Airplane icing detection method and system

Abstract

The invention provides an aircraft icing detection method and system, and belongs to the technical field of aircraft ice prevention and removal, and the method comprises the steps: constructing a data map library of different weather conditions of an aircraft icing monitoring region; based on the data image library, a deep learning algorithm is adopted for training, and an image recognition model is obtained; acquiring a real-time monitoring picture of the aircraft icing monitoring area; inputting the real-time monitoring picture into an image recognition model for processing and analysis, judging whether the current monitoring area is iced or not, and when the current monitoring area is iced, obtaining an icing parameter; according to the icing parameters, the icing grade is calculated; and triggering an alarm mechanism, and displaying differentiated response suggestions according to the icing grade. Through the processing scheme provided by the invention, the accuracy and timeliness of icing detection are improved, the working intensity of a pilot can be remarkably reduced, and a powerful guarantee is provided for flight safety.

Description

Technical Field

[0001] This application relates to the field of aircraft anti-icing technology, and in particular to an aircraft icing detection method and system. Background Technology

[0002] Aircraft icing refers to the phenomenon where atmospheric temperatures below 0 degrees Celsius, with supercooled water droplets in the air, impact the aircraft's windward surface and accumulate into an ice layer. Icing on the wings and tail alters the airfoil, reducing aerodynamic performance, decreasing the critical angle of attack, and increasing the risk of stall, seriously threatening flight safety. To address this issue, modern aviation technology has developed various icing detection methods. Currently used methods include thermal sensors, vibration feature recognition, microwave radar monitoring, and optical monitoring. These technologies capture the physical changes caused by icing accumulation, enabling real-time monitoring of the icing status on the aircraft surface. Optical monitoring uses cameras for real-time monitoring, with manual interpretation of the images to identify icing conditions. However, pilots have heavy workloads while flying and lack the time to constantly monitor the images, easily missing icing information. With the development of AI and intelligent monitoring, the goal is to build a more intelligent aircraft icing detection system to ensure flight safety. Summary of the Invention

[0003] In view of this, the present application provides an aircraft icing detection method and system, which at least partially solves the problem in the prior art that pilots may easily miss icing information when manually judging images.

[0004] In a first aspect, embodiments of this application provide an aircraft icing detection method, the method comprising: Construct a data library of different weather conditions in the aircraft icing monitoring area; Based on the data library, a deep learning algorithm is used for training to obtain an image recognition model; Acquire real-time monitoring images of the aircraft icing monitoring area; The real-time monitoring images are input into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area. When icing occurs, icing parameters are obtained. Calculate the icing level based on the icing parameters; Trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

[0005] According to a specific implementation of an embodiment of this application, the data library for different weather conditions includes: A database of images showing the different degrees of icing under varying lighting conditions during the day, night, sunny days, and rainy days.

[0006] According to a specific implementation of an embodiment of this application, the step of inputting the real-time monitoring image into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area includes: Construct a baseline background model in the image recognition model; Preprocess the real-time monitoring images; The preprocessed real-time monitoring image is compared pixel by pixel with the baseline background model, and the difference between the two is calculated. The number of pixels with a difference value greater than the first preset value; When the number of pixels within a region exceeds a second preset value, the region is determined to be an icing region. Calculate the icing parameters of the icing area.

[0007] According to a specific implementation of an embodiment of this application, the icing parameters include maximum icing thickness, icing coverage area ratio, and ice type characteristics.

[0008] According to one specific implementation of the embodiments of this application, the difference values ​​include grayscale difference values, color difference values, and texture difference values.

[0009] According to a specific implementation of an embodiment of this application, the step of calculating the icing level based on icing parameters includes: The corresponding icing level is calculated based on the maximum increase in icing thickness per minute.

[0010] According to a specific implementation of an embodiment of this application, the icing level includes weak icing, medium icing, strong icing, and extremely strong icing. The maximum icing thickness increment per minute corresponding to weak icing is less than 0.6 mm per minute, the maximum icing thickness increment per minute corresponding to medium icing is 0.6 to 1.0 mm per minute, the maximum icing thickness increment per minute corresponding to strong icing is 1.1 to 2.0 mm per minute, and the maximum icing thickness increment per minute corresponding to extremely strong icing is greater than 2.0 mm per minute.

[0011] According to one specific implementation of an embodiment of this application, the preprocessing includes noise reduction and contrast enhancement.

[0012] According to one specific implementation of an embodiment of this application, the deep learning algorithm employs a convolutional neural network model.

