Inspection report generation method, device, electronic device and storage medium

Abstract

The present invention discloses a method, device, electronic device and storage medium for generating an inspection report. The method comprises: obtaining inspection image information collected by an inspection robot within a preset time; screening the inspection image information according to a predetermined rule to determine target inspection image information; then inputting the target inspection image information into a historically trained neural network model for processing to determine an image recognition result; calculating corresponding geometric attribute information according to the image recognition result; and finally generating a corresponding inspection report according to the geometric attribute information. Thus, the image inspection data collected by the inspection robot can generate a corresponding inspection data report more accurately, so that the staff can view the problems in various areas of the substation through the inspection report and diagnose and repair them in time.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power inspection, and in particular to an inspection report generation method, device, electronic equipment and storage medium. Background Art

[0002] At present, when traditional manual inspection is adopted in the substation area, the workload is extremely large and repetitive, which can easily cause some inspectors to feel bored and the inspection work is not in place. To a certain extent, there are also a series of problems caused by no one on site to monitor and patrol in a timely manner. Therefore, the inspection robots currently used in the substation area realize various intelligent inspection tasks, but it is difficult to generate corresponding inspection data reports for the image inspection data collected by the inspection robots, which makes it difficult to diagnose problems in various areas of the substation in a timely manner. Summary of the invention

[0003] In a first aspect, the main purpose of the present invention is to provide a method for generating an inspection report, comprising:

[0004] Obtain inspection image information collected by the inspection robot within a preset time;

[0005] Screening the inspection image information according to a predetermined rule to determine target inspection image information;

[0006] Inputting the target inspection image information into a historically trained neural network model for processing to determine an image recognition result;

[0007] Calculating corresponding geometric attribute information according to the image recognition result;

[0008] Generate a corresponding inspection report according to the geometric attribute information.

[0009] Optionally, the obtaining of inspection image information collected by the inspection robot within a preset time includes:

[0010] Obtaining the inspection route of the inspection robot within a preset time;

[0011] According to the inspection route, a plurality of inspection image information corresponding to each inspection position in the inspection route is determined.

[0012] Optionally, the screening the inspection image information according to a predetermined rule to determine to obtain target inspection image information includes:

[0013] Determining the image quality of the plurality of inspection image information according to the plurality of inspection image information corresponding to each inspection position;

[0014] Screening is performed according to the image quality of the plurality of inspection image information to determine the target inspection image information with the best image quality.

[0015] Optionally, the step of inputting the target inspection image information into a historically trained neural network model for processing to determine a recognition result includes:

[0016] Inputting the target inspection image information into a neural network model to determine a global feature image;

[0017] Performing image segmentation on the global feature image to obtain an original local feature image and a deviation local feature image;

[0018] The original local feature image and the deviation local feature image are determined as the image recognition result.

[0019] Optionally, calculating corresponding geometric attribute information according to the image recognition result includes:

[0020] Performing semantic information fusion on the original local feature image and the deviation local feature image respectively to obtain a first local feature image and a second local feature image containing semantic information;

[0021] Calculating pixel points corresponding to the first local feature image and the second local feature image respectively;

[0022] An area calculation is performed based on the pixel points of the first local feature image to obtain first geometric attribute information; and an area calculation is performed based on the pixel points of the second local feature image to obtain second geometric attribute information.

[0023] Optionally, generating a corresponding inspection report according to the geometric attribute information includes:

[0024] Associating the first geometric attribute information and the first local feature image and inputting them into an inspection report template to obtain a first inspection report;

[0025] The second geometric attribute information and the second local feature image are associated and input into an inspection report template to obtain a second inspection report.

[0026] Optionally, after performing area calculation according to the pixel points of the second local feature image to obtain the second geometric attribute information, the method further comprises:

[0027] Determining whether the second geometric attribute information exceeds the preset geometric attribute information;

[0028] When the second geometric attribute information exceeds the preset geometric attribute information, determining that the target inspection image information is abnormal image information;

[0029] When the second geometric attribute information is lower than the preset geometric attribute information, it is determined that the target inspection image information is normal image information.

