Equipment inspection method and system based on unmanned aerial vehicle

Abstract

The invention provides an equipment inspection method and system based on an unmanned aerial vehicle, and the method comprises the steps: carrying out the inspection through the unmanned aerial vehicle according to a preset path, so as to obtain the equipment data of all equipment in the preset path; preprocessing and screening the equipment data of each piece of equipment, and taking the screened equipment data as equipment label data; importing the equipment label data into a fault analysis model to output a fault prediction curve of each piece of equipment; wherein the fault analysis model is obtained by training on the basis of the historical equipment data of each piece of equipment. Through the equipment inspection method based on the unmanned aerial vehicle provided by the invention, the fault detection efficiency can be improved.

Description

Technical Field

[0001] The present invention relates to the field of detection technology, and in particular to an equipment inspection method and system based on an unmanned aerial vehicle. Background Art

[0002] In the process of petrochemical engineering construction, it is necessary to ensure the technical performance of the equipment. If the equipment is not inspected regularly and in a timely manner, it may affect the normal operation of the equipment due to equipment aging, equipment performance degradation, equipment hidden dangers and other problems, and even cause major problems such as economic or personal injury. Therefore, it is necessary to conduct regular inspections of the equipment to detect problems and eliminate hidden dangers in a timely manner, thereby extending the service life of the equipment.

[0003] The existing intelligent inspection and management system for petrochemical equipment has high limitations in use and low data collection efficiency. Staff are required to manually adjust the collection path, and the operation process is cumbersome. In addition, the existing intelligent inspection and management system for petrochemical equipment cannot model itself, and the fault detection accuracy is low.

[0004] Providing an intelligent inspection method for petrochemical equipment to improve fault detection efficiency is a technical problem that needs to be solved at present. Summary of the invention

[0005] The present invention provides an equipment inspection method and system based on a drone, so as to solve the defect of low fault detection efficiency in the prior art.

[0006] The present invention provides an equipment inspection method based on a drone, comprising:

[0007] Performing inspections along a preset route using a drone to obtain equipment data of each device set in the preset route;

[0008] Preprocessing and screening the device data of each device, and using the screened device data as device tag data;

[0009] The device label data is imported into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

[0010] According to a drone-based equipment inspection method provided by the present invention, the training process of the fault analysis model includes:

[0011] Extracting multiple sets of historical device data of each device, selecting one set of the historical device data as a test set, and using the remaining historical device data as a training set;

[0012] Importing the training set into a convolutional neural network for learning and training to construct an initial fault analysis model;

[0013] Testing the initial fault analysis model through the test set, and if the test accuracy reaches a preset threshold, using the initial fault analysis model as the fault analysis model;

[0014] If the test accuracy rate does not reach the preset threshold, re-determine a set of historical device data from the multiple sets of historical device data as a test data set, and use the remaining historical device data as a training data set;

[0015] The initial fault analysis model is trained based on the training data set until the test accuracy of the trained initial fault analysis model tested by the test data set reaches the preset threshold, and the trained initial fault analysis model is used as the fault analysis model.

[0016] According to a drone-based equipment inspection method provided by the present invention, the preset path includes multiple flight paths inspected by the drone;

[0017] After the inspection is performed by the drone along the preset path, the method further includes:

[0018] The set of the multiple flight paths is used as a population, and a population matrix is ​​generated in combination with a genetic algorithm; wherein the population matrix includes multiple groups of individuals, and each group of individuals includes multiple flight paths;

[0019] Select any two groups of individuals in the population matrix, and select any flight path from each of the two groups of individuals for exchange, to obtain a population matrix after the first optimization;

[0020] Selecting a group of individuals from the first optimized population matrix, and selecting two flight paths from the group of individuals for exchange, to obtain a second optimized population matrix;

[0021] Select any section of the flight path in the population matrix after the second optimization, and traverse each node of the flight path in order from the end point to the starting point to optimize the flight path until the optimization of each section of the flight path is completed.

