Aircraft inspection method and equipment and storage medium

Abstract

The invention provides an aircraft inspection method and device and a storage medium. The method comprises the following steps: dynamically adjusting the inspection demand of each to-be-inspected aircraft according to the inspection auxiliary information of each to-be-inspected aircraft, wherein the inspection auxiliary information comprises historical flight data, historical inspection data and flight environment data. And according to the inspection requirements of the to-be-inspected aircrafts, obtaining inspection images of the to-be-inspected aircrafts in the aircraft inspection process. And inputting the inspection images of the to-be-inspected aircrafts into an inspection abnormality inspection model, and generating inspection reports of the to-be-inspected aircrafts, the inspection reports being used for indicating whether the surfaces of the to-be-inspected aircrafts are abnormal or not. If the inspection report comprises abnormal information, a processing decision corresponding to the abnormal information is obtained according to the abnormal information, and the abnormal information is used for indicating that the surface of the to-be-inspected aircraft is abnormal. According to the method, the accuracy and efficiency of aircraft inspection are improved.

Description

Technical Field

[0001] The present application relates to communication technology, and more particularly to an aircraft inspection method, equipment and storage medium. Background Art

[0002] With the rapid development of the aviation industry, aircraft safety and operational efficiency have become key considerations in the industry. As an important part of ensuring flight safety, aircraft inspections require accuracy and efficiency. Traditional aircraft inspection methods mainly rely on manual work, where inspectors need to visually inspect the surface of the aircraft. This method is not only time-consuming and labor-intensive, but also easily affected by human factors, and may result in missed inspections or false inspections.

[0003] Therefore, how to improve the accuracy and efficiency of aircraft inspections is an urgent problem to be solved. Summary of the invention

[0004] The present application provides an aircraft inspection method, device and storage medium to solve the problem of low accuracy and efficiency of aircraft inspection.

[0005] In a first aspect, the present application provides an aircraft inspection method, comprising:

[0006] Dynamically adjusting the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected, wherein the inspection auxiliary information includes historical flight data, historical inspection data, and flight environment data;

[0007] According to the inspection requirements of each of the aircraft to be inspected, obtaining inspection images of each of the aircraft to be inspected during the aircraft inspection process;

[0008] Inputting the inspection images of the aircraft to be inspected into the inspection anomaly detection model to generate an inspection report for each aircraft to be inspected, wherein the inspection report is used to indicate whether there is an abnormality on the aircraft surface of each aircraft to be inspected;

[0009] If the inspection report includes abnormal information, a processing decision corresponding to the abnormal information is obtained according to the abnormal information, and the abnormal information is used to indicate that an abnormality exists on the aircraft surface of the aircraft to be inspected.

[0010] Optionally, dynamically adjusting the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected includes:

[0011] Acquire the inspection auxiliary information and an inspection demand analysis network, wherein the inspection demand analysis network is used to determine the inspection demand of each of the aircraft to be inspected according to the inspection auxiliary information;

[0012] Performing abnormal trend analysis on the inspection auxiliary information according to the inspection demand analysis network, and determining a set of high-risk monitoring points of the aircraft to be inspected, and an inspection priority of the high-risk monitoring points in the set of high-risk monitoring points based on the historical flight data and / or the historical inspection data and / or the flight environment data;

[0013] The inspection requirements of each of the aircraft to be inspected are adjusted according to the high-risk monitoring point set and the inspection priorities of the high-risk monitoring points in the high-risk monitoring point set.

[0014] Optionally, performing abnormal trend analysis on the inspection auxiliary information according to the inspection demand analysis network, determining a set of high-risk monitoring points of the aircraft to be inspected based on the historical flight data and / or historical inspection data and / or flight environment data, and an inspection priority of high-risk monitoring points in the set of high-risk monitoring points, includes:

[0015] The inspection demand analysis network determines the abnormal score calculation data of each aircraft component feature node according to the inspection auxiliary information and the mapping relationship between the inspection auxiliary information and the aircraft component feature node;

[0016] Performing a correlation analysis on the abnormal score calculation data and the abnormal state of the aircraft component through a correlation analysis network to obtain the degree of deviation between the current state of the characteristic node of the aircraft component and the normal state, as well as the abnormal trend;

[0017] Performing fusion calculation on the degree of deviation between the current state of the aircraft component feature node and the normal state, as well as the abnormal trend, to generate an abnormal score for the aircraft component feature node;

[0018] According to the abnormality score of the aircraft component feature node and the abnormality score threshold, the aircraft component feature nodes with a score higher than the abnormality score threshold are used as a set of high-risk monitoring points of the aircraft to be inspected;

[0019] The inspection priority of the high-risk monitoring points in the high-risk monitoring point set is determined according to the abnormality score of each of the high-risk monitoring points in the high-risk monitoring point set.

[0020] Optionally, obtaining inspection images of each of the aircraft to be inspected during the aircraft inspection according to the inspection requirements of each of the aircraft to be inspected includes:

[0021] According to the inspection requirements of each of the aircraft to be inspected, determining the inspection image parameters required to be collected for the characteristic nodes of different aircraft components of each of the aircraft to be inspected, wherein the inspection image parameters include the number of the inspection images, and / or the angle of the inspection image, and / or the clarity of the inspection image;

[0022] Extracting acquisition features of the inspection image parameters through an intelligent image capture model, identifying and classifying the image features, and generating image acquisition parameters for acquiring inspection images of each of the aircraft to be inspected;

[0023] The inspection image of each of the aircraft to be inspected is collected according to the image collection parameters.

[0024] Optionally, the anomaly inspection model includes a visual recognition model and a deep learning model, and the inspection images of the aircraft to be inspected are input into the inspection anomaly inspection model to generate an inspection report for each aircraft to be inspected, including:

[0025] Extracting image features from the inspection image using a visual recognition model, wherein the image features include the shape of an aircraft component, the texture of an aircraft component, and the color of an aircraft component of the aircraft to be inspected;

[0026] Mapping the image features in the inspection image to the high-dimensional space of the deep learning model through the trained deep learning model to generate a feature vector corresponding to the image features;

[0027] Determine the abnormal features of the feature nodes of each aircraft component of the aircraft to be inspected according to the similarity calculation result between the feature vector corresponding to the image feature and the standard feature vector, and mark the abnormal state and / or abnormal position corresponding to the abnormal feature;

[0028] An inspection report for each of the aircraft to be inspected is generated according to the abnormal features of the feature nodes of each of the aircraft components of each of the aircraft to be inspected.

[0029] Optionally, obtaining a processing decision corresponding to the exception information according to the exception information includes:

[0030] Obtaining a first knowledge graph corresponding to the abnormal information and a second knowledge graph corresponding to the processing decision;

[0031] Determine, according to the association path of the abnormal information in the first knowledge graph, associated abnormal information of the abnormal information, where the associated abnormal information is abnormal information corresponding to a first abnormal entity that is abnormally associated with the abnormal information;

[0032] Determine at least one processing decision entity corresponding to the first abnormal entity and the second abnormal entity in the second knowledge graph according to the second abnormal entity corresponding to the abnormal information, the first abnormal entity, and the association between the visualized data matrix of the first knowledge graph and the visualized data matrix of the second knowledge graph;

[0033] Determine the processing decision corresponding to the abnormal information according to the processing decision corresponding to the processing decision entity;

[0034] The step of obtaining the first knowledge graph corresponding to the abnormal information and the second knowledge graph corresponding to the processing decision includes:

[0035] Pre-establishing an aviation field knowledge base, which contains knowledge data corresponding to aircraft components, common faults, maintenance procedures, and safety standards. The aviation field knowledge base is generated by organizing the aircraft manufacturer's technical documents, professional books in the aviation field, historical inspection records, and maintenance reports;

[0036] Using a natural language processing algorithm, perform entity recognition and relationship extraction on the abnormal information, determine key entity nouns in the abnormal information and the relationship between the key entity nouns, and generate a first processing result;

[0037] Based on the first processing result, construct a first knowledge graph, in which each node represents a key entity noun and each edge represents a relationship between the key entity nouns;

[0038] and, obtaining decision data related to the processing decision, the decision data being derived from maintenance manuals, best practice guidelines, and historical processing cases;

[0039] Using a natural language processing algorithm to perform entity recognition and relationship extraction on the decision data to generate a second processing result;

