Civil aviation radio station obstacle intelligent detection method, device and equipment

Abstract

The invention provides a civil aviation radio station obstacle detection method, device and equipment. The method comprises the following steps: acquiring a reference image, continuously acquiring a plurality of real-time images according to a time sequence, the plurality of real-time images being a plurality of images reflecting the change of a monitoring area along with time; respectively comparing the plurality of real-time images with the reference image to identify a newly added foreign matter contour compared with the reference image in each real-time image; and according to the newly-added foreign matter contour, determining the continuity of the newly-added foreign matter contour in the time sequence so as to determine whether the entity corresponding to the newly-added foreign matter contour is a long-term obstacle. According to the method provided by the invention, false alarms are remarkably reduced, and the detection accuracy is improved. According to the invention, the method achieves the intelligent judgment of the properties of the obstacles, improves the detection accuracy, enhances the automation and intelligence levels of the system, and provides powerful support for the safety management of civil aviation radio stations.

Description

Technical Field

[0001] The present invention relates to the field of civil aviation radio technology, and particularly to a method, device and equipment for detecting obstacles in civil aviation radio stations. Background Art

[0002] Civil aviation radio stations, such as radar stations and navigation stations, as an indispensable part of the air traffic management system, their site selection and construction need to strictly comply with the technical specifications of radio transmission to ensure the stable transmission of radio signals and flight safety. In the initial stage of construction, a comprehensive obstacle assessment will be carried out for these stations. The assessment mainly includes the height relationship between the location of the station and the heights of surrounding buildings, trees and other obstacles to ensure that they meet the established radio transmission requirements. However, with the continuous advancement of urban construction, the environment around radio stations often changes significantly. For example, new buildings may start construction in a certain direction, or existing buildings may be expanded, which may pose potential threats to radio stations. If these newly emerged obstacles exceed a certain height limit, they may interfere with the transmission of radio signals, thus posing a hidden danger to flight safety.

[0003] In order to detect and handle these potential obstacles in a timely manner, there are already some related video surveillance technologies on the market. These technologies are usually based on image processing algorithms and can achieve motion detection and foreign object detection of objects in the monitored area. However, although these technologies perform well in general scenarios, they have obvious limitations in the specific application scenario of detecting obstacles in civil aviation radio stations. For example: existing video surveillance technologies often adopt general foreign object detection methods in motion detection and foreign object detection, resulting in their inability to accurately distinguish different types of obstacles. For example, temporarily erected buildings or temporarily erected large vehicles, although they also belong to the category of foreign objects, in the detection of obstacles in civil aviation radio stations, these temporary obstacles do not pose a real threat. However, existing technologies cannot effectively identify and eliminate the interference of these temporary obstacles. In addition, for long-term obstacles that are "fixed and have growth changes", existing detection methods cannot accurately distinguish them.

[0004] In summary, the existing methods for detecting obstacles in radio stations at present cannot accurately distinguish temporary obstacles and long-term obstacles, resulting in a high false alarm rate. Summary of the Invention

[0005] The present invention provides a method, device and equipment for detecting obstacles in civil aviation radio stations to solve the defect in the prior art that temporary obstacles and long-term obstacles cannot be accurately distinguished, thereby improving the accuracy and efficiency of obstacle detection.

[0006] The present invention provides a method for detecting obstacles in civil aviation radio stations, including:

[0007] Obtain a reference image, where the reference image is an image of the monitoring area without obstacles or with a known static obstacle state;

[0008] Continuously obtain multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the changes in the monitoring area over time;

[0009] Compare each of the multiple real-time images with the reference image to identify the outline of newly added foreign objects in each real-time image compared with the reference image;

[0010] According to the outline of the newly added foreign object, determine its persistence in the time series to determine whether the entity corresponding to the outline of the newly added foreign object is a long-term obstacle.

[0011] According to a method for detecting obstacles in a civil aviation radio station provided by the present invention, the analyzing the persistence of the outline of the newly added foreign object in the time series to determine whether there is an obstacle that has long existed in the monitoring area includes:

[0012] Statistically analyze the frequency and / or duration of the appearance of the outline of the newly added foreign object in the time series;

[0013] When the frequency and / or duration exceed a preset frequency and / or preset time, it is determined that the outline of the newly added foreign object is a long-term obstacle.