[0022] According to a drone-based equipment inspection method provided by the present invention, a flight path is selected in the population matrix after the second optimization, and each node of the flight path is traversed in the order from the end point to the starting point to optimize the flight path, including:

[0023] Selecting any flight path in the population matrix after the second optimization, and traversing each node of the flight path in the order from the end point to the starting point;

[0024] If it is determined that the current node is connected to the starting point of the flight path without any obstacles, then a node between the current node and the starting point is determined to be a redundant node;

[0025] The redundant nodes in each section of the flight path are deleted to optimize each section of the flight path.

[0026] According to a drone-based equipment inspection method provided by the present invention, after outputting the fault prediction curve of each device, the method further includes:

[0027] The faulty device is determined based on the fault prediction curve, and a fault alarm prompt of the faulty device is issued.

[0028] According to a drone-based equipment inspection method provided by the present invention, after acquiring equipment data of each equipment arranged in the preset path, the method further includes:

[0029] The device data of each device is preprocessed into a unified format, and the device data preprocessed into the unified format is stored in blocks.

[0030] The present invention also provides an equipment inspection system based on a drone, comprising:

[0031] The collection and transmission module is used to inspect along a preset path by using a drone to obtain device data of each device set in the preset path;

[0032] A data processing module, used for preprocessing and screening the device data of each device, and using the screened device data as device tag data;

[0033] A fault analysis module is used to import the device label data into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on the historical device data of each device.

[0034] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, an equipment inspection method based on a drone as described above is implemented.

[0035] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described drone-based equipment inspection methods.

[0036] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned drone-based equipment inspection methods.

[0037] The present invention provides a drone-based equipment inspection method and system, which uses a drone to perform inspections along a preset path to obtain equipment data of each device set in the preset path; preprocesses and screens the equipment data of each device, and uses the screened equipment data as equipment label data. The equipment label data is imported into a fault analysis model trained based on historical equipment data of each device to quickly output a fault prediction curve for each device, thereby improving fault detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 This is one of the flow charts of the equipment inspection method based on drone provided by the present invention;

[0040] Figure 2 This is the second flow chart of the equipment inspection method based on drone provided by the present invention;

[0041] Figure 3 This is the third flow chart of the equipment inspection method based on drone provided by the present invention;

[0042] Figure 4 This is the fourth flow chart of the equipment inspection method based on drone provided by the present invention;

[0043] Figure 5 This is one of the structural schematic diagrams of the equipment inspection system based on drone provided by the present invention;

[0044] Figure 6 This is the second structural schematic diagram of the equipment inspection system based on drone provided by the present invention;

[0045] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0047] Combine the following Figure 1-Figure 7 The present invention describes an equipment inspection method and system based on a drone.

[0048] Figure 1 is one of the flow charts of the equipment inspection method based on drone provided in this embodiment, such as Figure 1 As shown, the drone-based equipment inspection method provided in this embodiment includes:

[0049] Step 100: Use a drone to conduct inspections along a preset path to obtain device data of each device set in the preset path.

[0050] It should be noted that the existing inspection of petrochemical equipment usually requires staff to manually adjust the collection path, which has low data collection efficiency and high limitations in the use of the management system; and the existing inspection management system of petrochemical equipment accurately outputs equipment failure problems through self-modeling. Based on this, this embodiment obtains equipment data of each device based on drone inspection.

[0051] Specifically, the drone inspects along a preset path and interacts with each device in the preset path, thereby obtaining device data of each device in the preset path.

[0052] Optionally, when the drone is inspecting along a preset route, it can detect surrounding environmental information through onboard sensors and make real-time adjustments to the flight path based on the environmental information to improve data collection efficiency.

[0053] Step 200: pre-processing and screening the device data of each device, and using the screened device data as device tag data.

[0054] Specifically, after receiving the equipment data of each device obtained by the drone inspection, the equipment data needs to be processed. The specific processing steps may include a preprocessing process and a screening process.

[0055] Preprocessing process: The device data of each device is preprocessed into a unified format, and then the device data of each device is converted into the range of 0 to 1 through the normalization method.