[0040] Based on the second processing result, construct a second knowledge graph, in which each node represents a processing strategy, a maintenance tool or a spare part, and an edge represents a logical relationship or a dependency relationship between nodes;

[0041] After generating the first knowledge graph and the second knowledge graph, establishing an association between the first knowledge graph and the second knowledge graph, specifically by adding cross-knowledge graph edges in the first knowledge graph and the second knowledge graph;

[0042] The step of determining the associated abnormal information of the abnormal information according to the associated path of the abnormal information in the first knowledge graph includes:

[0043] By matching the keywords in the abnormal information, locating the target node corresponding to the abnormal information in the first knowledge graph, and taking the target node as the starting point, starting from the starting point node, exploring potential association paths along the edges, wherein in the first knowledge graph, the nodes are connected by the edges, indicating the association relationship between the nodes;

[0044] During the exploration process, each node and edge passed through is recorded, as well as the degree of association between each node and edge and the starting node, and the degree of association is represented by a weight;

[0045] After exploring all potential associated paths, analyzing and recording all candidate nodes directly or indirectly connected to the starting point node, the candidate nodes representing other abnormal information associated with the abnormal information;

[0046] Sorting and filtering the associated abnormal information according to the association degree and the path length of the potential associated path, and generating filtered associated abnormal information;

[0047] The step of locating a target node corresponding to the abnormal information in the first knowledge graph by matching keywords in the abnormal information includes:

[0048] Extracting keywords from the abnormal information, wherein the keywords at least include aircraft component name, fault type, and abnormal phenomenon;

[0049] After extracting the keywords, preprocessing the keywords, the preprocessing includes removing stop words and stemming;

[0050] The word embedding algorithm is used to map the preprocessed keywords into a high-dimensional vector space to generate keyword vectors;

[0051] Traversing all nodes in the first knowledge graph by using an approximate nearest neighbor search algorithm, and calculating the cosine similarity between each node and the keyword vector;

[0052] If the cosine similarity between any node and the keyword vector is greater than a set threshold, the node is determined as a target node corresponding to the abnormal information;

[0053] And, if there are multiple nodes whose cosine similarities are all greater than a set threshold, sort them according to the size of the cosine similarities, and select the node with the highest similarity as the target node;

[0054] The step of taking the target node as the starting point and exploring potential associated paths along the edges starting from the starting point includes:

[0055] Initializing the exploration environment, specifically including loading the complete graph data of the first knowledge graph and setting corresponding exploration parameters, wherein the exploration parameters include the maximum exploration depth and the step length of each exploration;

[0056] Adopting a breadth-first search algorithm, starting from the target node, exploring along the edges connected to it, recording each node and edge passed through, as well as the degree of association between each node and edge and the starting node, and in the exploration process, screening association paths that meet the preset mining value based on set evaluation criteria, wherein the evaluation criteria include evaluation criteria based on node attributes, edge types or weights;

[0057] Visualize the explored association paths.

[0058] Optionally, determining the processing decision corresponding to the abnormal information according to the processing decision corresponding to the processing decision entity includes:

[0059] Obtaining maintenance history information, spare parts inventory status, and maintenance support information of the aircraft to be inspected;

[0060] adjusting the processing decision according to the maintenance history information, the spare parts inventory status, the maintenance support information, and the processing decision corresponding to the processing decision entity;

[0061] The adjusted processing decision is determined as the processing decision corresponding to the abnormal information.

[0062] Optionally, the method further includes:

[0063] If there is abnormal feedback information for the processing decision corresponding to the abnormal information, and a call request sent by the inspection terminal is received, then according to the call request and the state of the maintenance terminal, an idle maintenance terminal is called, and the call request is used to request a call with the maintenance terminal for abnormal feedback;

[0064] If the idle maintenance terminal passes the call request, a call connection is established between the inspection terminal and the idle maintenance terminal;

[0065] If the idle locomotive terminal does not respond to the call request, other idle locomotive terminals are called until the call connection between the inspection terminal and the other idle locomotive terminals is established.

[0066] In a second aspect, the present application provides an electronic device, comprising: a processor, a communication interface and a memory, wherein the processor is communicatively connected to the communication interface and the memory respectively;

[0067] The memory stores computer-executable instructions;

[0068] The communication interface communicates and interacts with external devices;

[0069] The processor executes the computer-executable instructions stored in the memory to implement the method as described in any one of the first aspects.

[0070] In a third aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the aircraft inspection method as described in any one of the first aspects.

[0071] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the aircraft inspection method described in any one of the first aspects.

[0072] The aircraft inspection method, device and storage medium provided in the present application dynamically adjust the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected, obtain the inspection images of each aircraft to be inspected during the aircraft inspection process according to the inspection requirements of each aircraft to be inspected, input the inspection images of each aircraft to be inspected into the inspection abnormality detection model, generate the inspection report of each aircraft to be inspected, and if the inspection report includes abnormal information, obtain the processing decision corresponding to the abnormal information according to the abnormal information. The method of the present application can effectively reduce the manual participation in the aircraft inspection process, reduce the manual workload, and improve the efficiency of the aircraft inspection process. In addition, through the image recognition method, the abnormal situation existing in the inspection image of the aircraft is automatically identified, and the corresponding processing decision is given for the abnormal situation, which further improves the accuracy of aircraft inspection and abnormal handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0074] Figure 1 A schematic diagram of a flow chart of an aircraft inspection method provided in an embodiment of the present application;

[0075] Figure 2 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0076] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0077] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0078] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below in conjunction with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0079] Figure 1 The following is a flow chart of an aircraft inspection method provided in an embodiment of the present application. Figure 1 As shown, the method may include:

[0080] S101 . Dynamically adjust the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected.

[0081] Among them, the inspection auxiliary information refers to various types of information that provide decision support for the server when conducting aircraft inspections, including but not limited to historical flight data, historical inspection data and flight environment data. The historical flight data contains historical records of the aircraft's flight time, route, flight altitude, speed, etc. The historical flight data can reflect the damage to the fuselage during the aircraft's historical flights. For example, when flying over a specific route, specific parts of the aircraft are prone to damage. In a possible case, it can also reflect the predicted cloud damage level of the aircraft. Historical inspection data refers to past inspection records, including inspection time, inspection results, problems found, etc. Combined with historical inspection data, the high-risk parts of the aircraft can be determined to determine a more targeted inspection plan for the high-position parts. Flight environment data includes information about the external environment in which the aircraft is flying, such as weather conditions, air quality, etc. In a possible case, the flight environment can also reflect the damage to the fuselage.

[0082] The server can flexibly modify the inspection plan based on the information collected in real time to meet the inspection needs of different aircraft and different situations. Specifically, it collects and integrates historical flight data, historical inspection data, and flight environment data. It analyzes these data to identify possible problem areas and risk points. Based on the analysis results, it adjusts the inspection focus, frequency, and method of each aircraft to be inspected.

[0083] For example, if the server receives data of five aircraft to be inspected, by analyzing historical flight data, it is found that two of the aircraft have recently flown for a long time and often fly in bad weather. At the same time, historical inspection data shows that the wings of these two aircraft have minor damage records. Based on this information, the server can increase the inspection frequency and accuracy of the wings of these two aircraft.

[0084] S102: Acquire inspection images of each of the aircraft to be inspected during an aircraft inspection process according to the inspection requirements of each of the aircraft to be inspected.

[0085] The inspection image refers to a high-definition image of each part of the aircraft taken during the inspection process, which is used for subsequent analysis and judgment. The inspection image can be an inspection image taken using a shooting position, angle and resolution determined according to inspection requirements. The inspection image can be taken using a high-definition camera or other image acquisition equipment.

[0086] In one possible implementation, inspection requirements may include standardized inspections and free inspections. Standard inspections refer to inspections of a certain type of aircraft with multiple designated inspection areas, and mandatory photography of certain key inspection areas. After each inspection of an area is completed, the inspection image of the area is uploaded and the area is marked as inspected. The inspection can only be completed when all areas have been inspected. Free inspections refer to inspections that are not restricted to specific inspection areas and photography requirements. Inspectors can determine the inspection process and whether to take photos by themselves.

[0087] Under standard inspection, the inspection image can be taken manually by the inspector, or the server can determine the shooting position, angle and resolution of the inspection image according to the inspection requirements, and send the shooting position, angle and resolution of the inspection image to the drone and other equipment, and the image acquisition device carried by these equipment will automatically take the image. After the inspection image is taken, the inspection image can be uploaded to the server as part of the inspection record for maintenance personnel to view, or for subsequent quality inspection and spot check.