[0014] According to a method for detecting obstacles in a civil aviation radio station provided by the present invention, the comparing each of the multiple real-time images with the reference image to identify the outline of newly added foreign objects in each real-time image compared with the reference image includes:

[0015] Obtain the multiple real-time images and the reference image;

[0016] Preprocess each of the multiple real-time images and the reference image to obtain the processed multiple real-time images and reference image;

[0017] Perform a pixel-level subtraction operation on each processed real-time image and the processed reference image to correspondingly obtain multiple difference images;

[0018] Perform binary processing on each difference image with a preset threshold to correspondingly obtain multiple foreground image masks;

[0019] According to the multiple foreground image masks, correspondingly obtain multiple background image masks of the real-time images;

[0020] Process the multiple foreground image masks and multiple background image masks respectively to obtain the processed multiple foreground image masks and multiple background image masks;

[0021] Use the processed multiple foreground image masks and multiple background image masks as weights to perform weighted synthesis on the corresponding foreground images and background images, so as to obtain the synthesized image of each real-time image;

[0022] Compare the synthesized image of each real-time image with the processed reference image respectively to identify the new foreign object contours in each real-time image compared with the reference image, and mark the new foreign object contours. The marking content includes: the timestamp when the foreign object contour appears, the number of foreign object contours, the position and the area.

[0023] According to a civil aviation radio station obstacle detection method provided by the present invention, the preprocessing of the multiple real-time images and the reference image respectively to obtain the processed multiple real-time images and reference image includes:

[0024] Perform grayscale processing on the multiple real-time images and the reference image respectively to obtain multiple grayscale real-time images and a grayscale reference image correspondingly;

[0025] Divide the multiple grayscale real-time images and the grayscale reference image respectively according to at least three different sizes of grids to obtain different sizes of grids of each grayscale real-time image and different sizes of grids of the grayscale reference image correspondingly;

[0026] Use the contrast limiting threshold to perform CLAHE enhancement on each size of grid one by one to obtain the corresponding local enhanced image blocks of each size of grid;

[0027] Fuse the local enhanced image blocks corresponding to each size of grid according to their corresponding weights to obtain multiple fused real-time images and a fused reference image correspondingly;

[0028] Perform normalization processing on each fused real-time image and the fused reference image to obtain the normalized multiple real-time images and the normalized reference image;

[0029] Perform Gamma correction on the normalized multiple real-time images and the normalized reference image respectively to obtain the processed multiple real-time images and reference image correspondingly.

[0030] According to a civil aviation radio station obstacle detection method provided by the present invention, the performing Gamma correction on the normalized multiple real-time images and the normalized reference image respectively includes:

[0031] Divide each of the normalized real-time images and the reference images into grid regions of a predetermined size, respectively, to obtain multiple grid regions of each real-time image and multiple grid regions of the reference images;

[0032] Determine the average luminance value of each grid region, and dynamically calculate the Gamma value applicable to each grid region according to the average luminance value of each grid region;

[0033] Generate a corresponding Gamma lookup table for each of the grid regions, and perform Gamma correction on the pixel values of each grid region using the generated Gamma lookup table, so as to obtain the corrected grid regions;

[0034] Process the adjacent corrected grid regions with smooth transition to finally obtain the processed multiple real-time images and reference images.

[0035] According to a civil aviation radio station obstacle detection method provided by the present invention, the processing of the multiple foreground image masks and the multiple background image masks respectively to obtain the processed multiple foreground image masks and the multiple background image masks includes:

[0036] Perform an erosion operation first and then a dilation operation on each of the foreground image masks to obtain multiple processed foreground image masks correspondingly;

[0037] Perform a bilateral filtering operation on each of the background image masks to obtain multiple processed background image masks correspondingly.

[0038] According to a civil aviation radio station obstacle detection method provided by the present invention, it further includes: regularly updating the reference image, and detecting civil aviation radio station obstacles according to the updated reference image;

[0039] The determining the persistence of the newly added foreign object contour in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle includes:

[0040] Mark the information of the newly added foreign object contour, and determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle according to the marked information and the persistence of the information in the time series;

[0041] The marked information includes: the time, position, height, growth situation, and the impact on radio wave propagation when the foreign object contour appears.

[0042] The present invention also provides a civil aviation radio station obstacle detection device, including:

[0043] An acquisition unit for acquiring a reference image, where the reference image is an image of the monitoring area without obstacles or with known static obstacle states;

[0044] The acquisition unit is further configured to continuously acquire multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the changes of the monitoring area over time;

[0045] A comparison unit for comparing the multiple real-time images with the reference image respectively;

[0046] An identification unit for identifying the contour of a foreign object newly added in each real-time image compared with the reference image;

[0047] A determination unit for determining the persistence of the newly added foreign object contour in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle.