[0013] Secondly, embodiments of this application also provide an aircraft icing detection system for implementing the aircraft icing detection method as described in any embodiment of the first aspect, the system comprising: The image library building module is used to build a data image library of different weather conditions in the aircraft icing monitoring area; The model training module is used to train an image recognition model based on a data library using deep learning algorithms. The data acquisition module is used to acquire real-time monitoring images of the aircraft icing monitoring area; The judgment module is used to input the real-time monitoring screen into the image recognition model for processing and analysis, to determine whether icing has occurred in the current monitoring area, and to obtain icing parameters when icing occurs. The icing level calculation module is used to calculate the icing level based on icing parameters; The alarm module is used to trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

[0014] Beneficial effects: The aircraft icing detection method and system in this application embodiment, based on AI image recognition software, constructs an icing condition database of the monitored area. By comparing the monitoring images in real time, the software intelligently detects icing and issues warnings. The intelligent icing detection system can monitor the icing status of the aircraft surface in real time, quickly identify icing risks through analysis, and promptly issue warning signals to help pilots take de-icing measures or adjust flight altitude and route, effectively preventing air disasters caused by icing. It effectively solves the problems of traditional optical icing detection relying on manual monitoring, which is time-consuming for pilots and prone to oversights. Through automated monitoring and data analysis, the intelligent system can quickly and accurately assess icing risks, reduce pilot workload, and improve the aircraft's intelligence level. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a logic diagram of an aircraft icing detection system according to an embodiment of the present invention. Detailed Implementation

[0017] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0018] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0020] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The illustrations only show the components related to this application and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0022] Firstly, embodiments of this application provide an aircraft icing detection method, referring to... Figure 1 The method includes: Construct a data library of different weather conditions in the aircraft icing monitoring area; Based on the data library, a deep learning algorithm is used for training to obtain an image recognition model; Acquire real-time monitoring images of the aircraft icing monitoring area; The real-time monitoring images are input into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area. When icing occurs, icing parameters are obtained. Calculate the icing level based on the icing parameters; Trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

[0023] In this embodiment, a comprehensive data library is constructed, covering different icing levels under various lighting conditions, including daytime, nighttime, sunny, and rainy days, providing a rich and diverse data foundation for training the image recognition model. Deep learning algorithms are used to train the data library, resulting in an image recognition model capable of accurately identifying icing conditions. After acquiring real-time monitoring images of the aircraft icing monitoring area, these images are input into the trained image recognition model for processing and analysis. This accurately determines whether icing has occurred in the current monitoring area and obtains icing parameters. Based on these parameters, the icing level is further calculated, triggering an alarm mechanism. Differentiated response suggestions are displayed according to different icing levels, helping pilots to take timely de-icing measures or adjust flight altitude and route, effectively preventing air disasters caused by icing. This method not only improves the accuracy and efficiency of icing detection but also significantly reduces the pilot's workload and enhances the aircraft's intelligence level.

[0024] In one embodiment, the database of data on different weather conditions includes: A database of images showing the different degrees of icing under varying lighting conditions during the day, night, sunny days, and rainy days.

[0025] In practice, a professional database of icing monitoring areas for key aircraft components such as wings and tail fins will be constructed. This process involves the comprehensive collection and classification of icing conditions on aircraft component surfaces under different weather conditions, including but not limited to light icing, severe icing, and different icing types. To ensure data quality and diversity, the images in the database will cover samples under different lighting conditions, such as day and night, sunny and rainy days, and will be labeled by professional aerospace engineers to form a standardized dataset, providing training samples for image recognition models.

[0026] In one embodiment, the step of inputting the real-time monitoring image into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area includes: Construct a baseline background model in the image recognition model; Preprocess the real-time monitoring images; The preprocessed real-time monitoring image is compared pixel by pixel with the baseline background model, and the difference between the two is calculated. The number of pixels with a difference value greater than the first preset value; When the number of pixels within a region exceeds a second preset value, the region is determined to be an icing region. Calculate the icing parameters of the icing area.

[0027] In practice, the baseline background model may change over time (e.g., due to changes in lighting or seasons), requiring periodic updates, such as using a moving average method. The first and second preset values ​​need to be adjusted based on the actual scene; the first preset value controls pixel-level sensitivity, and the second preset value controls region-level sensitivity.