[0030] In a second aspect, an embodiment of the present invention provides an inspection report generating device, including:

[0031] An acquisition module is used to acquire inspection image information collected by the inspection robot within a preset time;

[0032] A screening module, used to screen the inspection image information according to a predetermined rule to determine the target inspection image information;

[0033] A recognition module, used for inputting the target inspection image information into a historically trained neural network model for processing to determine an image recognition result;

[0034] A calculation module, used to calculate corresponding geometric attribute information according to the image recognition result;

[0035] A determination module is used to generate a corresponding inspection report according to the geometric attribute information.

[0036] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the inspection report generation method as described above when executing the computer program.

[0037] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the inspection report generation method as described above are implemented.

[0038] The above solution of the present invention includes at least the following beneficial effects:

[0039] The inspection report generation method provided by the present invention first obtains the inspection image information collected by the inspection robot within a preset time; and screens the inspection image information according to a predetermined rule to determine the target inspection image information; then inputs the target inspection image information into a historically trained neural network model for processing to determine the image recognition result; calculates the corresponding geometric attribute information according to the image recognition result; and finally generates the corresponding inspection report according to the geometric attribute information. In this way, the image inspection data collected by the inspection robot can generate the corresponding inspection data report more accurately, so that the staff can view the problems in various areas of the substation through the inspection report and diagnose and repair them in time. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0041] Figure 1 A schematic diagram of the overall process of the inspection report generation method provided by an embodiment of the present invention;

[0042] Figure 2 A schematic diagram of a specific flow chart of step S30 provided in an embodiment of the present invention;

[0043] Figure 3 A schematic diagram of a specific flow chart of step S40 provided in an embodiment of the present invention;

[0044] Figure 4 Another schematic diagram of the process of step S40 provided in an embodiment of the present invention;

[0045] Figure 5 A schematic diagram of a specific flow chart of step S50 provided in an embodiment of the present invention;

[0046] Figure 6 A structural block diagram of an inspection report generating device provided by an embodiment of the present invention;

[0047] Figure 7 This is a structural block diagram of an electronic device provided by an embodiment of the present invention.

[0048] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] The terms "first", "second", "third", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the term "comprising" and any variation thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0051] First, the scheme of the embodiment of the present application is introduced by way of example with reference to the relevant drawings.

[0052] like Figure 1 As shown, a specific embodiment of the present invention provides a method for generating an inspection report, comprising:

[0053] S10, obtaining inspection image information collected by the inspection robot within a preset time;

[0054] In this embodiment, the preset time can be 1 day, the inspection robot can be an inspection robot with functions such as autonomous walking, autonomous charging, and autonomous obstacle avoidance, or it can be a mechanical and electronic device with autonomous computing and execution capabilities, or it can be a combination of terminal equipment and mechanical mechanisms with control capabilities; the inspection image information can be in BMP format, jpg format, PNG format, etc., and the inspection image information can include power equipment such as transmission lines, transformers, and switch cabinets; the inspection robot can be controlled by a remote control platform, so that the remote control platform can receive the inspection image information collected by the inspection robot. Optionally, when collecting inspection image information, the inspection robot can patrol once every 1 hour within 1 day, and can collect multiple inspection image information for each location; for example, the regional location of the transformer can capture corresponding multiple inspection image information.

[0055] Specifically, the above-mentioned obtaining of inspection image information collected by the inspection robot within a preset time includes:

[0056] Step 1: Obtain the inspection route of the inspection robot within a preset time;

[0057] Step 2: According to the inspection route, multiple inspection image information corresponding to each inspection position in the inspection route is determined.

[0058] Among them, the inspection route can be a pre-set fixed route, for example, the inspection location can be a robot charging station, a high and low voltage distribution room, a switch cabinet room, a capacitor room, a transformer room, and a robot charging station. Therefore, when the inspection robot is controlled to perform an inspection, the route of the inspection robot can be starting from the robot charging station, and then passing through the high and low voltage distribution room, the switch cabinet room, the capacitor room, and the transformer room in turn, and finally returning to the robot charging station; and, during the inspection process, all transmission lines on the inspection route can be photographed in turn to obtain corresponding transmission line inspection image information, thereby, the collected inspection image information can include multiple inspection image information corresponding to the high and low voltage distribution room, multiple inspection image information corresponding to the switch cabinet room, multiple inspection image information corresponding to the capacitor room, multiple inspection image information corresponding to the transformer room, and multiple inspection image information on the inspection route; optionally, the inspection image information collected at each inspection location can have different image quality, so the inspection image information corresponding to each inspection location can be screened later.

