormal environment in the preset area, data support can be provided for estimating the hidden dangers of the power grid equipment.

[0055] Among them, the abnormal environment scene recognition model can set multiple abnormal scene recognition models according to different types of abnormal environment scenes. At the same time, when constructing the abnormal environment scene recognition model, the geographical environment information of the region also needs to be considered. Different geographical environment information also has different scene types. Therefore, when the second inspection image video data is used for environmental abnormal scene judgment, the longitude and latitude of the corresponding area in the second inspection image video data and the GIS geographical information can be first obtained to obtain the corresponding geographical environment information, and then different abnormal environment scene recognition models are obtained from the system according to different geographical environment information to realize the recognition of abnormal environment scenes.

[0056] In addition, when constructing the corresponding model, it is also updated according to the newly acquired inspection image video data. That is, after each inspection image video data is acquired, after the abnormal recognition type judgment result is performed, the inspection image video data is classified according to the judgment result, and then the key frame image is extracted and added to the data set of the corresponding abnormal recognition model to update the data set, and the abnormal recognition model is retrained according to the updated data set to improve the accuracy of the judgment result of the corresponding model through updating.

[0057] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0058] Embodiment 2, the second embodiment of the present invention, is different from the previous embodiment in that:

[0059] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0060] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0061] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0062] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0063] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0064] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for image and video scene recognition based on distance prediction information, characterized in that: include, Determine the satellite image of the power grid in the preset area, and mark it on the satellite image of the power grid according to the distribution location of the power grid in the preset area, so as to obtain the distribution area of ​​each power grid equipment in the preset area; By controlling the inspection equipment to inspect the preset area, the inspection image and video data in the preset area can be obtained. At the same time, during the inspection process, the inspection drone will obtain the positioning information of the inspection drone in real time; The inspection image video data is segmented according to the positioning information obtained by the inspection drone and the minimum distribution area of ​​each power grid device in the preset area to generate first inspection image video data and second inspection image video data; The first inspection image video data of each power grid device is converted into multiple inspection images after key frame image extraction, and the multiple inspection images are matched with the hidden danger scene recognition model of the power grid device constructed in the system to obtain the hidden danger scene of the power grid device; The acquired second inspection image video data is subjected to key frame image extraction to match it with the abnormal environment scene recognition model constructed in the system to determine the type of abnormal environment scene in the preset area.

2. The image and video scene recognition method based on distance prediction information as claimed in claim 1, characterized in that: The marking in the power grid satellite image map according to the distribution position of the power grid in the preset area includes identifying the distribution area of ​​each power grid device in the power grid satellite image map according to the distribution position of the power grid equipment in the power grid, and obtaining the distribution area of ​​each power grid device by marking the square area, and the distribution area includes the positioning information of the power grid equipment.

3. The image and video scene recognition method based on distance prediction information as claimed in claim 2, characterized in that: The distribution area includes, determined by the coordinate values ​​of the four corners of the square area, wherein the square area is set according to actual needs, including the minimum square area of ​​the power grid equipment, or the minimum circular area.

4. The image and video scene recognition method based on distance prediction information as claimed in claim 3, characterized in that: The acquisition of positioning information of the inspection drone includes the need to give the inspection drone a specific inspection area range, that is, the coordinate range of a preset area. The inspection drone inspects the preset area according to a predetermined route to obtain inspection image and video data within the preset area. The inspection image and video data include image and video data of power grid equipment and the surrounding environment without power grid equipment. At the same time, when shooting, the drone is also set with a corresponding preset shooting distance.

5. The method for image and video scene recognition based on distance prediction information as claimed in claim 4, characterized in that: Generating the first inspection image video data and the second inspection image video data includes, when the inspection drone is inspecting, recording the inspection image video data, obtaining positioning information, and segmenting the inspection image video data according to the acquisition time of the positioning information and the coordinate information of the positioning information to obtain two types of inspection image video data; The first inspection image video data is inspection image video data including power grid equipment, and the second inspection image video data refers to inspection image video data other than the first inspection image video data; When there are more than two power grid devices, each power grid device corresponds to one first inspection image video data while passing the first inspection image video data; When dividing the image and video data, it is necessary to divide it based on the preset shooting distance of the drone. When dividing the image and video data, it is necessary to consider the preset shooting distance of the drone. When the positioning information of the drone appears outside the area of ​​the power grid equipment, the image and video data taken are still in the distribution area of ​​the power grid equipment. Therefore, the image and video data are still divided into the first inspection image and video data; At the same time, during the division, images of drones at different shooting angles corresponding to each power grid device are constructed in the system, and the corresponding features are extracted to construct an image feature set of the power grid equipment. Based on the image feature set, auxiliary identification of the power grid equipment in the image video data is achieved.

6. The image and video scene recognition method based on distance prediction information as claimed in claim 5, characterized in that: The key frame image extraction of the acquired second inspection image video data includes: in the extraction scene, the types of abnormal environmental scenes include abnormal fire scenes, abnormal debris flow scenes, and abnormal scenes with overly high vegetation.

7. The method for image and video scene recognition based on distance prediction information as claimed in claim 6, characterized in that: The method of determining the type of abnormal environmental scenario in the preset area includes timely determining the scenario type of the environment surrounding the power grid equipment in the preset area, and determining the potential hidden danger scenario in which the power grid equipment may be located. When a mudslide occurs in the preset area, the tower equipment in the power grid may have a tower tilt problem. By determining the abnormal environment in the preset area, data support is provided for estimating the potential hidden dangers of the power grid equipment.

8. The method for image and video scene recognition based on distance prediction information as claimed in claim 7, characterized in that: The constructed environmental abnormal scene recognition model includes updating according to the newly acquired inspection image video data. After each inspection image video data is acquired, the abnormal recognition type judgment result is output, and the inspection image video data is classified according to the judgment result, and then the key frame image is extracted and added to the data set of the corresponding abnormal recognition model, and the abnormal recognition model is re-trained according to the updated data set.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

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 according to any one of claims 1 to 8 are implemented.

Images