[0056] Based on formula (1), the normalization of device data of each device can be achieved;

[0057]

[0058] Among them, x new represents the normalized data; x represents the extracted feature data; x max Indicates the maximum value of the feature data; x minIndicates the minimum value of the feature data.

[0059] Screening process: Calculate the variance coefficient of the normalized device data of each device. The larger the variance coefficient, the more important the device data is; otherwise, it is unimportant and should be eliminated.

[0060] Based on formula (2), the variance coefficient of the device data of each device can be calculated;

[0061]

[0062] Among them, CV represents the variance coefficient of the feature data; σ represents the standard deviation of the feature data; and μ represents the mean of the feature data.

[0063] The preprocessed and screened device data is used as device tag data, and the fault problem of the device corresponding to the device tag data can be confirmed based on the fault prediction of the device tag data.

[0064] Step 300: Importing the device tag data into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

[0065] Specifically, in this embodiment, the historical equipment data of each device is first trained to obtain a fault analysis model; secondly, the equipment label data of each device is standardized, such as the range standardization method, the Z-score standardization method, etc.; finally, the equipment label data of each device after the standardization is imported into the fault analysis model respectively, and the fault prediction curve of each device is output for the staff to view and analyze.

[0066] The above is a description of the steps of the drone-based equipment inspection method provided by the present invention. From the description of the above steps, it can be seen that according to the drone-based equipment inspection method provided by the present invention, the drone performs inspections along a preset path to obtain the equipment data of each device set in the preset path; the equipment data of each device is preprocessed and screened, and the screened equipment data is used as equipment label data, and the equipment label data is imported into the fault analysis model obtained by training based on the historical equipment data of each device to quickly output the fault prediction curve of each device, thereby improving the fault detection efficiency.

[0067] Based on the above embodiments, in this embodiment, Figure 2 This is the second flow chart of the equipment inspection method based on drone provided in this embodiment. Figure 2 As shown, the training process of the fault analysis model includes:

[0068] Step 210: extract multiple groups of historical device data of each device, select one group of the historical device data as a test set, and use the remaining historical device data as a training set.

[0069] Step 220: Import the training set into a convolutional neural network for learning and training to construct an initial fault analysis model.

[0070] Specifically, multiple groups of historical equipment data of each device are extracted, and one group of historical equipment data is selected as a test set to test the accuracy of the training model. The remaining historical equipment data is used as a training set. The training set is imported into the convolutional neural network and trained through input, convolution, pooling, full connection and output to obtain an initial fault analysis model.

[0071] Step 230: Testing the initial fault analysis model using the test set, and if the test accuracy reaches a preset threshold, using the initial fault analysis model as the fault analysis model.

[0072] Specifically, the initial fault analysis model is tested using the test set, and a threshold for the test accuracy is set in advance. For example, the preset threshold is set to 0.9. If the test accuracy of the current test set is equal to 0.9 or greater than 0.9, the initial fault analysis model is used as the fault analysis model, and the parameters of the initial fault analysis model are used as the parameters of the fault analysis model.

[0073] Step 240: If the test accuracy does not reach the preset threshold, a set of historical device data is re-determined from the multiple sets of historical device data as a test data set, and the remaining historical device data is used as a training data set.

[0074] Step 250: Train the initial fault analysis model based on the training data set until the test accuracy of the trained initial fault analysis model tested by the test data set reaches the preset threshold, and then use the trained initial fault analysis model as the fault analysis model.

[0075] It should be noted that if the test accuracy of the current test set is less than a preset threshold, the initial fault analysis model can continue to be trained to adjust parameters.

[0076] Specifically, a set of historical equipment data is re-determined from multiple sets of historical equipment data as a test data set, and the remaining historical equipment data is used as a training data set. The initial fault analysis model is continuously trained based on the training data set until the test accuracy of the trained initial fault analysis model tested by the test data set reaches a preset threshold. The trained initial fault analysis model is then used as the fault analysis model, and the parameters of the trained initial fault analysis model are used as the parameters of the fault analysis model.