[0059] S20: Screen the inspection image information according to a predetermined rule to determine target inspection image information.

[0060] In the present embodiment, the inspection image information is filtered according to predetermined rules to determine the target inspection image information, including: determining the image quality of the multiple inspection image information according to the multiple inspection image information corresponding to each inspection position; filtering according to the image quality of the multiple inspection image information to determine the target inspection image information with the best image quality.

[0061] Among them, the predetermined rule is expressed as determining the inspection image information with the best image quality as the target inspection image information; therefore, when screening multiple inspection image information of each inspection position, a unique target inspection image information can be determined for each type of power equipment in each inspection position, thereby generating an inspection report for the target inspection image information with the best image quality more accurately; for example, the high and low voltage distribution room includes a switchgear, a distribution room, an incoming line cabinet, a metering cabinet, etc. When the switchgear takes 20 pictures, the switchgear can determine the target inspection image information with the best image quality from the 20 pictures, and when the incoming line cabinet takes 30 pictures, the incoming line cabinet can determine the target inspection image information with the best image quality from the 30 pictures, thereby ensuring that the target inspection image information corresponding to each type of power equipment at each inspection position is more accurate.

[0062] Optionally, each type of electrical equipment at each inspection position may correspond to multiple inspection image information, and the inspection image information of each type of electrical equipment has corresponding image noise, such as strong light, low light, blurred and jittery images. By screening the inspection image information of each type of electrical equipment, the multiple inspection image information corresponding to each inspection position can be determined, and then a corresponding inspection report can be generated for each inspection position; it can be understood that when screening the inspection image information, an image noise threshold can be pre-set. When the noise of the inspection image information is greater than the pre-set image noise threshold, it can be deleted. When the noise of the inspection image information is less than the pre-set image noise threshold, it can be deleted. , it can be retained, that is, the greater the image noise, the lower the image quality of the inspection image information, and the smaller the image noise, the higher the image quality of the inspection image; for example, for the switchgear and incoming line cabinet in the high and low voltage distribution room, when the switchgear takes 50 pictures, the switchgear can determine the image noise threshold that is less than the preset image noise threshold from the 50 pictures as the target inspection image information, and when the incoming line cabinet takes 60 pictures, the incoming line cabinet can determine the image noise threshold that is less than the preset image noise threshold from the 60 pictures as the target inspection image information, thereby ensuring that the target inspection image information corresponding to each power equipment at each inspection position is more accurate.

[0063] S30, inputting the target inspection image information into the historically trained neural network model for processing to determine the image recognition result.

[0064] In this embodiment, the neural network model can be an FCN (Fully Convolutional Networks) network model. The FCN network model is represented by pixel-level classification of the image. The feature map of the last convolution layer is upsampled through the deconvolution layer to restore the feature map to the same size as the input image, so that a prediction can be generated for each pixel, while retaining the spatial information in the original input image, and finally pixel-by-pixel classification is performed on the upsampled feature map; therefore, when training the neural network model, multiple inspection image information obtained in historical time can be input into the neural network model for training. Therefore, when processing the target inspection image information through the neural network model, the corresponding semantic information in the target inspection image information can be determined, and whether there is a problem with the power equipment in the target inspection image information can be determined based on the corresponding semantic information, so that the corresponding data when generating the inspection report is more comprehensive.

[0065] like Figure 2 As shown, the target inspection image information is input into the historically trained neural network model for processing to determine the recognition result, including:

[0066] S31, inputting the target inspection image information into the neural network model to determine the global feature image;

[0067] S32, performing image segmentation on the global feature image to obtain an original local feature image and a deviation local feature image;

[0068] S33: Determine the original local feature image and the deviation local feature image as image recognition results.

