A power line fault early warning method and system based on drone measurement

Abstract

The present invention discloses a power line fault early warning method and system based on drone measurement, which relates to the technical field of power line inspection, including: obtaining at least one power line block; obtaining at least one line block classification; forming at least one early warning feature by classifying the line blocks; obtaining at least one infrared image of an obstruction existing in the power line and an ordinary image of the obstruction; establishing an infrared image superposition model; obtaining at least one actual obstruction existing in the actual ordinary image; obtaining an infrared preprocessing image; obtaining all early warning features appearing in the infrared preprocessing image as target early warning features; and issuing an early warning based on the number of target early warning features. By setting a classification module, a feature acquisition module, a model building module, and a data processing module, it is ensured that there will be no omissions or misjudgments in identification, and all early warning information can be correctly identified, thereby improving the accuracy of early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of power line inspection, and in particular to a power line fault early warning method and system based on unmanned aerial vehicle measurement. Background Art

[0002] The current operation and maintenance mode and inspection method of high-voltage transmission lines in the power grid mainly relies on maintenance personnel relying on ground transportation or walking, using handheld instruments or the naked eye to inspect facilities and deal with defects. This can no longer meet the development and safe operation needs of modern power grids. Ultra-high and extra-high voltage power grids are in urgent need of safe, advanced, scientific and efficient power inspection methods.

[0003] In recent years, power grid companies have gradually adopted the method of using drones as a platform and equipped with sensor terminals to conduct aerial inspections of transmission line equipment and corridor environments, which can partially replace manual line inspections and significantly reduce the labor intensity of personnel. At present, drones are usually equipped with some or all equipment and sensors including stable platforms, positioning and attitude systems, laser scanners, infrared thermal imagers, ultraviolet imagers, video recorders, visible light cameras, etc. to carry out inspections, which greatly improves inspection efficiency and reduces manual inspection costs.

[0004] When existing drones monitor power lines, they are prone to encounter situations where the power lines are blocked. Blocked power lines will interfere with infrared images, which in turn will affect the accuracy of judgment and easily lead to misjudgment of early warnings. Summary of the invention

[0005] In order to solve the above technical problems, a power line fault warning method and system based on drone measurement are provided. The technical solution solves the problem that when the existing drones mentioned in the above background technology are performing power line monitoring, the power lines are easily blocked, and the blocked power lines will interfere with the infrared image, which will affect the accuracy of the judgment and easily lead to misjudgment of the warning.

[0006] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0007] A power line fault early warning method based on drone measurement, comprising:

[0008] Dividing the power line into regions to obtain at least one power line block;

[0009] Acquire sample infrared images of power line blocks without faults or obstructions;

[0010] Classifying at least one power line block to obtain at least one line block classification;

[0011] Classifying the line blocks to form at least one early warning feature, where the early warning feature is an infrared image acquired under unobstructed conditions;

[0012] Acquire at least one infrared image and a common image of an obstruction existing in the power line, and make one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction;

[0013] Based on the infrared image and warning features of the obstruction, an infrared image superposition model is established;

[0014] The drone monitors the power lines in real time and obtains the actual infrared images and actual ordinary images of the power line blocks;

[0015] Based on the actual ordinary image, obtaining at least one actual occluder existing in the actual ordinary image;

[0016] Based on at least one actual obstruction and the infrared image superposition model, removing the obstruction from the actual infrared image to obtain an infrared preprocessed image;

[0017] Obtain all warning features appearing in the infrared pre-processed image as target warning features;

[0018] When the number of target warning features is 0, there is no fault that needs to be warned in the power line block. When the number of target warning features is not 0, it is necessary to warn the fault corresponding to the target warning feature in the power line block.

[0019] Preferably, said dividing the power line into regions to obtain at least one power line block comprises the following steps:

[0020] Based on the composition structure of the power line, the power line is divided, and the different composition structures in the power line are divided into different power line blocks.

[0021] Preferably, the step of classifying at least one power line block to obtain at least one line block classification comprises the following steps:

[0022] Based on the functions of the power line blocks, the power line blocks are classified to obtain at least one line block classification, wherein the power line blocks in the same line block classification realize the same function.