[0077] Optionally, in this embodiment, parameter adjustment can also be performed based on the root mean square error. Specifically, a group of historical device data is determined as a test data set from multiple groups of historical device data, and the remaining historical device data is used as a training data set. The initial fault analysis model is trained using the training data set, and the trained initial fault analysis model is tested using the test data set. After the test is repeated for a preset number of times, the root mean square error and corresponding parameters of the trained initial fault analysis model are recorded. The above steps are repeated until each group of historical device data is tested as a test data set, and the parameters of the initial fault analysis model are replaced with the parameters corresponding to the model with the smallest root mean square error.

[0078] The drone-based equipment inspection method provided in this embodiment obtains a fault analysis model by learning and training the historical equipment data of each device, and ensures the accuracy of fault detection by adjusting the parameters of the fault analysis model.

[0079] Based on the above embodiments, in this embodiment, Figure 3 This is the third flow chart of the equipment inspection method based on drone provided in this embodiment. Figure 3 As shown, the preset path includes multiple flight paths for the drone inspection; after the drone performs the inspection along the preset path in step 100, the method further includes:

[0080] Step 310: taking the set of the multiple flight paths as a population and combining it with a genetic algorithm to generate a population matrix; wherein the population matrix includes multiple groups of individuals, and each group of individuals includes multiple flight paths.

[0081] It should be noted that, when the drone is inspecting along a preset path, the onboard sensors can detect the surrounding environmental information and adjust the flight path in real time based on the environmental information to improve data collection efficiency. Specifically, the preset path of the drone inspection includes multiple flight paths.

[0082] Based on formula (3), the flight path of the UAV is expressed;

[0083]

[0084] in, represents the flight path of the UAV; τ k,n It represents a single flight path represented by the number of n nodes where the UAV takes off for the kth time; k represents the kth takeoff of the UAV after completing the collection of all flight paths.

[0085] Step 320: Select any two groups of individuals in the population matrix, and select any flight path from each of the two groups of individuals for exchange, to obtain the population matrix after the first optimization.

[0086] Step 330: Select a group of individuals from the first optimized population matrix, select two flight paths from the group of individuals for exchange, and obtain a second optimized population matrix.

[0087] Specifically, a collection of multiple flight paths is used as a population, and a population matrix is ​​generated in combination with a genetic algorithm. For example, the population matrix is ​​generated through steps such as encoding, population initialization, calculation of individual fitness, evolutionary calculation, and decoding. This embodiment does not impose any special restrictions on this. Two groups of individuals are randomly selected from the population matrix, and then a flight path is selected from each of the two groups of individuals for exchange, thereby obtaining two new groups of individuals, based on which the population matrix after the first optimization is obtained. Then, a group of individuals is randomly selected, and two flight paths in the individuals are randomly selected for exchange, based on which the population matrix after the second optimization is obtained.

[0088] Step 340: Select any section of the flight path in the population matrix after the second optimization, and traverse each node of the flight path in order from the end point to the starting point to optimize the flight path until the optimization of each section of the flight path is completed.

[0089] Optionally, after the population matrix is ​​optimized as a whole, a specific flight path needs to be optimized separately to generate an optimal flight path. Specifically, a flight path is selected from the population matrix after the second optimization, and each node of the flight path is traversed in the order from the end point to the starting point to optimize the flight path, and each flight path is optimized to achieve a complete optimization process.

[0090] The drone-based equipment inspection method provided in this embodiment generates a population matrix by combining a set of flight paths with a genetic algorithm, and continuously optimizes the population matrix and each flight path so that the drone can obtain equipment data based on the optimized path. This can improve data collection efficiency and does not require staff to manually adjust the collection path, making the operation simple and convenient.

[0091] Based on the above embodiments, in this embodiment, Figure 4 This is a fourth flow chart of the equipment inspection method based on drone provided in this embodiment. Figure 4 As shown, in step 340, a flight path is selected in the population matrix after the second optimization, and each node of the flight path is traversed in the order from the end point to the starting point to optimize the flight path, including:

[0092] Step 410: Select any flight path in the population matrix after the second optimization, and traverse each node of the flight path in the order from the end point to the starting point.