[0048] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the civil aviation radio station obstacle detection method as described in any one of the above.

[0049] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the civil aviation radio station obstacle detection method as described in any one of the above.

[0050] The civil aviation radio station obstacle detection method, device, and equipment provided by the present invention can accurately identify and exclude temporary obstacles, such as temporarily built buildings or large vehicles, by introducing time series analysis, thereby significantly reducing false alarms and improving the detection accuracy. Moreover, it realizes the intelligent judgment of the nature of obstacles. This intelligent detection method not only improves the detection accuracy but also enhances the automation and intelligent level of the system, providing strong support for the safety management of civil aviation radio stations. Description of the Drawings

[0051] Figure 1 It is a schematic flowchart of the civil aviation radio station obstacle detection method provided by the present invention;

[0052] Figure 2 It is a schematic diagram of the reference image taken in the embodiment of the present invention;

[0053] Figure 3 It is a schematic diagram of the real-time image taken in the embodiment of the invention;

[0054] Figure 4 It is a schematic flowchart of the method for identifying the foreign object contour in the embodiment of the present invention;

[0055] Figure 5 Schematic flow chart of the method for preprocessing real-time image and reference image in an embodiment of the present invention;

[0056] Figure 6 Schematic flow chart of the method for performing Gamma correction on real-time image and reference image in an embodiment of the present invention;

[0057] Figure 7 Block diagram of the civil aviation radio station obstacle detection device provided in an embodiment of the present invention;

[0058] Figure 8 Schematic structural diagram of the electronic device provided by the present invention. Detailed implementation manners

[0059] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0060] The present invention performs intelligent detection on obstacles in the surrounding environment of civil aviation radio stations (such as radar stations, navigation stations, etc.). The aim is to monitor the dynamic changes in the surrounding environment of radio stations in real time or regularly, and timely discover and give early warnings of newly added or growing fixed obstacles to ensure the normal operation of radio stations and flight safety.

[0061] The detection method provided by the present invention captures a real-time picture of the external station of the radio station every day at a fixed time, compares it with the reference picture respectively, and finds out whether there are differences in the contours. If there are differences, the difference results are saved, so as to detect and record the daily environmental changes of the external station, including what objects are added, what objects are missing, and on which day and at which position the additions and deletions occur, all of which are recorded. For temporary obstacles, such as an antenna is temporarily erected on a certain building on the third day. Then, in the recorded pictures, there are no differences between the first day and the second day and the reference picture, and differences start to appear on the third day, that is, the contour of the temporary antenna is added. The difference results will show on which day the temporary antenna starts to be erected and at which position it is erected. If the temporary antenna is removed on the fourth day, it indicates that the antenna is temporarily built and is a temporary obstacle. For long-term obstacles, such as a certain building within the monitoring range, a building is built on the roof. Similarly, the daily changing contour of the building is recorded. Starting from the day when the anomaly is first discovered, the change situation of the building is continuously monitored. If the height of the building exceeds the preset height, corresponding measures should be taken in time to prevent affecting the communication of civil aviation radio stations.

[0062] Figure 1 Schematic diagram of the obstacle detection method for civil aviation radio stations provided by the present invention, as Figure 1 shown, the method includes the following steps:

[0063] Step 1: Obtain a reference image, which is an image of the monitoring area without obstacles or with known static obstacle states.

[0064] Specifically, the so-called reference image refers to a high-definition image of the monitoring area without obstacles or only containing known static obstacle states. The state of "without obstacles" here refers to the initial environmental state that meets the radio emission requirements after obstacle assessment at the initial stage of the station construction; while the "known static obstacles" include buildings, terrain features, etc. that already existed at the initial stage of construction and will not change (or the change can be ignored) subsequently.

[0065] Step 2: Continuously obtain multiple real-time images in time series, which are multiple images reflecting the changes of the monitoring area over time.

[0066] Specifically, deploy high-definition cameras at key positions around the civil aviation radio station to ensure that the areas where potential obstacles may appear can be fully covered. And use high-definition cameras to capture images to ensure that the images are clear and can accurately reflect the detailed features of the monitoring area, including the shape, size, position, etc. of the obstacles. Then configure the relevant information of the camera (the IP address of the camera, the username and password for logging in to the camera), and use the relevant information of the camera to construct the URL of the camera. Finally, start the corresponding camera, and continuously obtain multiple real-time images in time series through the camera. In the embodiment of the present invention, the so-called continuously obtaining multiple real-time images in time series means capturing real-time images through the camera at a fixed time point every day (such as in the early morning, evening, etc. when the light conditions are relatively stable), and saving the captured real-time images, thereby forming an image set in time series. Among them, each captured real-time image is marked with the capture time.