[0088] Optionally, when the inspection image is taken manually by the inspection personnel, if there is a serious problem in an area outside the area covered by the inspection requirements, the inspection personnel can additionally take inspection images of the area and upload them to the server as emergency inspection images. The uploaded emergency inspection images can be displayed in the maintenance monitoring list with a higher priority, so that the emergency inspection images can obtain processing opinions with a higher priority.

[0089] S103: Input the inspection images of the aircraft to be inspected into an inspection anomaly detection model to generate inspection reports for the aircraft to be inspected.

[0090] The inspection report is used to indicate whether there are any abnormalities on the aircraft surface of each of the aircraft to be inspected. The inspection anomaly detection model is a machine learning-based model that can automatically identify abnormalities in the inspection image. The inspection report is a document that records the inspection results of each aircraft in detail, including information such as whether there are abnormalities, the location and type of abnormalities, etc.

[0091] The inspection image is input into the inspection anomaly detection model for processing. The model automatically identifies the aircraft parts in the image and detects whether there are any anomalies. An inspection report containing the inspection results is generated. Exemplarily, the server can input the inspection image obtained from the aforementioned step S102 into the pre-trained inspection anomaly detection model. The model performs a deep analysis of the image, automatically identifies a tiny crack in the wing, and records the location, size, shape and other information of the crack in detail in the inspection report.

[0092] S104: If the inspection report includes abnormal information, obtaining a processing decision corresponding to the abnormal information according to the abnormal information.

[0093] The abnormal information is used to indicate that there is an abnormality on the aircraft surface of the aircraft to be inspected. The abnormal information may include abnormal conditions on the aircraft surface recorded in the inspection report, such as cracks, corrosion, etc. The processing decision is a solution or maintenance suggestion proposed for the abnormal information.

[0094] Based on the inspection report, the server determines the abnormal location, abnormal situation and other information of the inspected aircraft, and queries the processing decision corresponding to the information based on the abnormal location, abnormal situation and other information, and outputs the processing decision to the maintenance personnel. The maintenance personnel repair the aircraft in a timely manner based on the processing decision. For example, the server finds a small crack in the wing part through the inspection report. Based on the size and location of the crack, the server queries the processing decision in its database and determines the maintenance plan corresponding to the abnormality. Subsequently, the server sends the maintenance plan to the maintenance personnel and guides them to repair the crack in a timely manner.

[0095] The method provided in the embodiment of the present application dynamically adjusts the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected, obtains the inspection images of each aircraft to be inspected during the aircraft inspection process according to the inspection requirements of each aircraft to be inspected, inputs the inspection images of each aircraft to be inspected into the inspection abnormality detection model, generates the inspection report of each aircraft to be inspected, and if the inspection report includes abnormal information, obtains the processing decision corresponding to the abnormal information according to the abnormal information. The method of the present application can effectively reduce the manual participation in the aircraft inspection process, reduce the manual workload, and improve the efficiency of the aircraft inspection process. In addition, through the image recognition method, the abnormal situation existing in the inspection image of the aircraft is automatically identified, and the corresponding processing decision is given for the abnormal situation, which further improves the accuracy of aircraft inspection and abnormal handling.

[0096] In a possible implementation, the aforementioned step S101 may include:

[0097] S1011. Acquire the inspection auxiliary information and the inspection demand analysis network.

[0098] The inspection requirement analysis network is used to determine the inspection requirement of each of the aircraft to be inspected according to the inspection auxiliary information.

[0099] Among them, the inspection demand analysis network is a neural network model that, through training and learning, can intelligently determine the specific inspection requirements of each aircraft to be inspected based on the input inspection auxiliary information.

[0100] The server can extract or receive inspection auxiliary information from the database, including historical flight data (such as flight duration, route, etc.), historical inspection data (past inspection records and problems found), and flight environment data (such as weather conditions, specific environmental factors in the flight area). Load or access the pre-trained inspection demand analysis network model to prepare for subsequent data analysis.

[0101] For example, the server retrieves relevant data of five aircraft to be inspected from the internal database, including their historical flight records, past inspection reports, and recent flight environment information. At the same time, the server loads a trained inspection demand analysis network that can predict possible problems and inspection priorities for each part of the aircraft based on this data.

[0102] S1012. Perform abnormal trend analysis on the inspection auxiliary information according to the inspection demand analysis network, and determine the set of high-risk monitoring points of the aircraft to be inspected based on the historical flight data and / or historical inspection data and / or flight environment data, as well as the inspection priority of the high-risk monitoring points in the set of high-risk monitoring points.

[0103] Among them, the server can use the inspection demand analysis network to conduct in-depth analysis of the collected inspection auxiliary information to predict possible abnormal conditions or failure trends of various aircraft components. According to the abnormal conditions or failure trends that may occur in various aircraft components, the set of aircraft components or areas that need to be focused on and inspected during the inspection process is determined through analysis. For each point in the set of high-risk monitoring points, a priority is assigned according to its potential risk level and urgency to guide the order and focus of the inspection.

[0104] For example, the server can input the collected data into the inspection demand analysis network, and after analysis, the network outputs the high-risk monitoring points that need to be focused on for each aircraft, such as a specific area of ​​the wing, the connection of the engine, etc. At the same time, the network also assigns different inspection priorities to these monitoring points. For example, a certain area of ​​the wing is given the highest inspection priority because it has multiple damage records in history.

[0105] Specifically, step S1012 may include:

[0106] S10121. The inspection demand analysis network may determine abnormal score calculation data of each aircraft component feature node according to the inspection auxiliary information and a mapping relationship between the inspection auxiliary information and the aircraft component feature node.

[0107] The aircraft component feature nodes refer to the feature points of each key component or structure of the aircraft, which can reflect the status and performance of the component. The abnormal score calculation data: based on the mapping relationship between the inspection auxiliary information and the aircraft component feature nodes, the data used to calculate the abnormal score is obtained.

[0108] The server has a mapping relationship between the acquired inspection auxiliary information and the aircraft component feature nodes. The server uses these data and mapping relationships to extract the abnormal score calculation data related to each aircraft component feature node. For example, for the wing component of the aircraft, the server will extract the wind speed and wind direction information from the historical flight data related to the component, the wing damage record from the historical inspection data, and the temperature, humidity and other information from the flight environment data as the basis for calculating the abnormal score of the wing component.

[0109] S10122. Perform a correlation analysis on the abnormal score calculation data and the abnormal state of the aircraft component through a correlation analysis network to obtain the degree of deviation between the current state of the characteristic node of the aircraft component and the normal state, as well as the abnormal trend.

[0110] The correlation analysis network may be a neural network model for analyzing the correlation between the abnormal score calculation data and the abnormal state of the aircraft component. The deviation degree indicates the degree of difference between the current state of the aircraft component and the normal state. The abnormal trend is a component performance change trend predicted based on historical data and current data.

[0111] The server inputs the extracted abnormal score calculation data into the correlation analysis network. The network will perform correlation analysis on these data and the abnormal status of aircraft components. Taking wing components as an example, the network will analyze the correlation between environmental factors such as wind speed, wind direction, temperature, humidity and wing damage, calculate the degree of deviation between the current state of the wing component and the normal state, and predict abnormal trends. For example, the network may find that high wind speed and specific wind direction are highly correlated with wing damage, thereby determining the degree of deviation of the wing component in the current environment and possible abnormal trends.

[0112] S10123. Perform fusion calculation on the degree of deviation between the current state of the aircraft component feature node and the normal state, as well as the abnormal trend, to generate an abnormal score for the aircraft component feature node.

[0113] The server generates an anomaly score for each aircraft component feature node based on the fusion calculation results of the deviation degree and abnormal trend. This score reflects the abnormal risk and potential problems of the component in the current environment. If the deviation degree of the aircraft component is large and the abnormal trend is obvious, then its anomaly score will be high.

[0114] For example, the server uses a weighted average algorithm to perform fusion calculation based on the deviation degree and abnormal trend of the wing feature nodes, and obtains a specific abnormality score (e.g., 75 points, the full score is 100 points, the higher the score, the better the status). This score intuitively reflects the current abnormality degree and potential risks of the wing.