[0023] Preferably, the step of classifying the lines into blocks to form at least one early warning feature comprises the following steps:

[0024] Based on the historical data, obtaining at least one fault feature occurring in the power line block in the line block classification;

[0025] The early warning feature is obtained before the fault feature appears in the power line block, and the interval between the time of obtaining the early warning feature and the time of the fault feature appearing in the power line block is a preset time.

[0026] Preferably, the step of establishing the infrared image superposition model comprises the following steps:

[0027] Obtain the value range of pixel values ​​of pixel points in the infrared image of the obstruction and the warning feature as the infrared value range;

[0028] Divide the infrared value range at equal intervals to obtain at least one infrared point;

[0029] Combining and pairing at least one infrared spot to obtain at least one infrared spot group;

[0030] Obtain the pixel value of the infrared superposition point after the infrared points in the infrared point group are superimposed;

[0031] The infrared point groups are paired and summarized with the infrared overlay points to obtain the infrared image overlay model.

[0032] Preferably, obtaining at least one actual obstruction existing in the actual ordinary image based on the actual ordinary image comprises the following steps:

[0033] Contour recognition is performed in an actual ordinary image to obtain at least one contour. When contour recognition is performed, a difference value between adjacent pixel points in the actual ordinary image is calculated. When the difference value between adjacent pixel points is greater than a preset pixel value, one of the adjacent pixel points is used as a contour point, and the contour points are aggregated to form a contour.

[0034] Classifying the pixels within the contour to obtain at least one contour partition;

[0035] Average the pixel points in the contour partition to obtain the pixel value of the contour partition;

[0036] Calculate the proportion of the contour partition within the contour as the contour partition ratio;

[0037] Classifying the pixels in the ordinary image of the occluder to obtain at least one occluder partition;

[0038] Average the pixel points in the occluder partition to obtain the pixel value of the occluder partition;

[0039] Calculate the proportion of the occluder partition within the outline as the occluder partition ratio;

[0040] In at least one contour partition within the contour, matching the contour partition having the smallest difference between the contour partition pixel value and the occluder partition pixel value of the occluder partition to the occluder partition;

[0041] Determine whether the differences between the contour partition ratios of the contour partitions within the contour and the occlusion partition ratios of the corresponding occlusion partitions are both smaller than a first preset value; if so, take the occlusion corresponding to the occlusion ordinary image as the actual occlusion, and take the contour as the position where the actual occlusion appears; if not, do nothing.

[0042] Preferably, removing obstructions from the actual infrared image to obtain the infrared preprocessed image comprises the following steps:

[0043] Obtaining an infrared image of an actual obstruction existing in an actual ordinary image;

[0044] Obtain the position of the contour of the actual occluder in the actual ordinary image as the feature position;

[0045] The position in the actual infrared image that is the same as the feature position in the actual ordinary image is used as the target position;

[0046] Pair the target position with the pixel points at the same position of the corresponding occluder in the infrared image of the occluder;

[0047] Acquire an infrared superposition point whose pixel value difference from the pixel point at the target position is less than a second preset value as the target infrared superposition point;

[0048] Obtain the infrared point group corresponding to the target infrared superposition point as the target infrared point group;

[0049] Acquire an infrared point in the target infrared point group, the infrared point of the infrared image of the obstruction corresponding to the pixel point at the target position having a pixel difference greater than a second preset value, as the target infrared point;

[0050] The target infrared point is used to replace the pixel point at the target position to obtain the infrared preprocessed image.

[0051] Preferably, the step of acquiring all warning features appearing in the infrared preprocessed image as target warning features comprises the following steps:

[0052] Acquire at least one warning feature corresponding to the line block classification of the power line block corresponding to the infrared preprocessed image as a target warning feature;

[0053] uniformly setting at least one sampling point in the infrared preprocessed image;

[0054] Obtaining the number of at least one target warning feature as a benchmark number;

[0055] Combining at least one target warning feature to form a target warning feature group, wherein the number of target warning features included in the target warning feature group is any number from 0 to a reference number;

[0056] Acquire a sample infrared image of the power line block corresponding to the infrared preprocessing image as a target sample infrared image;

[0057] Randomly selecting a position from at least one sampling point to arrange the warning features in the warning feature group, the warning features in the warning feature group and the infrared image of the target sample form an infrared combined image;

[0058] When the difference between the infrared combined image and the infrared pre-processed image is smaller than a third preset value, the warning feature in the warning feature group corresponding to the infrared combined image is used as the target warning feature.