[0069] Among them, global features refer to the overall attributes of the image. Common global features include color features, texture features, and shape features, such as intensity histograms, etc. Local features are features extracted from local areas of the image, including edges, corners, lines, curves, and areas with special attributes, etc. Therefore, when processing the target inspection image information, the corresponding global feature image can be determined after feature processing through the neural network model, and the global feature image can be segmented to determine the original local feature image and the deviation local feature image. The original local feature image can be represented as the regional feature image of the power equipment; the deviation local feature image can be represented as the regional feature image of the foreign body. Of course, the original local feature image can also be represented as the regional feature image of the transmission line, and the deviation local feature image can be represented as the regional feature image of the foreign body; for example, in winter, when the transmission line is frozen, the transmission line corresponds to the original local feature image, and the ice cube corresponds to the deviation local feature image. Therefore, it is possible to determine whether the power equipment or the transmission line is abnormal through the target inspection image information, and then generate the corresponding inspection report.

[0070] S40. Calculate corresponding geometric attribute information according to the image recognition result.

[0071] In this embodiment, the image recognition result includes the original local feature image and the deviation local feature image, and the geometric attribute information may include pixel area and pixel direction. The pixel area of ​​the original local feature image can be expressed as the image area of ​​the power equipment, and the pixel direction of the original local feature image can be expressed as the edge direction of the power equipment. The pixel area of ​​the deviation local feature image can be expressed as the image area of ​​the foreign object, and the pixel direction of the deviation local feature image can be expressed as the edge direction of the foreign object. By respectively determining the geometric attribute information of the original local feature image and the geometric attribute information of the deviation local feature image, an inspection report of the power equipment and an inspection report of the corresponding foreign objects on the power equipment can be generated according to the geometric attribute information, so that it can be determined in the inspection report whether the power equipment has abnormalities and other problems.

[0072] like Figure 3 As shown, the corresponding geometric attribute information is calculated based on the image recognition result, including:

[0073] S41, performing semantic information fusion on the original local feature image and the deviation local feature image respectively to obtain a first local feature image and a second local feature image containing semantic information;

[0074] S42, respectively calculating the pixel points corresponding to the first local feature image and the second local feature image;

[0075] S43, performing area calculation according to the pixel points of the first local feature image to obtain first geometric attribute information; and performing area calculation according to the pixel points of the second local feature image to obtain second geometric attribute information.

[0076] Among them, the semantic information can be determined by the neural network model. Therefore, the original local feature image and its corresponding semantic information can be fused through the neural network model, as well as the deviation local feature image and its corresponding semantic information. The semantic information of the original local feature image can be the name corresponding to the power equipment; the semantic information of the deviation local feature image can be the name corresponding to the foreign object; a variety of different semantic information can be determined by the neural network model, and then the corresponding original local feature images and the corresponding deviation local feature images are fused; it can be understood that the first local image and the second local image determined after the fusion both contain corresponding semantic information, and the first geometric attribute information is determined by calculating the pixel area of ​​the first local feature image, and the second geometric attribute information is determined by calculating the pixel area of ​​the second local feature image, and then the inspection report corresponding to the power equipment can be generated by the first geometric attribute information, and the inspection report corresponding to the foreign object can be generated by the second geometric attribute information.

[0077] like Figure 4 As shown, after the area calculation is performed based on the pixel points of the second local feature image to obtain the second geometric attribute information, the following steps are included:

[0078] S45, determining whether the second geometric attribute information exceeds the preset geometric attribute information;

[0079] S46, when the second geometric attribute information exceeds the preset geometric attribute information, determining that the target inspection image information is abnormal image information;

[0080] S47. When the second geometric attribute information is lower than the preset geometric attribute information, determine that the target inspection image information is normal image information.