[0115] S10124. According to the abnormality score of the aircraft component feature node and the abnormality score threshold, the aircraft component feature nodes having a score higher than the abnormality score threshold are used as a set of high-risk monitoring points of the aircraft to be inspected.

[0116] The anomaly score threshold is a preset score value used to determine whether an aircraft component is in a high-risk state. The server compares the anomaly score of each aircraft component feature node with the preset anomaly score threshold. Component nodes above the threshold are determined to be high-risk monitoring points and form a high-risk monitoring point set. For example, if the anomaly score of a wing component exceeds the threshold, it will be added to the high-risk monitoring point set.

[0117] Exemplarily, the anomaly score (75 points) of the wing feature node is compared with a preset anomaly score threshold (80 points), and it is found that the score is lower than the threshold, so the wing feature node can be added to the high-risk monitoring point set.

[0118] S10125. Determine the inspection priority of the high-risk monitoring points in the set of high-risk monitoring points according to the abnormality score of each of the high-risk monitoring points in the set of high-risk monitoring points.

[0119] In the set of high-risk monitoring points, the server can determine their inspection priority based on the anomaly score of each high-risk monitoring point. The higher the anomaly score of a component node, the higher its inspection priority. For example, if the wing component has the highest anomaly score among all high-risk monitoring points, it will be given the highest inspection priority to ensure that it can be inspected and processed first during the inspection process.

[0120] For example, in the high-risk monitoring point set, in addition to the wing feature node, there is also a feature node for the engine, whose anomaly score is 70 points. The server sorts the two nodes according to their anomaly scores and determines that the inspection priority of the wing feature node is lower than that of the engine feature node, because the engine feature node has a lower anomaly score, indicating that its status is worse or the risk is higher. Therefore, this feature node of the engine will be checked first during the inspection.

[0121] S1013: Adjust the inspection requirements of each of the aircraft to be inspected according to the high-risk monitoring point set and the inspection priority of the high-risk monitoring points in the high-risk monitoring point set.

[0122] After the server has completed the analysis of the inspection auxiliary information of the aircraft to be inspected and determined the set of high-risk monitoring points of each aircraft and the inspection priority of these monitoring points, it is necessary to adjust the inspection plan of each aircraft based on this information.

[0123] Taking a specific aircraft to be inspected as an example, suppose the server finds through analysis that the wing joint and the engine air intake of this aircraft are two high-risk monitoring points, and the inspection priority of the wing joint is higher than that of the engine air intake. Based on these findings, the server can make at least one of the following adjustments:

[0124] Adjust the focus of image acquisition: For the wing joints, the server can instruct the inspection equipment (such as drones or inspection robots) or inspection personnel to increase the frequency and clarity of shooting at this part when collecting inspection images. For example, the drone can be set to perform multi-angle, high-resolution image acquisition of the wing joints during the inspection process. For engine blades, although their inspection priority is relatively low, it is still necessary to ensure the coverage and clarity of image acquisition, but the frequency and angle of acquisition can be slightly less than that of the wing joints.

[0125] Optimizing image processing algorithms: The server can adjust or optimize image processing algorithms to more accurately identify and analyze anomalies at wing joints and engine blades. This may include enhancing algorithms such as edge detection, color analysis, or shape recognition to increase the accuracy of automatically identifying potential problems.

[0126] Customized image recognition models: Based on historical data and current high-risk monitoring points, the server may train or adjust specialized image recognition models to more accurately identify cracks, corrosion, or other damage on wing joints and engine blades.

[0127] Set image recognition priority: When performing automatic image recognition, the server will process the images of the wing joints first to ensure that high-priority monitoring points can receive faster analysis and feedback.

[0128] The method provided in the embodiment of the present application uses the inspection demand analysis network to deeply analyze and process the inspection auxiliary information to intelligently determine the specific inspection needs of each aircraft to be inspected, realize personalized inspection plan formulation, optimize inspection resource allocation, and reduce blindness and repetitiveness in the inspection process. Through abnormal trend analysis, the method accurately determines the set of high-risk monitoring points of the aircraft to be inspected and the inspection priority of these monitoring points, so that the inspection work can be more focused on areas with higher potential risks, and possible problems can be discovered and handled in a timely manner, thereby improving the safety and reliability of aircraft operations.

[0129] In a possible implementation, the aforementioned step S102 may include:

[0130] S1021. Determine inspection image parameters required to be collected for feature nodes of different aircraft components of each of the aircraft to be inspected according to the inspection requirements of each of the aircraft to be inspected.

[0131] The inspection image parameters include the number of the inspection images, and / or the angle of the inspection image, and / or the clarity of the inspection image.

[0132] The server can analyze the inspection requirements of each aircraft to be inspected. These requirements may include special attention to the characteristic nodes of different aircraft components, such as the connection of the wing, the engine air intake, etc. Based on these requirements, the server determines the specific parameters for collecting inspection images of these characteristic nodes.

[0133] For example, for the joint of the wing, due to its complex structure and easy to be damaged, the server may determine that it is necessary to collect multiple high-definition images to show the details of the joint from different angles. Therefore, the inspection image parameters for this feature node may include: the number of images is 10, each taken from different angles (such as the front, side, oblique angle, etc.), and the clarity of each image must reach 1080P or above.

[0134] S1022. Extracting acquisition features of the inspection image parameters through an intelligent image capture model, identifying and classifying the image features, and generating image acquisition parameters for acquiring inspection images of each of the aircraft to be inspected.

[0135] The intelligent image capture model is a machine learning-based model that can analyze and extract image features to optimize the image acquisition process. The image features are key information in the image that can reflect its content and attributes, such as edges, textures, colors, etc. The image acquisition parameters include the number of inspection images set in order to obtain images that meet inspection requirements, and / or the angle of the inspection image, and / or the clarity of the inspection image, etc.

[0136] The server can use the intelligent image capture model to further refine the acquisition parameters of the inspection image. This model can identify the unique acquisition features of the characteristic nodes of different aircraft components, such as specific shapes, textures, or colors. Taking the wing joint as an example, the intelligent image capture model may identify key components such as bolts and sealing strips at the joint, and determine the best shooting angle and lighting conditions based on the characteristics of these components. In this way, the generated image acquisition parameters will be more accurate, ensuring that the collected inspection images can clearly show the status of these key components.

[0137] S1023: Collect inspection images of each of the aircraft to be inspected according to the image acquisition parameters.

[0138] Based on the precise image acquisition parameters generated by the intelligent image capture model, the server can guide inspection equipment (such as drones or inspection robots) or inspection personnel to collect inspection images. In this process, the server will ensure that the collected images strictly meet the previously determined parameter requirements, such as quantity, angle, and clarity.

[0139] Taking the example of a server guiding a drone inspection device, the server can instruct the drone to shoot at a predetermined angle and distance to ensure that each image can clearly show the details of the connection. At the same time, the server will also perform real-time quality checks on the collected images to ensure that the image quality meets the needs of subsequent analysis.

[0140] Taking the example of the server guiding the inspectors, the server can instruct the inspectors to use image acquisition devices equipped with high-resolution lenses to take pictures of various characteristic nodes of the wing according to the image acquisition parameters provided by the intelligent image capture model. After the shooting is completed, the inspectors can immediately check the quality of the image to ensure that the acquired image meets the requirements of the image acquisition parameters.

[0141] The method provided in the embodiment of the present application determines the inspection image parameters required to be collected for the characteristic nodes of different aircraft components of each of the aircraft to be inspected according to the inspection requirements of each of the aircraft to be inspected, extracts the collection features of the inspection image parameters through an intelligent image capture model, identifies and classifies the image features, generates image collection parameters for collecting inspection images of each of the aircraft to be inspected, and collects inspection images of each of the aircraft to be inspected according to the image collection parameters. Thus, the collection parameters for collecting inspection images of the aircraft to be inspected are adaptively determined according to the inspection requirements of the aircraft to be inspected, thereby improving the collection efficiency and accuracy of inspection images.

[0142] In a possible implementation, the anomaly detection model includes a visual recognition model and a deep learning model, and the aforementioned step S103 may include:

[0143] S1031. Extract image features from the inspection image using a visual recognition model.