[0093] Step 420: If it is determined that the current node is unimpededly connected to the starting point of the flight path, a node between the current node and the starting point is determined to be a redundant node.

[0094] It should be noted that after the overall optimization of the population matrix, the specific flight path needs to be optimized separately to generate the optimal flight path. Specifically, a flight path is selected in the population matrix after the second optimization, and each node of the flight path is traversed in the order from the end point to the starting point. If it is determined that the current node and the starting point of the flight path can be connected without obstacles, the node between the current node and the starting point is determined as a redundant node. Traverse each node of each flight path to determine all redundant nodes.

[0095] Step 430: Delete the redundant nodes in each section of the flight path to optimize each section of the flight path.

[0096] It should be noted that the fitness function is a function used to measure the adaptability of individuals in the population matrix.

[0097] Specifically, after determining the redundant nodes of each flight path, all redundant nodes are deleted and the fitness function of the path is recalculated. Each flight path is optimized through continuous iteration.

[0098] The drone-based equipment inspection method provided in this embodiment generates a population matrix by combining a set of flight paths with a genetic algorithm, and continuously optimizes the population matrix and each flight path so that the drone can obtain equipment data based on the optimized path. This can improve data collection efficiency and does not require staff to manually adjust the collection path, making the operation simple and convenient.

[0099] Based on the above embodiment, in this embodiment, after outputting the fault prediction curve of each device in step 300, the method further includes:

[0100] The faulty device is determined based on the fault prediction curve, and a fault alarm prompt of the faulty device is issued.

[0101] Specifically, after outputting the fault prediction curves of each device, the faulty device is determined based on the fault prediction curve, and a fault alarm prompt of the faulty device is issued to the staff, so that the staff can quickly locate the position of the faulty device and handle the faulty device in time.

[0102] The drone-based equipment inspection method provided in this embodiment can send out fault alarm prompts for faulty equipment, so that workers can quickly locate the position of the faulty equipment and handle the faulty equipment in a timely manner, thereby reducing safety hazards.

[0103] Based on the above embodiment, in this embodiment, after obtaining the device data of each device set in the preset path in step 100, the method further includes:

[0104] The device data of each device is preprocessed into a unified format, and the device data preprocessed into the unified format is stored in blocks.

[0105] It should be noted that after obtaining the device data of each device, the device data needs to be stored in blocks to meet the needs of data decentralization, while reducing data processing energy consumption and preventing malicious tampering of data.

[0106] Specifically, first, the device data of each device is preprocessed into a unified format and divided into qualified blocks. When joining the network, each node in the blockchain network generates a local public-private key pair as its own identification code in the network. When a node waits for its local role to become a candidate node, it broadcasts a leadership application to other nodes in the network and sends it; secondly, after the application is approved, the candidate node becomes the leader node, and the other nodes become follower nodes. The leader node then broadcasts the block to record the information. After receiving the information, the follower node broadcasts the received information to other follower nodes and records the number of repetitions. The information with the most repetitions is used to generate a block header, and a verification application is sent to the leader node; finally, after the verification is passed, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the dormant period ends. After receiving the confirmation information, the follower node adds each newly generated group of blocks to the blockchain and returns the candidate identity.

[0107] The drone-based equipment inspection method provided in this embodiment stores equipment data in blocks to meet the needs of decentralized data maintenance, while reducing data processing energy consumption and preventing malicious tampering of data.

[0108] The drone-based equipment inspection system provided by the present invention is described below. The drone-based equipment inspection system described below and the drone-based equipment inspection method described above can be referenced to each other.