[0059] A power line fault early warning system based on drone measurement, used to implement the above-mentioned power line fault early warning method based on drone measurement, comprising:

[0060] A region division module, wherein the region division module divides the power line into regions to obtain at least one power line block;

[0061] A sample acquisition module, wherein the sample acquisition module acquires a sample infrared image of a power line block under the condition that there is no fault and no obstruction;

[0062] A classification module, wherein the classification module classifies at least one power line block to obtain at least one line block classification;

[0063] A feature acquisition module, wherein the feature acquisition module forms at least one early warning feature by classifying the line blocks, and the early warning feature is an infrared image acquired under unobstructed conditions;

[0064] An image acquisition module, wherein the image acquisition module acquires at least one infrared image and a common image of an obstruction existing in the power line, and makes a one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction;

[0065] A model building module, wherein the model building module builds an infrared image superposition model based on the infrared image of the obstruction and the warning feature;

[0066] A monitoring module, wherein the monitoring module drone performs real-time monitoring of the power line and obtains actual infrared images and actual ordinary images of the power line blocks;

[0067] A data processing module, wherein the data processing module obtains at least one actual occluder existing in the actual ordinary image based on the actual ordinary image, removes the occluder from the actual infrared image based on the at least one actual occluder and the infrared image superposition model to obtain an infrared preprocessed image, and obtains all warning features appearing in the infrared preprocessed image as target warning features;

[0068] A judgment module is used to perform early warning judgment of faults in power line blocks.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] By setting up a classification module, a feature acquisition module, a model building module and a data processing module, it is possible to perform occlusion recognition on the acquired image, and with the help of the established infrared image superposition model, remove the occlusion in the image, thereby obtaining an unobstructed infrared pre-processed image, and by retrieving all warning features in the line block classification of the power line block corresponding to the infrared pre-processed image, it is ensured that there will be no omissions or misjudgments in the recognition, and then all the warning information can be correctly identified, thereby improving the accuracy of the warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of the power line fault early warning method based on drone measurement of the present invention;

[0072] Figure 2 A schematic diagram of a process of forming at least one early warning feature by classifying line blocks according to the present invention;

[0073] Figure 3 A schematic diagram of the process of establishing an infrared image superposition model of the present invention;

[0074] Figure 4 A schematic diagram of a process of obtaining at least one actual obstruction existing in an actual ordinary image based on an actual ordinary image according to the present invention;

[0075] Figure 5 It is a schematic diagram of the process of removing obstructions from an actual infrared image to obtain an infrared preprocessed image according to the present invention;

[0076] Figure 6 The present invention is a flow chart of obtaining all warning features appearing in an infrared preprocessed image as target warning features. DETAILED DESCRIPTION

[0077] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0078] Reference Figure 1 As shown, a power line fault early warning method based on drone measurement includes:

[0079] Dividing the power line into regions to obtain at least one power line block;

[0080] Acquire sample infrared images of power line blocks without faults or obstructions;

[0081] Classifying at least one power line block to obtain at least one line block classification;

[0082] Classifying the line blocks to form at least one early warning feature, where the early warning feature is an infrared image acquired under unobstructed conditions;

[0083] Acquire at least one infrared image and a common image of an obstruction existing in the power line, and make one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction;

[0084] Based on the infrared image and warning features of the obstruction, an infrared image superposition model is established;

[0085] The drone monitors the power lines in real time and obtains the actual infrared images and actual ordinary images of the power line blocks;

[0086] Based on the actual ordinary image, obtaining at least one actual occluder existing in the actual ordinary image;

[0087] Based on at least one actual obstruction and the infrared image superposition model, removing the obstruction from the actual infrared image to obtain an infrared preprocessed image;

[0088] Obtain all warning features appearing in the infrared pre-processed image as target warning features;

[0089] When the number of target warning features is 0, there is no fault that needs to be warned in the power line block. When the number of target warning features is not 0, it is necessary to warn the fault corresponding to the target warning feature in the power line block.