[0144] The image features include the shape of the aircraft parts, the texture of the aircraft parts, and the color of the aircraft parts of the aircraft to be inspected. The visual recognition model can be any existing algorithm model for visual recognition, which is used to automatically extract key information from the inspection image. In the present application, the key information includes information for judging whether there are any inspection anomalies in the aircraft to be inspected. For example, the visual recognition model can extract key information such as whether the shape of the aircraft parts in the inspection image is deformed, whether the color of the aircraft parts changes, and whether there are any abnormalities in the texture of the aircraft parts, so as to determine whether the aircraft parts of the aircraft to be inspected may be damaged or other abnormal conditions.

[0145] The inspection image is feature extracted through the visual recognition model to obtain the image features of the aircraft component shape, aircraft component texture, aircraft component color, etc. of the aircraft to be inspected. Based on these image features, it is possible to determine whether the aircraft component meets the standards to determine whether the aircraft component can work normally. For example, whether the shape of the wing meets the standard profile, whether the texture of the engine blade is regular, whether the paint color of the aircraft surface is uniform, etc., to provide a data basis for subsequent abnormality detection.

[0146] S1032. Map the image features in the inspection image to the high-dimensional space of the deep learning model through the trained deep learning model to generate a feature vector corresponding to the image feature.

[0147] After extracting the image features, the server can input these image features into a trained deep learning model, which is used to map these image features into a high-dimensional space and generate a unique feature vector for each feature. In the high-dimensional space, these feature vectors can more clearly show the subtle differences between the components, thereby helping the server to more accurately identify anomalies. Generating feature vectors corresponding to the image features in the high-dimensional space through the deep learning model can improve the sensitivity of anomaly detection and provide richer and more detailed information for subsequent data analysis.

[0148] S1033. Determine abnormal features of feature nodes of each aircraft component of the aircraft to be inspected based on similarity calculation results between feature vectors corresponding to the image features and standard feature vectors, and mark abnormal states and / or abnormal positions corresponding to the abnormal features.

[0149] The standard feature vector represents the feature vector of the aircraft component in a normal state, and the standard feature vector can be obtained through training with a large number of positive samples.

[0150] After generating the feature vectors, the server can compare these vectors with pre-stored standard feature vectors. Standard feature vectors are constructed based on the features of normal aircraft parts and represent the "standard appearance" of aircraft parts under normal working conditions. By calculating the similarity between the feature vector to be detected and the standard feature vector, the server can determine which parts may have abnormalities. For example, if the color feature vector of a part is significantly different from the standard vector, then the part may have problems such as paint peeling or rust. The server will accurately mark these abnormal features and indicate their corresponding abnormal states and locations so that maintenance personnel can quickly locate and deal with the problem.

[0151] For example, if the similarity between a feature vector and a standard feature vector is lower than a set similarity threshold, the aircraft component corresponding to the feature vector may be determined to be abnormal, and the abnormal state and / or abnormal position may be marked. For example, if the texture features of a certain area of ​​the wing are significantly different from the standard features, then this area may be marked as abnormal.

[0152] S1034: Generate an inspection report for each of the aircraft to be inspected based on the abnormal features of the feature nodes of each of the aircraft components of each of the aircraft to be inspected.

[0153] After completing the anomaly detection of all aircraft component feature nodes, the server can generate an inspection report based on the results of the anomaly detection. The inspection report lists in detail all the marked abnormal features on each aircraft to be inspected, including information such as the type, location, and severity of the anomaly. In addition, the inspection report can also sort the anomalies according to the urgency and impact of the anomalies to ensure that maintenance personnel can prioritize those issues that have the greatest impact on flight safety. This inspection report not only provides clear guidance for maintenance personnel, but also greatly improves the efficiency and accuracy of aircraft maintenance. Through this report, airlines can also better understand the overall condition of the aircraft and formulate more reasonable maintenance plans.

[0154] The method provided in the embodiment of the present application utilizes a visual recognition model to extract image features from the inspection image, maps the image features in the inspection image to the high-dimensional space of the deep learning model through a trained deep learning model to generate a feature vector corresponding to the image feature, determines the abnormal features of the feature nodes of each aircraft component of the aircraft to be inspected based on the similarity calculation result between the feature vector corresponding to the image feature and the standard feature vector, marks the abnormal state and / or abnormal position corresponding to the abnormal features, and generates an inspection report for each aircraft to be inspected based on the abnormal features of the feature nodes of each aircraft component of each aircraft to be inspected, thereby realizing efficient, accurate and intelligent processing of the aircraft inspection process and improving the efficiency and safety of aircraft inspection.

[0155] In a possible implementation, the aforementioned step S104 may include:

[0156] S1041. Obtain a first knowledge graph corresponding to the exception information and a second knowledge graph corresponding to the processing decision.

[0157] The first knowledge graph and / or the second knowledge graph may be obtained from a pre-built knowledge base, for example, from a pre-built aviation domain knowledge base, which contains knowledge data corresponding to aircraft parts, common faults, maintenance procedures, and safety standards, and which may be generated by organizing technical documents of aircraft manufacturers, professional books in the aviation field, historical inspection records, and maintenance reports.

[0158] Specifically, taking the first knowledge graph and / or the second knowledge graph being obtained from a pre-built aviation field knowledge base as an example, step 1041 may specifically include:

[0159] Using a natural language processing algorithm, entity recognition and relationship extraction are performed on the abnormal information to determine key entity nouns in the abnormal information and the relationship between the key entity nouns, and generate a first processing result. Among them, the key entity nouns may include, for example, the names of aircraft parts such as wings and noses, and may also include fault types such as cracks and dents. The relationship between the key entity nouns may be, for example, an association relationship between the names of aircraft parts and the fault types.

[0160] Based on the first processing result, a first knowledge graph is constructed, in which each node represents a key entity noun, and the edge represents the relationship between the key entity nouns. Exemplarily, in the first knowledge graph, each node can represent a key entity noun, such as "wing", "crack", etc. The edge between nodes can represent the fault type corresponding to the aircraft part, for example, the edge between the wing node and the crack node represents the existence of a crack on the wing.

[0161] In one possible example, the entity recognition result of the natural language processing algorithm on the abnormal information may include three entity nouns: airplane, left wing, and crack. The relationship extraction result of the natural language processing algorithm on the abnormal information may include: airplane-contains-left wing, left wing-appears-crack. Correspondingly, in the first knowledge graph, the nodes include "airplane", "left wing", and "crack", and the edges represent the relationship between each two nodes (for example, the edge between airplane and left wing represents the inclusion relationship).

[0162] Further, decision data related to the processing decision is obtained, and the decision data comes from a maintenance manual, a best practice guide, a historical processing case, etc. A natural language processing algorithm is used to perform entity recognition and relationship extraction on the decision data to generate a second processing result. Based on the second processing result, a second knowledge graph is constructed, in which each node represents a processing strategy, a maintenance tool or a spare part, and an edge represents a logical relationship or dependency relationship between nodes.

[0163] The entity nouns of the decision data are nodes of the second knowledge graph, and may include nouns such as processing strategy, maintenance tool or spare parts. For example, the processing strategy is left wing crack repair, the maintenance tool is a repair tool, and the material is a crack repair material.

[0164] The relationship between the entity nouns is used as the edge between the nodes of the second knowledge graph. For example, the relationship extraction result may include: left wing crack repair-needs-repair tools, left wing crack repair-needs-crack repair materials.

[0165] After generating the first knowledge graph and the second knowledge graph, an association between the first knowledge graph and the second knowledge graph is established, specifically by adding cross-knowledge-graph edges to the first knowledge graph and the second knowledge graph.

[0166] Exemplarily, continuing with the above example, for instance, an edge that crosses the first knowledge graph and the second knowledge graph can be added between the "crack" node and the "left-wing crack repair" node, indicating that when a left-wing crack is detected, the processing decision of left-wing crack repair should be taken.

[0167] S1042. Determine the associated abnormal information of the abnormal information according to the association path of the abnormal information in the first knowledge graph, where the associated abnormal information is the abnormal information corresponding to the first abnormal entity that has an abnormal association with the abnormal information.

[0168] Exemplarily, after obtaining the first knowledge graph, the server can analyze the associated abnormal information of the wing crack. By traversing the association paths in the graph, the server discovers the first abnormal entity - "wing structure fatigue" - that has an abnormal association with the wing crack. This is because wing cracks are often a direct manifestation of structural fatigue. By further exploring the abnormal information corresponding to "wing structure fatigue", it is determined that this fatigue may be caused by long-term flight stress, environmental factors, or improper maintenance. These information provide important background for subsequent processing decisions.