[0081] Among them, the preset geometric attribute information may include a preset pixel area and a preset pixel direction. By judging whether the pixel area of ​​the foreign object exceeds the preset pixel area and whether the pixel direction of the foreign object exceeds the preset pixel direction, it can be determined whether the foreign object will cause abnormality in the power equipment; therefore, when the second geometric attribute exceeds the preset geometric attribute information, it indicates that the size of the foreign object is relatively large, so it can be determined that the target inspection image information is abnormal image information, such as ice cubes appearing on the transmission line, debris appearing on the power equipment, etc.; when the second geometric attribute does not exceed the preset geometric attribute information, it indicates that the size of the foreign object is relatively small, so it can be determined that the target inspection image information is normal image information, such as leaves appearing on the power equipment, etc.

[0082] S50. Generate a corresponding inspection report according to the geometric attribute information.

[0083] In this embodiment, the inspection report may include an inspection report on power equipment and an inspection report on foreign objects, and the inspection report may be generated in a tabular form. Therefore, the inspection report may include each inspection location, multiple power equipment corresponding to each inspection location, and the inspection results corresponding to each power equipment, such as the result that the power equipment is in a normal state and the result that the power equipment is in an abnormal state. The target image information of the power equipment may also be used as the inspection result. Of course, the first geometric attribute information and the second geometric attribute information may also be used as the inspection result for each power equipment.

[0084] like Figure 5 As shown, the above-mentioned inspection report generated according to the geometric attribute information includes:

[0085] S51, associating the first geometric attribute information and the first local feature image and inputting them into an inspection report template to obtain a first inspection report;

[0086] S52: Associating the second geometric attribute information and the second local feature image and inputting them into an inspection report template to obtain a second inspection report.

[0087] Among them, the first inspection report can be expressed as an inspection report of the power equipment, the second inspection report can be expressed as an inspection report of the foreign object, the inspection report template can be a pre-set inspection report, etc., the first geometric attribute information and the first local feature image are associated to indicate that the pixel area and the pixel direction are marked on the first local feature image, thereby, the first inspection report can be determined in the associated input inspection report template, and the first inspection report may include the size of the power equipment; the second geometric attribute information and the second local feature image are associated to indicate that the pixel area and the pixel direction are marked on the second local feature image, thereby, the second inspection report can be determined in the associated input inspection report template, and the second inspection report may include the size of the foreign object and the direction of the foreign object, and the status of each power equipment can be accurately determined through the first inspection report and the second inspection report.

[0088] The inspection report generation method provided by the present invention first obtains the inspection image information collected by the inspection robot within a preset time; and screens the inspection image information according to a predetermined rule to determine the target inspection image information; then inputs the target inspection image information into a historically trained neural network model for processing to determine the image recognition result; calculates the corresponding geometric attribute information according to the image recognition result; and finally generates the corresponding inspection report according to the geometric attribute information. In this way, the image inspection data collected by the inspection robot can generate the corresponding inspection data report more accurately, so that the staff can view the problems in various areas of the substation through the inspection report and diagnose and repair them in time.

[0089] like Figure 6 As shown, an embodiment of the present invention provides an inspection report generating device 10, comprising:

[0090] The acquisition module 11 is used to acquire the inspection image information collected by the inspection robot within a preset time;

[0091] A screening module 12 is used to screen the inspection image information according to a predetermined rule to determine the target inspection image information;

[0092] Recognition module 13, used for inputting target inspection image information into the historically trained neural network model for processing to determine the image recognition result;

[0093] A calculation module 14, used to calculate corresponding geometric attribute information according to the image recognition result;

[0094] The determination module 15 is used to generate a corresponding inspection report according to the geometric attribute information.

[0095] The inspection report generating device 10 provided by the present invention first obtains the inspection image information collected by the inspection robot within a preset time; and screens the inspection image information according to a predetermined rule to determine the target inspection image information; then inputs the target inspection image information into the historically trained neural network model for processing to determine the image recognition result; calculates the corresponding geometric attribute information according to the image recognition result; and finally generates the corresponding inspection report according to the geometric attribute information. In this way, the image inspection data collected by the inspection robot can generate the corresponding inspection data report more accurately, so that the staff can view the problems in various areas of the substation through the inspection report and diagnose and repair them in time.