[0028] In this embodiment, constructing a baseline background model within the image recognition model is a crucial step, providing a stable reference for subsequent icing determination. Preprocessing the real-time monitoring image, such as denoising and contrast enhancement, effectively improves image quality and reduces interference. Comparing the preprocessed real-time monitoring image pixel-by-pixel with the baseline background model and calculating the difference between the two allows for precise capture of subtle changes in the image. The number of pixels with a difference greater than a first preset value is counted, and further, if the number of pixels within a region exceeds a second preset value, that region is identified as an icing area. Finally, by calculating icing parameters of the icing area, such as maximum icing thickness, icing coverage area percentage, and ice type characteristics, accurate data support is provided for subsequent icing level calculations and alarm triggering. This series of steps not only improves the accuracy and reliability of icing detection but also provides pilots with timely and effective icing information, contributing to flight safety.

[0029] Furthermore, the icing parameters include maximum icing thickness, icing coverage area percentage, and ice type characteristics.

[0030] Furthermore, the difference values ​​include grayscale difference values, color difference values, and texture difference values.

[0031] In practice, for calculating grayscale difference values, the real-time monitoring image and the baseline background model are first converted from color images to grayscale images. For each pixel location, the grayscale difference value is defined as the absolute difference between the current grayscale value and the background grayscale value. Grayscale difference mainly captures brightness changes; icy areas may experience grayscale value changes due to reflections or color brightening. For calculating color difference values, Euclidean distance can be used. Color difference can capture color shifts caused by icing; for example, ice may appear blue or white. For calculating texture difference values, texture feature extraction algorithms are used, such as Local Binary Pattern (LBP), Gray-Level Co-occurrence Matrix (GLCM), or Gabor filters. Texture difference can capture the smoothness or crystallization patterns of icy areas; ice surfaces are often smoother or have specific textures than normal surfaces. Finally, when calculating the difference values ​​between the real-time monitoring image and the baseline background model, different weights can be assigned to the three parameters—grayscale difference, color difference, and texture difference—and linearly weighted fusion can be performed. The weight allocation ratio for each parameter is determined using training data.

[0032] In practical implementation, advanced AI viewing software is used to intelligently analyze real-time monitoring images. This embodiment employs deep learning algorithms, based on a convolutional neural network (CNN) image recognition model, to process and analyze the images captured by the icing monitor. The system first preprocesses the real-time images, including noise reduction and contrast enhancement, to improve recognition accuracy. Then, a data comparison method is used to extract key features such as the shape and size of the ice formation and the area covered by ice. The data comparison method mainly uses video frame difference and background modeling technology. The system first captures multiple frames of images of the monitored scene in an icy state, and generates a stable "baseline background model" through algorithms to eliminate dynamic and static interferences such as lighting, shadows, raindrops, and wing vibration. The real-time captured image frame (the current frame) is compared pixel by pixel with the baseline background model, and the differences in grayscale values, colors, or textures are calculated. When the area, color, or other parameters of the difference region exceed preset values ​​(forming ice), the system determines that ice has formed. Finally, the system compares the size and shape of the ice after it freezes with key features such as the ice-covered area with samples in the image library in multiple dimensions. The main comparisons are the height parameter of the ice shape protruding from the wing surface normal, the area parameter of the icing area covering the leading edge of the wing, and the ice shape. Criterion parameters such as maximum ice thickness, the proportion of the ice-covered area (as a percentage of the total area of ​​the leading edge of the wing), and ice shape characteristics (such as horn ice and frosted ice) are given.

[0033] In one embodiment, calculating the icing level based on icing parameters includes: The corresponding icing level is calculated based on the maximum increase in icing thickness per minute.

[0034] Furthermore, the icing levels include weak icing, medium icing, strong icing, and extremely strong icing. The maximum icing thickness increment per minute corresponding to weak icing is less than 0.6 mm per minute, the maximum icing thickness increment per minute corresponding to medium icing is 0.6 to 1.0 mm per minute, the maximum icing thickness increment per minute corresponding to strong icing is 1.1 to 2.0 mm per minute, and the maximum icing thickness increment per minute corresponding to extremely strong icing is greater than 2.0 mm per minute.

[0035] In practice, the system can calculate the corresponding icing level based on the maximum icing thickness increment per minute criterion parameter and immediately trigger an alarm mechanism. The system issues alarm commands through multiple methods: including sending visual cues (level, icing coverage area percentage, ice type characteristics) to the cockpit, playing voice warnings, and displaying specific icing levels, icing coverage area percentages, and ice types on the flight control panel. It provides differentiated response suggestions based on different icing levels, thus providing a scientific basis for pilot decision-making. Simultaneously, interfaces with other aircraft systems (such as anti-icing and de-icing systems) are reserved to facilitate automated response and closed-loop control.