[0169] Specifically, the target node corresponding to the abnormal information can be located in the first knowledge graph by matching the keywords in the abnormal information. This step can be achieved by the following method:

[0170] Extract keywords from the abnormal information, where the keywords at least include aircraft component names, fault types, abnormal phenomena, etc. Precisely extract keywords from the abnormal information. These keywords mainly include aircraft component names (such as wings, engines, etc.), fault types (such as cracks, fractures, etc.), and abnormal phenomena (such as oil leakage, overheating, etc.). The purpose of this step is to screen out the most critical and representative vocabulary from a large amount of text information for subsequent precise matching in the knowledge graph.

[0171] After extracting the keywords, preprocess the keywords, where the preprocessing includes removing stop words and stemming. Among them, stop words refer to words that frequently appear in the text but contribute little to the meaning of the text, such as "de" (的), "le" (了), etc. These words will generate noise during the search process and thus need to be removed. Stemming is to restore the vocabulary to its basic form. For example, restoring "flying" to "fly", which can improve the accuracy and efficiency of the search.

[0172] The word embedding algorithm is used to map the preprocessed keywords into a high-dimensional vector space to generate keyword vectors. For example, a word embedding algorithm (such as Word2Vec, GloVe, etc.) can be used to map the preprocessed keywords into a high-dimensional vector space to generate corresponding keyword vectors. The keyword vectors generated by the word embedding algorithm can capture the semantic relationship between words, so that semantically similar words are closer in the vector space.

[0173] All nodes in the first knowledge graph are traversed by an approximate nearest neighbor search algorithm, and the cosine similarity between each node and the keyword vector is calculated. Among them, cosine similarity is an indicator to measure the similarity of two vectors in direction, and its value range is between -1 and 1. The closer the value is to 1, the more similar the two vectors are. For example, all nodes in the first knowledge graph can be traversed, and the cosine similarity between each node and the keyword vector can be calculated during the traversal process. If the cosine similarity between any node and the keyword vector is greater than a set threshold, the node is determined as a target node corresponding to the abnormal information.

[0174] Optionally, if there are multiple nodes whose cosine similarities are all greater than a set threshold, they are sorted according to the size of the cosine similarities, and the node with the highest similarity is selected as the target node.

[0175] After locating the target node corresponding to the abnormal information, the target node can be used as a starting point to explore potential associated paths along the edges starting from the starting point. This method can be implemented as follows:

[0176] Initialize the exploration environment, specifically including loading the complete graph data of the first knowledge graph, which includes all nodes (such as aircraft parts, fault types, etc.) and edges (representing the relationship between nodes), and ensure that all data is correctly loaded into memory for subsequent exploration operations.

[0177] Set the corresponding exploration parameters, including the maximum exploration depth (i.e., starting from a node, the maximum number of nodes that can be traversed to find the associated path) and the step size of each exploration (the number of nodes in the next layer that can be traversed from the current node during an exploration process). The setting of these parameters can help control the scope and accuracy of the exploration.

[0178] A breadth-first search algorithm is adopted, with the target node as the starting point, and exploration begins along the edges connected to it, recording each node and edge passed through, as well as the degree of association between each node and edge and the starting point node. During the exploration process, associated paths that meet the preset mining value are screened based on set evaluation criteria, and the evaluation criteria include evaluation criteria based on node attributes, edge types or weights.

[0179] Specifically, starting from the target node, the breadth-first search algorithm is used to start the exploration. First, all the adjacent nodes of the starting node are visited, and then the adjacent nodes of these adjacent nodes are visited, and so on. During the exploration process, each node and edge passed through is recorded, as well as their degree of association with the starting node. The degree of association can be measured in many ways, such as the weight of the edge, the similarity between nodes, etc. This information is very helpful for the subsequent analysis of the importance of the associated path.

[0180] At the same time, during the exploration process, it is also necessary to screen association paths that meet the preset mining value based on the set evaluation criteria. These evaluation criteria can include node attributes (such as node importance, activity, etc.), edge types (such as edges representing causal relationships, edges representing similarity relationships, etc.) or edge weights. By setting these criteria, it can be ensured that the explored association paths are meaningful rather than worthless random paths. Optionally, the explored association paths can also be visualized.

[0181] After exploring all potential associated paths, all candidate nodes directly or indirectly connected to the starting point node are analyzed and recorded, and the candidate nodes represent other abnormal information associated with the abnormal information. The associated abnormal information is sorted and screened according to the degree of association and the path length of the potential associated path to generate screened associated abnormal information.

[0182] S1043. Determine at least one processing decision entity in the second knowledge graph corresponding to the first abnormal entity and the second abnormal entity based on the second abnormal entity corresponding to the abnormal information, the first abnormal entity, and the correlation between the visualization data matrix of the first knowledge graph and the visualization data matrix of the second knowledge graph.

[0183] The server needs to determine the processing decision entity in the second knowledge graph based on the second abnormal entity corresponding to the abnormal information (for example, "wing crack"), the first abnormal entity ("wing structure fatigue"), and the correlation between the two knowledge graphs.

[0184] By comparing the visualized data matrices of the two graphs, the server found that "wing crack repair program" is highly correlated with "wing crack" and "wing structural fatigue". In addition, "structural fatigue detection and prevention" is also closely related to these two abnormal entities. Therefore, the server identified these two processing decision entities as the key to subsequent processing.

[0185] S1044. Determine a processing decision corresponding to the abnormal information according to the processing decision corresponding to the processing decision entity.

[0186] The server determines the specific processing decision for the wing crack anomaly based on the determined processing decision entities (such as "wing crack repair procedure" and "structural fatigue detection and prevention").

[0187] For example, for a "wing crack repair procedure", the server can retrieve detailed step-by-step instructions, a list of required tools and materials, and safe operation instructions. This information will directly guide maintenance personnel to perform crack repair work.

[0188] Optionally, the “Structural Fatigue Detection and Prevention” treatment decision can also provide a systematic set of detection methods and preventive measures aimed at avoiding the recurrence of similar crack anomalies in the future.

[0189] The technical solution described in the embodiment of the present method, by combining knowledge graph and natural language processing technology, provides intelligent processing decision support for abnormal information found during aircraft inspection, thereby improving the efficiency and safety of aircraft maintenance.

[0190] In a possible implementation, determining the processing decision corresponding to the abnormal information according to the processing decision corresponding to the processing decision entity may also include:

[0191] Obtain the maintenance history information, spare parts inventory, and maintenance support information of the aircraft to be inspected. The maintenance history information of the aircraft to be inspected includes but is not limited to the aircraft's past maintenance records, replacement of parts, results of regular inspections, etc. By analyzing this information, it is possible to understand which parts of the aircraft often have problems and which parts have or are about to reach their service life, which helps to predict possible problems in the future. The spare parts inventory is crucial for understanding the types, quantities, and storage status of currently available spare parts. If a processing decision requires the replacement of a specific part, but the part is not in inventory, then the decision cannot be executed immediately. Maintenance support information may include the number of available maintenance personnel, skill levels, and available maintenance equipment and tools. If a processing decision requires specific skills or equipment, and these resources are currently unavailable, then the decision cannot be implemented either.

[0192] Adjust the processing decision according to the maintenance history information, the spare parts inventory status, the repair support information, and the processing decision corresponding to the processing decision entity. Compare and analyze the collected maintenance history information, spare parts inventory status and repair support information with the preliminary processing decision corresponding to the processing decision entity, evaluate the feasibility and efficiency of the preliminary processing decision, and make necessary adjustments to the decision considering the actual situation. Based on the above analysis, if it is found that the preliminary processing decision does not match the current actual situation, it is necessary to make adjustments. For example, if the parts that need to be replaced in the preliminary decision are currently out of stock, it may be necessary to change to other available spare parts, or adjust the maintenance plan to wait for the spare parts to arrive. For example, if the maintenance personnel or equipment required in the preliminary decision are currently unavailable, it may also be necessary to adjust the work plan or seek external support.

[0193] The adjusted processing decision is determined as the processing decision corresponding to the abnormal information. After the above adjustment, the obtained processing decision will be more in line with the current actual situation, and it can be determined as the final processing decision of the abnormal information. This decision not only takes into account the condition of the aircraft itself, but also comprehensively considers actual operational factors such as maintenance resources and spare parts inventory, so it is more executable and efficient.