[0096] It should be noted that the inspection report generating device 10 provided in the specific embodiment of the present invention is a device corresponding to the above-mentioned inspection report generating method, and all embodiments of the above-mentioned inspection report generating method are applicable to the inspection report generating device 10. The above-mentioned inspection report generating device 10 embodiments all have corresponding modules corresponding to the steps in the above-mentioned inspection report generating method, and can achieve the same or similar beneficial effects. In order to avoid excessive repetition, each module in the inspection report generating device 2 will not be described in detail here.

[0097] like Figure 7 As shown, a specific embodiment of the present invention further provides an electronic device 20, including a memory 202, a processor 201, and a computer program stored in the memory 202 and executable on the processor 201, wherein the processor 201 implements the steps of the above-mentioned inspection report generation method when executing the computer program.

[0098] Specifically, the processor 201 is used to call the computer program stored in the memory 202 to perform the following steps:

[0099] Obtain inspection image information collected by the inspection robot within a preset time;

[0100] Screening the inspection image information according to a predetermined rule to determine the target inspection image information;

[0101] Input the target inspection image information into the historically trained neural network model for processing to determine the image recognition result;

[0102] Calculate the corresponding geometric attribute information according to the image recognition result;

[0103] Generate corresponding inspection reports based on geometric attribute information.

[0104] Optionally, the processor 201 executes the step of obtaining inspection image information collected by the inspection robot within a preset time, including:

[0105] Obtain the inspection route of the inspection robot within the preset time;

[0106] According to the inspection route, multiple inspection image information corresponding to each inspection position in the inspection route is determined.

[0107] Optionally, the processor 201 performs screening of the inspection image information according to a predetermined rule to determine to obtain target inspection image information, including:

[0108] Determine the image quality of the multiple inspection image information according to the multiple inspection image information corresponding to each inspection position;

[0109] Screening is performed based on the image quality of multiple inspection image information to determine the target inspection image information with the best image quality.

[0110] Optionally, the processor 201 inputs the target inspection image information into a historically trained neural network model for processing to determine a recognition result, including:

[0111] Input the target inspection image information into the neural network model to determine the global feature image;

[0112] Perform image segmentation on the global feature image to obtain an original local feature image and a deviation local feature image;

[0113] The original local feature image and the deviation local feature image are determined as image recognition results.

[0114] Optionally, the processor 201 calculates corresponding geometric attribute information according to the image recognition result, including:

[0115] Performing semantic information fusion on the original local feature image and the deviation local feature image respectively to obtain a first local feature image and a second local feature image containing semantic information;

[0116] Calculate the pixel points corresponding to the first local feature image and the second local feature image respectively;

[0117] An area calculation is performed based on the pixel points of the first local feature image to obtain the first geometric attribute information; and an area calculation is performed based on the pixel points of the second local feature image to obtain the second geometric attribute information.

[0118] Optionally, the processor 201 generates a corresponding inspection report according to the geometric attribute information, including:

[0119] Associating the first geometric attribute information and the first local feature image and inputting them into an inspection report template to obtain a first inspection report;

[0120] The second geometric attribute information and the second local feature image are associated and input into the inspection report template to obtain a second inspection report.

[0121] Optionally, after the processor 201 performs area calculation according to the pixels of the second local feature image to obtain the second geometric attribute information, the following steps may be performed:

[0122] Determining whether the second geometric attribute information exceeds the preset geometric attribute information;

[0123] When the second geometric attribute information exceeds the preset geometric attribute information, determining that the target inspection image information is abnormal image information;

[0124] When the second geometric attribute information is lower than the preset geometric attribute information, it is determined that the target inspection image information is normal image information.

[0125] That is, in a specific embodiment of the present invention, the processor 201 of the electronic device 20 implements the steps of the above-mentioned inspection report generation method when executing the computer program, so that the image inspection data collected by the inspection robot can more accurately generate corresponding inspection data reports, so that the staff can view the problems in various areas of the substation through the inspection report, and diagnose and repair them in time.

[0126] It should be noted that since the processor 201 of the electronic device 20 implements the steps of the above-mentioned inspection report generation method when executing a computer program, all embodiments of the above-mentioned inspection report generation method are applicable to the electronic device 20 and can achieve the same or similar beneficial effects.