[0090] Dividing the power line into regions to obtain at least one power line block comprises the following steps:

[0091] Based on the composition structure of the power line, the power line is divided, and the different composition structures in the power line are divided into different power line blocks.

[0092] The faults of power lines are different due to their different structures. Classification can enable targeted fault identification and improve the accuracy of fault identification. At the same time, there is no need to search for faults that will not occur in the structure, thereby saving search time.

[0093] Classifying at least one power line block to obtain at least one line block classification includes the following steps:

[0094] Based on the functions of the power line blocks, the power line blocks are classified to obtain at least one line block classification, wherein the power line blocks in the same line block classification realize the same function.

[0095] By classifying the power line blocks, the early warning features of the power line blocks in the same classification can be shared, because the power line blocks in the same line block classification implement the same functions, and therefore their faults are also similar. Therefore, at least one early warning feature formed by the line block classification may appear in the power line blocks of the line block classification.

[0096] Reference Figure 2 As shown, the process of classifying the line blocks to form at least one early warning feature includes the following steps:

[0097] Based on the historical data, obtaining at least one fault feature occurring in the power line block in the line block classification;

[0098] The early warning feature is obtained before the fault feature appears in the power line block, and the interval between the time of obtaining the early warning feature and the time of the fault feature appearing in the power line block is a preset time.

[0099] There will inevitably be signs before fault characteristics appear in the power line block. In order to provide early warning, the time for obtaining the warning characteristics is set to a preset time before the fault characteristics occur. As a result, maintenance can be carried out within the preset time to avoid the fault affecting the operation of the power line.

[0100] Reference Figure 3 As shown, establishing the infrared image superposition model includes the following steps:

[0101] Obtain the value range of pixel values ​​of pixel points in the infrared image of the obstruction and the warning feature as the infrared value range;

[0102] Divide the infrared value range at equal intervals to obtain at least one infrared point;

[0103] Combining and pairing at least one infrared spot to obtain at least one infrared spot group;

[0104] Obtain the pixel value of the infrared superposition point after the infrared points in the infrared point group are superimposed;

[0105] The infrared point groups are paired and summarized with the infrared overlay points to obtain the infrared image overlay model.

[0106] The purpose of establishing an infrared image superposition model is to obtain the superposition effect of two pixels under infrared conditions. Therefore, according to the superposition effect and the identification of the occlusion, the occlusion can be removed to obtain an image without occlusion, thereby removing interference.

[0107] Reference Figure 4 As shown, based on the actual ordinary image, obtaining at least one actual occluder existing in the actual ordinary image includes the following steps:

[0108] Contour recognition is performed in an actual ordinary image to obtain at least one contour. When contour recognition is performed, a difference value between adjacent pixel points in the actual ordinary image is calculated. When the difference value between adjacent pixel points is greater than a preset pixel value, one of the adjacent pixel points is used as a contour point, and the contour points are aggregated to form a contour.

[0109] Classifying the pixels within the contour to obtain at least one contour partition;

[0110] Average the pixel points in the contour partition to obtain the pixel value of the contour partition;

[0111] Calculate the proportion of the contour partition within the contour as the contour partition ratio;

[0112] Classifying the pixels in the ordinary image of the occluder to obtain at least one occluder partition;

[0113] Average the pixel points in the occluder partition to obtain the pixel value of the occluder partition;

[0114] Calculate the proportion of the occluder partition within the outline as the occluder partition ratio;

[0115] In at least one contour partition within the contour, matching the contour partition having the smallest difference between the contour partition pixel value and the occluder partition pixel value of the occluder partition to the occluder partition;

[0116] Determine whether the differences between the contour partition ratios of the contour partitions within the contour and the occlusion partition ratios of the corresponding occlusion partitions are both smaller than a first preset value; if so, take the occlusion corresponding to the occlusion ordinary image as the actual occlusion, and take the contour as the position where the actual occlusion appears; if not, do nothing.

[0117] Here, at least one actual occluder in the actual ordinary image is identified. Since the occluder may cause changes in the shooting angle, direct image comparison cannot be performed. However, the change in angle of view will not affect the proportion of each part in the image, while the proportion of each part in different objects will be different. Therefore, the actual occluder can be identified based on this.