[0194] The method provided in the embodiment of the present application can ensure that the processing decisions made for each abnormal information are based on comprehensive information analysis and actual situation considerations, thereby improving the accuracy and efficiency of maintenance work.

[0195] Optionally, the method may also include content of multi-party terminal interaction to provide a function capable of manually handling emergency exceptions in real time or handling existing exceptions automatically.

[0196] Specifically, if there is abnormal feedback information for the processing decision corresponding to the abnormal information, and a call request sent by the inspection terminal is received, then according to the call request and the state of the maintenance terminal, an idle maintenance terminal is called, and the call request is used to request a call with the maintenance terminal for abnormal feedback. If the idle maintenance terminal passes the call request, a call connection between the inspection terminal and the idle maintenance terminal is established. If the idle maintenance terminal does not respond to the call request, other idle maintenance terminals are called until the call connection between the inspection terminal and the other idle maintenance terminals is established.

[0197] The embodiment of the present application also provides an aircraft inspection system. The aircraft inspection system may include an inspection terminal, a server, and a maintenance terminal. Among them, the server can implement the method steps performed by the server described in the aforementioned method embodiment. The inspection terminal is used by inspection personnel to inspect the aircraft to be inspected, collect inspection images, etc. The inspection terminal uploads the inspection images and information collected during the inspection process to the server, and the server generates an inspection report based on the inspection images and information, and sends the inspection report to the maintenance terminal. The maintenance terminal is used for back-end maintenance personnel to monitor and assist the aircraft inspection process, processing decision execution process, etc. according to the inspection report.

[0198] In a possible implementation, the server may include a proxy application for a mobile inspection real-time service, a core routing service, a mobile inspection real-time main service, a background application of an inspection management platform, a file transfer service, a recording service, etc.

[0199] The proxy application of the mobile inspection real-time service is an intermediate layer that uses a load balancing architecture to ensure stable and reliable services under high concurrency. The main task of this proxy application is to receive connection requests from the inspection terminal and forward them to the mobile inspection real-time main service of the maintenance terminal. The proxy application of the mobile inspection real-time service allows the inspection terminal to establish a persistent connection with the server through the WebSocket protocol to achieve two-way transmission of real-time data.

[0200] Core Routing Service The core routing service is the central nervous system of the entire system, responsible for managing user login and logout, state switching, client routing requests, client connection management, and LiveKit rooms (rooms for audio and video calls) and events (receiving and processing webhook events from LiveKit, such as changes in room status). It is also responsible for receiving and applying the system's dynamic configuration from Redis, and publishing various status events to Redis for use by other applications or services.

[0201] The mobile inspection real-time main service is the core service for processing real-time inspection data. It provides a WebSocket interface to support real-time data exchange, including the exchange of audio and video streams, text messages, etc. Inspection terminals can connect to the mobile inspection real-time service through the agent application of the mobile inspection real-time service.

[0202] The backend application of the inspection management platform is the core system for managing inspection tasks and data. It adopts a load balancing architecture to ensure high performance and scalability, and provides HTTP interfaces for front-end or other service calls. For example, it can provide interfaces for managing inspection tasks, querying inspection data, configuring inspection plans, etc.

[0203] The file transfer service is responsible for processing file upload and download requirements on the server, especially for processing pictures (such as inspection images) and files (such as inspection-related files reported by inspection personnel). It provides picture upload function and returns the uniform resource locator (URL) of the picture for subsequent use. At the same time, it also supports file download service.

[0204] The recording service is responsible for recording the audio and video content during the inspection process. It provides two recording modes: room-wide recording and track-independent recording. Room-wide recording combines the video and audio of all participants in the inspection into one audio and video file. Track-independent recording records the audio and video streams of each participant separately.

[0205] Optionally, in the above-mentioned aircraft inspection system, a multi-tenant management function can also be supported. That is, the aircraft inspection system adopts a multi-tenant architecture, and the tenant division range includes databases, inspection terminals, inspection images taken by inspection terminals, videos recorded by servers, servers, etc. Inspection database instances can be named according to the identifiers of tenants, and different tenants can use different types of database instances or the same database instance.

[0206] Tenant division of the inspection terminal can be achieved by setting the tenant number, and the tenant number can be set or changed through the initial page or setting interface of the inspection terminal. When uploading and downloading the inspection pictures taken by the inspection terminal, the message used can carry the token obtained after logging in. The server obtains the tenant information based on the token, creates a tenant directory under the root directory, and implements isolation between tenants. Similarly, when the server records the audio and video of the room call, it can also implement tenant isolation through a similar mechanism.

[0207] In a possible implementation, the server may distinguish tenants through a URL, and the URL may carry a tenant mask, which may be either a tenant ID or a custom disguised ID.

[0208] Figure 2 The structure diagram of an electronic device provided in an embodiment of the present application. The electronic device is used to execute the aforementioned aircraft inspection method, for example, it can be the aforementioned server. Figure 2 As shown, the electronic device 200 may include: at least one processor 201 , a memory 202 , and a communication interface 203 .

[0209] The memory 202 is used to store programs. Specifically, the programs may include program codes, and the program codes include computer operation instructions.

[0210] The memory 202 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0211] The processor 201 is used to execute the computer-executable instructions stored in the memory 202 to implement the method described in the above method embodiment. The processor 201 may be a CPU, or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0212] The processor 201 can communicate and interact with external devices through the communication interface 203. The external device can be, for example, the inspection terminal or the maintenance terminal mentioned above. In a specific implementation, if the communication interface 203, the memory 202 and the processor 201 are implemented independently, the communication interface 203, the memory 202 and the processor 201 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc., but it does not mean that there is only one bus or one type of bus.

[0213] Optionally, in a specific implementation, if the communication interface 203, the memory 202 and the processor 201 are integrated on a chip, the communication interface 203, the memory 202 and the processor 201 can communicate through an internal interface.

[0214] The present application also provides a computer-readable storage medium, which may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes. Specifically, the computer-readable storage medium stores program instructions, and the program instructions are used for the methods in the above embodiments.

[0215] The present application also provides a program product, which includes an execution instruction, which is stored in a readable storage medium. At least one processor of a computing device can read the execution instruction from the readable storage medium, and at least one processor executes the execution instruction so that the computing device implements the above-mentioned aircraft inspection method.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application 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 replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An aircraft inspection method, characterized in that: The method comprises: Dynamically adjusting the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected, wherein the inspection auxiliary information includes historical flight data, historical inspection data, and flight environment data; According to the inspection requirements of each of the aircraft to be inspected, obtaining inspection images of each of the aircraft to be inspected during the aircraft inspection process; Inputting the inspection images of the aircraft to be inspected into the inspection anomaly detection model to generate an inspection report for each aircraft to be inspected, wherein the inspection report is used to indicate whether there is an abnormality on the aircraft surface of each aircraft to be inspected; If the inspection report includes abnormal information, a processing decision corresponding to the abnormal information is obtained according to the abnormal information, and the abnormal information is used to indicate that an abnormality exists on the aircraft surface of the aircraft to be inspected.

2. The method according to claim 1, characterized in that The dynamically adjusting the inspection requirements of each aircraft to be inspected according to the inspection auxiliary information of each aircraft to be inspected includes: Acquire the inspection auxiliary information and an inspection demand analysis network, wherein the inspection demand analysis network is used to determine the inspection demand of each of the aircraft to be inspected according to the inspection auxiliary information; Performing abnormal trend analysis on the inspection auxiliary information according to the inspection demand analysis network, and determining a set of high-risk monitoring points of the aircraft to be inspected, and an inspection priority of the high-risk monitoring points in the set of high-risk monitoring points based on the historical flight data and / or the historical inspection data and / or the flight environment data; The inspection requirements of each of the aircraft to be inspected are adjusted according to the high-risk monitoring point set and the inspection priorities of the high-risk monitoring points in the high-risk monitoring point set.