[0109] Figure 5 is one of the structural schematic diagrams of the equipment inspection system based on drone provided in this embodiment, such as Figure 5 As shown, the drone-based equipment inspection system provided in this embodiment includes:

[0110] The collection and transmission module 501 is used to inspect along a preset path by using a drone to obtain device data of each device set in the preset path;

[0111] The data processing module 502 is used to pre-process and filter the device data of each device, and use the filtered device data as device tag data;

[0112] The fault analysis module 503 is used to import the device label data into the fault analysis model to output the fault prediction curve of each device; wherein the fault analysis model is trained based on the historical device data of each device.

[0113] The drone-based equipment inspection system provided in this embodiment uses the drone to conduct inspections along a preset path to obtain equipment data of each device set in the preset path; the equipment data of each device is preprocessed and screened, and the screened equipment data is used as equipment label data. The equipment label data is imported into a fault analysis model trained based on the historical equipment data of each device to quickly output the fault prediction curve of each device, thereby improving the fault detection efficiency.

[0114] Based on the above embodiments, in this embodiment, Figure 6 This is the second structural diagram of the equipment inspection system based on drone provided in this embodiment. Figure 6 As shown, the system further includes a parameter adjustment module 504, which is specifically used to:

[0115] Extracting multiple sets of historical device data of each device, selecting one set of the historical device data as a test set, and using the remaining historical device data as a training set;

[0116] Importing the training set into a convolutional neural network for learning and training to construct an initial fault analysis model;

[0117] Testing the initial fault analysis model through the test set, and if the test accuracy reaches a preset threshold, using the initial fault analysis model as the fault analysis model;

[0118] If the test accuracy rate does not reach the preset threshold, re-determine a set of historical device data from the multiple sets of historical device data as a test data set, and use the remaining historical device data as a training data set;

[0119] The initial fault analysis model is trained based on the training data set until the test accuracy of the trained initial fault analysis model tested by the test data set reaches the preset threshold, and the trained initial fault analysis model is used as the fault analysis model.

[0120] Based on the above embodiment, in this embodiment, the preset path includes multiple flight paths of the drone inspection;

[0121] The system further includes a path planning module 505, which is specifically used for:

[0122] After the drone performs inspection along the preset path, the set of the multiple flight paths is used as a population, and a population matrix is ​​generated in combination with a genetic algorithm; wherein the population matrix includes multiple groups of individuals, and each group of individuals includes multiple flight paths;

[0123] Select any two groups of individuals in the population matrix, and select any flight path from each of the two groups of individuals for exchange, to obtain a population matrix after the first optimization;

[0124] Selecting a group of individuals from the first optimized population matrix, and selecting two flight paths from the group of individuals for exchange, to obtain a second optimized population matrix;

[0125] Select any section of the flight path in the population matrix after the second optimization, and traverse each node of the flight path in order from the end point to the starting point to optimize the flight path until the optimization of each section of the flight path is completed.

[0126] Based on the above embodiments, in this embodiment, the path planning module 505 is specifically used for:

[0127] Selecting any flight path in the population matrix after the second optimization, and traversing each node of the flight path in the order from the end point to the starting point;

[0128] If it is determined that the current node is connected to the starting point of the flight path without any obstacles, then a node between the current node and the starting point is determined to be a redundant node;

[0129] The redundant nodes in each section of the flight path are deleted to optimize each section of the flight path.

[0130] Based on the above embodiment, in this embodiment, the system further includes an alarm positioning module 506, which is specifically used for:

[0131] After the fault prediction curves of each device are output, the faulty device is determined based on the fault prediction curve, and a fault alarm prompt of the faulty device is issued.

[0132] Based on the above embodiment, in this embodiment, the system further includes a block storage module 507, which is specifically used for:

[0133] After acquiring the device data of each device arranged in the preset path, the device data of each device is preprocessed into a unified format, and the device data preprocessed into the unified format is stored in blocks.