[0118] Reference Figure 5 As shown, removing the obstruction from the actual infrared image to obtain the infrared preprocessed image includes the following steps:

[0119] Obtaining an infrared image of an actual obstruction existing in an actual ordinary image;

[0120] Obtain the position of the contour of the actual occluder in the actual ordinary image as the feature position;

[0121] The position in the actual infrared image that is the same as the feature position in the actual ordinary image is used as the target position;

[0122] Pair the target position with the pixel points at the same position of the corresponding occluder in the infrared image of the occluder;

[0123] Acquire an infrared superposition point whose pixel value difference from the pixel point at the target position is less than a second preset value as the target infrared superposition point;

[0124] Obtain the infrared point group corresponding to the target infrared superposition point as the target infrared point group;

[0125] Acquire an infrared point in the target infrared point group, the infrared point of the infrared image of the obstruction corresponding to the pixel point at the target position having a pixel difference greater than a second preset value, as the target infrared point;

[0126] The target infrared point is used to replace the pixel point at the target position to obtain the infrared preprocessed image.

[0127] The superposition of two infrared points in the target infrared point group constitutes a pixel point at the target position. The two infrared points in the target infrared point group are respectively a pixel point without an obstruction and a pixel point with an obstruction. The difference between the pixel point of the obstruction and the pixel point of the target position corresponding to the pixel point of the infrared image of the obstruction is less than a second preset value. Therefore, the point where the difference between the pixel point of the infrared image of the obstruction corresponding to the pixel point of the target position is greater than the second preset value is a pixel point without an obstruction. Therefore, using this point for replacement can obtain an infrared preprocessed image.

[0128] Reference Figure 6 As shown, obtaining all the warning features appearing in the infrared preprocessing image as target warning features includes the following steps:

[0129] Acquire at least one warning feature corresponding to the line block classification of the power line block corresponding to the infrared preprocessed image as a target warning feature;

[0130] uniformly setting at least one sampling point in the infrared preprocessed image;

[0131] Obtaining the number of at least one target warning feature as a benchmark number;

[0132] Combining at least one target warning feature to form a target warning feature group, wherein the number of target warning features included in the target warning feature group is any number from 0 to a reference number;

[0133] Acquire a sample infrared image of the power line block corresponding to the infrared preprocessing image as a target sample infrared image;

[0134] Randomly selecting a position from at least one sampling point to arrange the warning features in the warning feature group, the warning features in the warning feature group and the infrared image of the target sample form an infrared combined image;

[0135] When the difference between the infrared combined image and the infrared pre-processed image is smaller than a third preset value, the warning feature in the warning feature group corresponding to the infrared combined image is used as the target warning feature.

[0136] Since the warning features appearing in the infrared preprocessed image are unknown, all possibilities need to be exhaustively enumerated and searched and identified, thereby obtaining an infrared combined image that is similar to the infrared preprocessed image, wherein the degree of similarity is determined by a third preset value. Since the formation process of the infrared combined image is known, the warning features contained therein are known, and therefore, at least one target warning feature of the infrared preprocessed image can be obtained.

[0137] The difference between the infrared combined image and the infrared pre-processed image is the sum of the absolute values ​​of the pixel value differences of the pixel points at corresponding positions in the infrared combined image and the infrared pre-processed image.

[0138] A power line fault early warning system based on drone measurement, used to implement the above-mentioned power line fault early warning method based on drone measurement, comprising:

[0139] A region division module, wherein the region division module divides the power line into regions to obtain at least one power line block;

[0140] A sample acquisition module, wherein the sample acquisition module acquires a sample infrared image of a power line block under the condition that there is no fault and no obstruction;

[0141] A classification module, wherein the classification module classifies at least one power line block to obtain at least one line block classification;

[0142] A feature acquisition module, wherein the feature acquisition module forms at least one early warning feature by classifying the line blocks, and the early warning feature is an infrared image acquired under unobstructed conditions;

[0143] An image acquisition module, wherein the image acquisition module acquires at least one infrared image and a common image of an obstruction existing in the power line, and makes a one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction;

[0144] A model building module, wherein the model building module builds an infrared image superposition model based on the infrared image of the obstruction and the warning feature;

[0145] A monitoring module, wherein the monitoring module drone performs real-time monitoring of the power line and obtains actual infrared images and actual ordinary images of the power line blocks;

[0146] A data processing module, wherein the data processing module obtains at least one actual occluder existing in the actual ordinary image based on the actual ordinary image, removes the occluder from the actual infrared image based on the at least one actual occluder and the infrared image superposition model to obtain an infrared preprocessed image, and obtains all warning features appearing in the infrared preprocessed image as target warning features;

[0147] A judgment module is used to perform early warning judgment of faults in power line blocks.