3. The method according to claim 2, characterized in that The abnormal trend analysis of the inspection auxiliary information is performed according to the inspection demand analysis network, and a set of high-risk monitoring points of the aircraft to be inspected is determined based on the historical flight data and / or the historical inspection data and / or the flight environment data, as well as the inspection priority of the high-risk monitoring points in the set of high-risk monitoring points, including: The inspection demand analysis network determines the abnormal score calculation data of each aircraft component feature node according to the inspection auxiliary information and the mapping relationship between the inspection auxiliary information and the aircraft component feature node; Performing a correlation analysis on the abnormal score calculation data and the abnormal state of the aircraft component through a correlation analysis network to obtain the degree of deviation between the current state of the characteristic node of the aircraft component and the normal state, as well as the abnormal trend; Performing fusion calculation on the degree of deviation between the current state of the aircraft component feature node and the normal state, as well as the abnormal trend, to generate an abnormal score for the aircraft component feature node; According to the abnormality score of the aircraft component feature node and the abnormality score threshold, the aircraft component feature nodes with a score higher than the abnormality score threshold are used as a set of high-risk monitoring points of the aircraft to be inspected; The inspection priority of the high-risk monitoring points in the high-risk monitoring point set is determined according to the abnormality score of each of the high-risk monitoring points in the high-risk monitoring point set.

4. The method according to claim 3, characterized in that The obtaining, according to the inspection requirements of each of the aircraft to be inspected, inspection images of each of the aircraft to be inspected during the aircraft inspection process includes: According to the inspection requirements of each of the aircraft to be inspected, determining the inspection image parameters required to be collected for the characteristic nodes of different aircraft components of each of the aircraft to be inspected, wherein the inspection image parameters include the number of the inspection images, and / or the angle of the inspection image, and / or the clarity of the inspection image; Extracting acquisition features of the inspection image parameters through an intelligent image capture model, identifying and classifying the image features, and generating image acquisition parameters for acquiring inspection images of each of the aircraft to be inspected; The inspection image of each of the aircraft to be inspected is collected according to the image collection parameters.

5. The method according to claim 4, characterized in that The abnormality inspection model includes a visual recognition model and a deep learning model. The inspection image of each aircraft to be inspected is input into the inspection abnormality inspection model to generate an inspection report for each aircraft to be inspected, including: Extracting image features from the inspection image using a visual recognition model, wherein the image features include the shape of an aircraft component, the texture of an aircraft component, and the color of an aircraft component of the aircraft to be inspected; Mapping the image features in the inspection image to the high-dimensional space of the deep learning model through the trained deep learning model to generate a feature vector corresponding to the image feature; Determine the abnormal features of the feature nodes of each aircraft component of the aircraft to be inspected according to the similarity calculation result between the feature vector corresponding to the image feature and the standard feature vector, and mark the abnormal state and / or abnormal position corresponding to the abnormal feature; An inspection report for each of the aircraft to be inspected is generated according to the abnormal features of the feature nodes of each of the aircraft components of each of the aircraft to be inspected.

6. The method according to claim 1, characterized in that The obtaining, according to the abnormal information, a processing decision corresponding to the abnormal information, comprises: Obtaining a first knowledge graph corresponding to the abnormal information and a second knowledge graph corresponding to the processing decision; Determine, according to the association path of the abnormal information in the first knowledge graph, associated abnormal information of the abnormal information, where the associated abnormal information is abnormal information corresponding to a first abnormal entity that is abnormally associated with the abnormal information; Determine at least one processing decision entity corresponding to the first abnormal entity and the second abnormal entity in the second knowledge graph according to the second abnormal entity corresponding to the abnormal information, the first abnormal entity, and the association between the visualized data matrix of the first knowledge graph and the visualized data matrix of the second knowledge graph; Determine the processing decision corresponding to the abnormal information according to the processing decision corresponding to the processing decision entity; The step of obtaining the first knowledge graph corresponding to the abnormal information and the second knowledge graph corresponding to the processing decision includes: Pre-establishing an aviation field knowledge base, which contains knowledge data corresponding to aircraft components, common faults, maintenance procedures, and safety standards. The aviation field knowledge base is generated by organizing technical documents of aircraft manufacturers, professional books in the aviation field, historical inspection records, and maintenance reports; Using a natural language processing algorithm, perform entity recognition and relationship extraction on the abnormal information, determine key entity nouns in the abnormal information and the relationship between the key entity nouns, and generate a first processing result; Based on the first processing result, construct a first knowledge graph, in which each node represents a key entity noun and each edge represents a relationship between the key entity nouns; and, obtaining decision data related to the processing decision, the decision data being derived from maintenance manuals, best practice guidelines, and historical processing cases; Using a natural language processing algorithm to perform entity recognition and relationship extraction on the decision data to generate a second processing result; Based on the second processing result, construct a second knowledge graph, in which each node represents a processing strategy, a maintenance tool or a spare part, and an edge represents a logical relationship or a dependency relationship between nodes; After generating the first knowledge graph and the second knowledge graph, establishing an association between the first knowledge graph and the second knowledge graph, specifically by adding cross-knowledge graph edges in the first knowledge graph and the second knowledge graph; The step of determining the associated abnormal information of the abnormal information according to the associated path of the abnormal information in the first knowledge graph includes: By matching the keywords in the abnormal information, locating the target node corresponding to the abnormal information in the first knowledge graph, and taking the target node as the starting point, starting from the starting point node, exploring potential association paths along the edges, wherein in the first knowledge graph, the nodes are connected by the edges, indicating the association relationship between the nodes; During the exploration process, each node and edge passed through is recorded, as well as the degree of association between each node and edge and the starting node, and the degree of association is represented by a weight; After exploring all potential associated paths, analyzing and recording all candidate nodes directly or indirectly connected to the starting point node, the candidate nodes representing other abnormal information associated with the abnormal information; Sorting and filtering the associated abnormal information according to the association degree and the path length of the potential associated path, and generating filtered associated abnormal information; The step of locating a target node corresponding to the abnormal information in the first knowledge graph by matching keywords in the abnormal information includes: Extracting keywords from the abnormal information, wherein the keywords at least include aircraft component name, fault type, and abnormal phenomenon; After extracting the keywords, preprocessing the keywords, the preprocessing includes removing stop words and stemming; The word embedding algorithm is used to map the preprocessed keywords into a high-dimensional vector space to generate keyword vectors; Traversing all nodes in the first knowledge graph by using an approximate nearest neighbor search algorithm, and calculating the cosine similarity between each node and the keyword vector; If the cosine similarity between any node and the keyword vector is greater than a set threshold, the node is determined as a target node corresponding to the abnormal information; And, if there are multiple nodes whose cosine similarities are all greater than a set threshold, sort them according to the size of the cosine similarities, and select the node with the highest similarity as the target node; The step of taking the target node as the starting point and exploring potential associated paths along the edges starting from the starting point includes: Initializing the exploration environment, specifically including loading the complete graph data of the first knowledge graph and setting corresponding exploration parameters, wherein the exploration parameters include the maximum exploration depth and the step length of each exploration; Adopting a breadth-first search algorithm, starting from the target node, exploring along the edges connected to it, recording each node and edge passed through, as well as the degree of association between each node and edge and the starting node, and in the exploration process, screening association paths that meet the preset mining value based on set evaluation criteria, wherein the evaluation criteria include evaluation criteria based on node attributes, edge types or weights; Visualize the explored association paths.

7. The method according to claim 6, characterized in that The determining, according to the processing decision corresponding to the processing decision entity, a processing decision corresponding to the abnormal information includes: Obtaining maintenance history information, spare parts inventory status, and maintenance support information of the aircraft to be inspected; adjusting the processing decision according to the maintenance history information, the spare parts inventory status, the maintenance support information, and the processing decision corresponding to the processing decision entity; The adjusted processing decision is determined as the processing decision corresponding to the abnormal information.

8. The method according to any one of claims 1 to 7, characterized in that: The method further comprises: If there is abnormal feedback information for the processing decision corresponding to the abnormal information, and a call request sent by the inspection terminal is received, then according to the call request and the state of the maintenance terminal, an idle maintenance terminal is called, and the call request is used to request a call with the maintenance terminal for abnormal feedback; If the idle maintenance terminal passes the call request, a call connection is established between the inspection terminal and the idle maintenance terminal; If the idle locomotive terminal does not respond to the call request, other idle locomotive terminals are called until the call connection between the inspection terminal and the other idle locomotive terminals is established.

9. An electronic device, characterized in that: include: A processor, a communication interface and a memory, wherein the processor is communicatively connected to the communication interface and the memory respectively; The memory stores computer-executable instructions; The communication interface communicates and interacts with external devices; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the aircraft inspection method according to any one of claims 1 to 8.

Images