[0127] The computer-readable storage medium provided in the embodiment of the present invention stores a computer program. When the computer program is executed by the processor, it implements the various processes of the inspection report generation method or the application-side inspection report generation method provided in the embodiment of the present invention, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0128] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0129] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0130] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for generating an inspection report, characterized in that: include: Obtain inspection image information collected by the inspection robot within a preset time; Screening the inspection image information according to a predetermined rule to determine target inspection image information; Inputting the target inspection image information into the historically trained neural network model for processing to determine the image recognition result, including: inputting the target inspection image information into the neural network model to determine the global feature image; performing image segmentation on the global feature image to obtain an original local feature image and a deviation local feature image; determining the original local feature image and the deviation local feature image as the image recognition result; the original local feature is represented as a power equipment area feature image, and the deviation local feature is represented as a foreign body area feature image; Calculating the corresponding geometric attribute information according to the image recognition result, including: fusing the original local feature image with its corresponding semantic information, and fusing the deviation local feature image with its corresponding semantic information to obtain a first local feature image and a second local feature image containing semantic information; respectively calculating the pixel points corresponding to the first local feature image and the second local feature image; performing area calculation according to the pixel points of the first local feature image to obtain the first geometric attribute information; and performing area calculation according to the pixel points of the second local feature image to obtain the second geometric attribute information; Generate a corresponding inspection report according to the geometric attribute information.

2. The inspection report generation method according to claim 1, characterized in that: The obtaining of inspection image information collected by the inspection robot within a preset time includes: Obtaining the inspection route of the inspection robot within a preset time; According to the inspection route, a plurality of inspection image information corresponding to each inspection position in the inspection route is determined.

3. The inspection report generation method according to claim 2, characterized in that: The screening of the inspection image information according to a predetermined rule to determine to obtain target inspection image information includes: Determining the image quality of the plurality of inspection image information according to the plurality of inspection image information corresponding to each inspection position; Screening is performed based on the image quality of the plurality of inspection image information to determine target inspection image information with the best image quality.

4. The inspection report generation method according to claim 1, characterized in that: The generating a corresponding inspection report according to the geometric attribute information includes: Associating the first geometric attribute information and the first local feature image and inputting them into an inspection report template to obtain a first inspection report; The second geometric attribute information and the second local feature image are associated and input into an inspection report template to obtain a second inspection report.

5. The inspection report generation method according to claim 1, characterized in that: After performing area calculation according to the pixel points of the second local feature image to obtain the second geometric attribute information, the method further comprises: Determining whether the second geometric attribute information exceeds the preset geometric attribute information; When the second geometric attribute information exceeds the preset geometric attribute information, determining that the target inspection image information is abnormal image information; When the second geometric attribute information is lower than the preset geometric attribute information, it is determined that the target inspection image information is normal image information.

6. A patrol inspection report generating device, characterized in that: include: An acquisition module is used to acquire inspection image information collected by the inspection robot within a preset time; A screening module, used to screen the inspection image information according to a predetermined rule to determine the target inspection image information; The recognition module is used to input the target inspection image information into the historically trained neural network model for processing to determine the image recognition result; including: inputting the target inspection image information into the neural network model to determine the global feature image; performing image segmentation on the global feature image to obtain an original local feature image and a deviation local feature image; determining the original local feature image and the deviation local feature image as the image recognition result; the original local feature is represented as a power equipment area feature image, and the deviation local feature is represented as a foreign body area feature image; A calculation module, used to calculate the corresponding geometric attribute information according to the image recognition result, including: fusing the original local feature image with its corresponding semantic information, and fusing the deviation local feature image with its corresponding semantic information to obtain a first local feature image and a second local feature image containing semantic information; respectively calculating the pixel points corresponding to the first local feature image and the second local feature image; performing area calculation according to the pixel points of the first local feature image to obtain the first geometric attribute information; and performing area calculation according to the pixel points of the second local feature image to obtain the second geometric attribute information; A determination module is used to generate a corresponding inspection report according to the geometric attribute information.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the inspection report generating method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the inspection report generating method according to any one of claims 1 to 5 are implemented.

Images