[0148] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is called, the above-mentioned power line fault early warning method based on drone measurement is executed.

[0149] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0150] To sum up, the advantages of the present invention are: by setting a classification module, a feature acquisition module, a model building module and a data processing module, it is possible to perform occlusion identification on the acquired image, and with the help of the established infrared image superposition model, the occlusion in the image can be removed, thereby obtaining an infrared preprocessed image without occlusion, and by retrieving all warning features in the line block classification of the power line block corresponding to the infrared preprocessed image, it is ensured that there will be no omissions or misjudgments in the identification, and then all warning information can be correctly identified, thereby improving the accuracy of the warning.

[0151] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A power line fault early warning method based on drone measurement, characterized in that: include: Dividing the power line into regions to obtain at least one power line block; Acquire sample infrared images of power line blocks without faults or obstructions; Classifying at least one power line block to obtain at least one line block classification; Classifying the line blocks to form at least one early warning feature, where the early warning feature is an infrared image acquired under unobstructed conditions; Acquire at least one infrared image and a common image of an obstruction existing in the power line, and make one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction; Based on the infrared image and warning features of the obstruction, an infrared image superposition model is established; The drone monitors the power lines in real time and obtains the actual infrared images and actual ordinary images of the power line blocks; Based on the actual ordinary image, obtaining at least one actual occluder existing in the actual ordinary image; Based on at least one actual obstruction and the infrared image superposition model, removing the obstruction from the actual infrared image to obtain an infrared preprocessed image; Obtain all warning features appearing in the infrared pre-processed image as target warning features; When the number of target warning features is 0, there is no fault that needs warning in the power line block. When the number of target warning features is not 0, it is necessary to warn the fault corresponding to the target warning feature in the power line block. The infrared image superposition model is established, comprising the following steps: Obtain the value range of pixel values ​​of pixel points in the infrared image of the obstruction and the warning feature as the infrared value range; Divide the infrared value range at equal intervals to obtain at least one infrared point; Combining and pairing at least one infrared spot to obtain at least one infrared spot group; Obtain the pixel value of the infrared superposition point after the infrared points in the infrared point group are superimposed; Pair and summarize the infrared point group with the infrared overlay points to obtain an infrared image overlay model; The process of removing obstructions from the actual infrared image to obtain the infrared preprocessed image comprises the following steps: Obtaining an infrared image of an actual obstruction existing in an actual ordinary image; Obtain the position of the contour of the actual occluder in the actual ordinary image as the feature position; The position in the actual infrared image that is the same as the feature position in the actual ordinary image is used as the target position; Pair the target position with the pixel points at the same position of the corresponding occluder in the infrared image of the occluder; Acquire an infrared superposition point whose pixel value difference from the pixel point at the target position is less than a second preset value as the target infrared superposition point; Obtain the infrared point group corresponding to the target infrared superposition point as the target infrared point group; Acquire an infrared point in the target infrared point group, the infrared point of the infrared image of the obstruction corresponding to the pixel point at the target position having a pixel difference greater than a second preset value, as the target infrared point; The target infrared point is used to replace the pixel point at the target position to obtain the infrared preprocessed image.

2. The power line fault early warning method based on drone measurement according to claim 1 is characterized in that: The step of dividing the power line into regions to obtain at least one power line block comprises the following steps: Based on the composition structure of the power line, the power line is divided, and the different composition structures in the power line are divided into different power line blocks.

3. The power line fault early warning method based on drone measurement according to claim 2 is characterized in that: The classifying at least one power line block to obtain at least one line block classification comprises the following steps: Based on the functions of the power line blocks, the power line blocks are classified to obtain at least one line block classification, wherein the power line blocks in the same line block classification realize the same function.