[0036] Furthermore, the preprocessing includes noise reduction and contrast enhancement.

[0037] Furthermore, the deep learning algorithm employs a convolutional neural network model.

[0038] Secondly, embodiments of this application also provide an aircraft icing detection system for implementing the aircraft icing detection method as described in any embodiment of the first aspect, the system comprising: The image library building module is used to build a data image library of different weather conditions in the aircraft icing monitoring area; The model training module is used to train an image recognition model based on a data library using deep learning algorithms. The data acquisition module is used to acquire real-time monitoring images of the aircraft icing monitoring area; The judgment module is used to input the real-time monitoring screen into the image recognition model for processing and analysis, to determine whether icing has occurred in the current monitoring area, and to obtain icing parameters when icing occurs. The icing level calculation module is used to calculate the icing level based on icing parameters; The alarm module is used to trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

[0039] The embodiments provided by this invention utilize AI image recognition software to establish a database of icing conditions in the monitored area. By comparing real-time monitoring images, it quickly and intelligently identifies potential icing hazards and sends alarm signals. This assists pilots in performing de-icing operations or changing flight altitude and route, effectively preventing air disasters caused by icing, reducing pilot workload, and enhancing the aircraft's intelligence. It significantly improves the accuracy and timeliness of icing detection while markedly reducing pilot workload, providing solid support for flight safety. This pioneering solution fully demonstrates the deep integration of artificial intelligence technology and aeronautical engineering, foreshadowing the future development trend of aircraft safety monitoring systems.

[0040] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting aircraft icing, characterized in that, The method includes: Construct a data library of different weather conditions in the aircraft icing monitoring area; Based on the data library, a deep learning algorithm is used for training to obtain an image recognition model; Acquire real-time monitoring images of the aircraft icing monitoring area; The real-time monitoring images are input into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area. When icing occurs, icing parameters are obtained. Calculate the icing level based on the icing parameters; Trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

2. The aircraft icing detection method according to claim 1, characterized in that, The database of data on different weather conditions includes: A database of images showing the different degrees of icing under varying lighting conditions during the day, night, sunny days, and rainy days.

3. The aircraft icing detection method according to claim 1, characterized in that, The step of inputting real-time monitoring images into an image recognition model for processing and analysis to determine whether icing has occurred in the current monitoring area includes: Construct a baseline background model in the image recognition model; Preprocess the real-time monitoring images; The preprocessed real-time monitoring image is compared pixel by pixel with the baseline background model, and the difference between the two is calculated. The number of pixels with a difference value greater than the first preset value; When the number of pixels within a region exceeds a second preset value, the region is determined to be an icing region. Calculate the icing parameters of the icing area.

4. The aircraft icing detection method according to claim 3, characterized in that, The icing parameters include maximum icing thickness, icing coverage area percentage, and ice type characteristics.

5. The aircraft icing detection method according to claim 3, characterized in that, The difference values ​​include grayscale difference values, color difference values, and texture difference values.

6. The aircraft icing detection method according to claim 4, characterized in that, The calculation of the icing level based on icing parameters includes: The corresponding icing level is calculated based on the maximum increase in icing thickness per minute.

7. The aircraft icing detection method according to claim 6, characterized in that, The icing levels include weak icing, medium icing, strong icing, and extremely strong icing. The maximum icing thickness increment per minute for weak icing is less than 0.6 mm per minute, for medium icing it is 0.6 to 1.0 mm per minute, for strong icing it is 1.1 to 2.0 mm per minute, and for extremely strong icing it is greater than 2.0 mm per minute.

8. The aircraft icing detection method according to claim 3, characterized in that, The preprocessing includes noise reduction and contrast enhancement.

9. The aircraft icing detection method according to claim 1, characterized in that, The deep learning algorithm uses a convolutional neural network model.

10. An aircraft icing detection system, used to implement the aircraft icing detection method as described in any one of claims 1-9, characterized in that, The system includes: The image library building module is used to build a data image library of different weather conditions in the aircraft icing monitoring area; The model training module is used to train an image recognition model based on a data library using deep learning algorithms. The data acquisition module is used to acquire real-time monitoring images of the aircraft icing monitoring area; The judgment module is used to input the real-time monitoring screen into the image recognition model for processing and analysis, to determine whether icing has occurred in the current monitoring area, and to obtain icing parameters when icing occurs. The icing level calculation module is used to calculate the icing level based on icing parameters; The alarm module is used to trigger the alarm mechanism and display differentiated response suggestions based on the icing level.

Images