[0134] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the equipment inspection method based on the drone, and the method includes:

[0135] Performing inspections along a preset route using a drone to obtain equipment data of each device set in the preset route;

[0136] Preprocessing and screening the device data of each device, and using the screened device data as device tag data;

[0137] The device label data is imported into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

[0138] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0139] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the equipment inspection method based on a drone provided by the above methods, the method includes:

[0140] Performing inspections along a preset route using a drone to obtain equipment data of each device set in the preset route;

[0141] Preprocessing and screening the device data of each device, and using the screened device data as device tag data;

[0142] The device label data is imported into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

[0143] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the equipment inspection method based on a drone provided by the above methods, the method comprising:

[0144] Performing inspections along a preset route using a drone to obtain equipment data of each device set in the preset route;

[0145] Preprocessing and screening the device data of each device, and using the screened device data as device tag data;

[0146] The device label data is imported into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

[0147] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0148] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for equipment inspection based on drones, It is characterized in that include: Performing inspections along a preset route using a drone to obtain device data of each device set in the preset route; Preprocessing and screening the device data of each device, and using the screened device data as device tag data; The device label data is imported into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on historical device data for each device.

2. The equipment inspection method based on drone according to claim 1, It is characterized in that The training process of the fault analysis model includes: Extracting multiple sets of historical device data of each device, selecting one set of the historical device data as a test set, and using the remaining historical device data as a training set; Importing the training set into a convolutional neural network for learning and training to build an initial fault analysis model; Testing the initial fault analysis model through the test set, and if the test accuracy reaches a preset threshold, using the initial fault analysis model as the fault analysis model; If the test accuracy rate does not reach the preset threshold, re-determine a set of historical device data from the multiple sets of historical device data as a test data set, and use the remaining historical device data as a training data set; The initial fault analysis model is trained based on the training data set until the test accuracy of the trained initial fault analysis model tested by the test data set reaches the preset threshold, and the trained initial fault analysis model is used as the fault analysis model.

3. The equipment inspection method based on drone according to claim 1, It is characterized in that The preset path includes multiple flight paths for the drone inspection; After the inspection is performed by the drone along the preset path, the method further includes: The set of the multiple flight paths is used as a population, and a population matrix is ​​generated in combination with a genetic algorithm; wherein the population matrix includes multiple groups of individuals, and each group of individuals includes multiple flight paths; Select any two groups of individuals in the population matrix, and select any flight path from each of the two groups of individuals for exchange, to obtain a population matrix after the first optimization; Selecting a group of individuals from the first optimized population matrix, and selecting two flight paths from the group of individuals for exchange, to obtain a second optimized population matrix; Select any section of the flight path in the population matrix after the second optimization, and traverse each node of the flight path in order from the end point to the starting point to optimize the flight path until the optimization of each section of the flight path is completed.

4. The equipment inspection method based on drone according to claim 3, It is characterized in that The step of selecting a flight path from the population matrix after the second optimization, and traversing each node of the flight path in the order from the end point to the starting point to optimize the flight path includes: Selecting any flight path in the population matrix after the second optimization, and traversing each node of the flight path in the order from the end point to the starting point; If it is determined that the current node is connected to the starting point of the flight path without any obstacles, then a node between the current node and the starting point is determined to be a redundant node; The redundant nodes in each section of the flight path are deleted to optimize each section of the flight path.

5. The equipment inspection method based on drone according to claim 1, It is characterized in that After outputting the failure prediction curves of each device, the method further includes: The faulty device is determined based on the fault prediction curve, and a fault alarm prompt of the faulty device is issued.

6. The equipment inspection method based on drone according to claim 1, It is characterized in that After acquiring the device data of each device set in the preset path, the method further includes: The device data of each device is preprocessed into a unified format, and the device data preprocessed into the unified format is stored in blocks.

7. An equipment inspection system based on drones, It is characterized in that include: The collection and transmission module is used to inspect along a preset path by using a drone to obtain device data of each device set in the preset path; A data processing module, used for preprocessing and screening the device data of each device, and using the screened device data as device tag data; A fault analysis module is used to import the device label data into a fault analysis model to output a fault prediction curve for each device; wherein the fault analysis model is trained based on the historical device data of each device.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the program, the drone-based equipment inspection method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the drone-based equipment inspection method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, the drone-based equipment inspection method according to any one of claims 1 to 6 is implemented.

Images