4. The power line fault early warning method based on drone measurement according to claim 3 is characterized in that: The step of classifying the line blocks to form at least one early warning feature comprises the following steps: Based on the historical data, obtaining at least one fault feature occurring in the power line block in the line block classification; The early warning feature is obtained before the fault feature appears in the power line block, and the interval between the time of obtaining the early warning feature and the time of the fault feature appearing in the power line block is a preset time.

5. The power line fault early warning method based on drone measurement according to claim 4 is characterized in that: The obtaining, based on the actual ordinary image, at least one actual occluder existing in the actual ordinary image comprises the following steps: Contour recognition is performed in an actual ordinary image to obtain at least one contour. When contour recognition is performed, a difference value between adjacent pixel points in the actual ordinary image is calculated. When the difference value between adjacent pixel points is greater than a preset pixel value, one of the adjacent pixel points is used as a contour point, and the contour points are aggregated to form a contour. Classifying the pixels within the contour to obtain at least one contour partition; Average the pixel points in the contour partition to obtain the pixel value of the contour partition; Calculate the proportion of the contour partition within the contour as the contour partition ratio; Classifying the pixels in the ordinary image of the occluder to obtain at least one occluder partition; Average the pixel points in the occluder partition to obtain the pixel value of the occluder partition; Calculate the proportion of the occluder partition within the outline as the occluder partition ratio; In at least one contour partition within the contour, matching the contour partition having the smallest difference between the contour partition pixel value and the occluder partition pixel value of the occluder partition to the occluder partition; Determine whether the differences between the contour partition ratios of the contour partitions within the contour and the occlusion partition ratios of the corresponding occlusion partitions are both smaller than a first preset value; if so, take the occlusion corresponding to the occlusion ordinary image as the actual occlusion, and take the contour as the position where the actual occlusion appears; if not, do nothing.

6. The power line fault early warning method based on drone measurement according to claim 5 is characterized in that: The method of obtaining all the warning features appearing in the infrared pre-processed image as target warning features comprises the following steps: Acquire at least one warning feature corresponding to the line block classification of the power line block corresponding to the infrared preprocessed image as a target warning feature; uniformly setting at least one sampling point in the infrared preprocessed image; Obtaining the number of at least one target warning feature as a benchmark number; Combining at least one target warning feature to form a target warning feature group, wherein the number of target warning features included in the target warning feature group is any number from 0 to a reference number; Acquire a sample infrared image of the power line block corresponding to the infrared preprocessing image as a target sample infrared image; Randomly selecting a position from at least one sampling point to arrange the warning features in the warning feature group, the warning features in the warning feature group and the infrared image of the target sample form an infrared combined image; When the difference between the infrared combined image and the infrared pre-processed image is smaller than a third preset value, the warning feature in the warning feature group corresponding to the infrared combined image is used as the target warning feature.

7. A power line fault early warning system based on drone measurement, used to implement the power line fault early warning method based on drone measurement as claimed in any one of claims 1 to 6, characterized in that: include: A region division module, wherein the region division module divides the power line into regions to obtain at least one power line block; A sample acquisition module, wherein the sample acquisition module acquires a sample infrared image of a power line block under the condition that there is no fault and no obstruction; A classification module, wherein the classification module classifies at least one power line block to obtain at least one line block classification; A feature acquisition module, wherein the feature acquisition module forms at least one early warning feature by classifying the line blocks, and the early warning feature is an infrared image acquired under unobstructed conditions; An image acquisition module, wherein the image acquisition module acquires at least one infrared image and a common image of an obstruction existing in the power line, and makes a one-to-one correspondence between the infrared image and the common image of the obstruction corresponding to the same obstruction; A model building module, wherein the model building module builds an infrared image superposition model based on the infrared image of the obstruction and the warning feature; A monitoring module, wherein the monitoring module drone performs real-time monitoring of the power line and obtains actual infrared images and actual ordinary images of the power line blocks; A data processing module, wherein the data processing module obtains at least one actual occluder existing in the actual ordinary image based on the actual ordinary image, removes the occluder from the actual infrared image based on the at least one actual occluder and the infrared image superposition model to obtain an infrared preprocessed image, and obtains all warning features appearing in the infrared preprocessed image as target warning features; A judgment module is used to perform early warning judgment of faults in power line blocks.

Images