[0067] Step 3: Compare each of the multiple real-time images with the reference image to identify the new foreign object contours in each real-time image compared with the reference image.

[0068] Specifically, arrange the real-time images captured every day in chronological order, and then compare each real-time image with the reference image respectively to identify whether there are new foreign object contours. For example, Figure 2 Schematic diagram of the reference image taken in the embodiment of the present invention, Figure 3 is a schematic diagram of the real-time image taken in the embodiment of the invention. Compare Figure 2 with Figure 3 If inFigure 3 If a foreign object contour appears in the square, mark the foreign object contour.

[0069] Step 4: According to the newly added foreign object contour, determine its persistence in the time series to determine whether the newly added foreign object contour is a long-term obstacle, that is, whether there is an obstacle that has long existed in the monitoring area.

[0070] Specifically, taking the reference image as the standard, compare the real-time image taken on the second day with the reference image. If there is a newly added foreign object contour, mark the foreign object contour; then compare the real-time image taken on the third day with the reference image. If the foreign object contour disappears, it means that the foreign object contour detected on the second day appears temporarily, that is, a temporary obstacle. If in the following several days, by comparing the taken real-time images with the reference image, the foreign object contour always exists, then the entity corresponding to the foreign object contour is a long-term obstacle.

[0071] In addition, information marking needs to be performed on the identified foreign object contour, which may specifically include key information such as the time, location, height, growth situation, and possible impact on radio wave propagation of the foreign object contour. Thus, it provides decision-making support for the civil aviation management department, promptly stops relevant construction behaviors, and ensures the normal operation of radio stations and flight safety.

[0072] The method provided by the present invention can accurately identify and eliminate temporary obstacles, such as temporarily erected buildings or large vehicles, by introducing time series analysis, thereby significantly reducing false alarms and improving the accuracy of detection. Moreover, it realizes the intelligent judgment of the nature of obstacles. This intelligent detection method not only improves the accuracy of detection, but also enhances the automation and intelligence level of the system, providing strong support for the safety management of civil aviation radio stations.

[0073] Furthermore, the identification of the newly added foreign object contour by the present invention further includes: setting a minimum contour area threshold, detecting the contour, using a square to mark the difference area, and saving the reference image and the real-time image with the square.

[0074] Specifically, the contour drawing is based on the processing of contour detection results and is achieved by calculating the contour area and drawing the minimum bounding rectangle containing the contour. The contour detection algorithm can identify the significant shapes or region boundaries in the image, while the area threshold filtering is used to remove the noise contours with small areas. Drawing the rectangle provides an intuitive way to display the detected difference regions, facilitating the user to quickly locate and analyze. Each contour is described by a set of coordinate points, representing a possible obstacle or change region. The area of each contour is calculated, and the noise contours with small areas are filtered according to the preset area threshold, thus avoiding marking meaningless small differences (such as the swaying of leaves, light and shadow changes, etc.). For each contour filtered by the area threshold, the minimum bounding rectangle containing the contour is calculated to draw the bounding rectangle for the detected contour. The upper left corner coordinates (x, y) and the width and height (w, h) of the rectangle are used to draw the rectangle frame on the image, as shown in formula (1).

[0075] R(x,y,w,h)=rectangle(x,y,x+w,y+h) (1)

[0076] Where R(x, y, w, h) is the drawn bounding rectangle.

[0077] In addition, the present invention can draw the rectangle frame in red color to visually display the detected difference regions. Red is usually used to represent warnings or important information, which helps the user to quickly locate the differences. The present invention also marks the difference regions, specifically including the timestamp, the number of difference regions, the location and the area, and then saves the reference image with the marked differences for comparison with the real-time images collected subsequently. By saving the real-time images with the marked differences, it helps to track the historical records of the changes and provides a visual reference for further investigation or processing. Specifically, by setting the minimum contour area threshold, small difference regions can be filtered out, reducing false alarms. For example, small area change regions such as the swaying of leaves and light and shadow changes will not be marked as valid difference regions if their areas are less than the set threshold; in the obstacle detection of civil aviation radio stations, this threshold can not affect the small area changes that do not interfere with radio signals, such as small stones and tree branches on the ground, so as to focus on detecting large area obstacles that may interfere with the signals.

[0078] In addition, the present invention only detects the outer contours. By only retaining the end points of the contours, the contour representation can be simplified and the calculation amount can be reduced. Specifically, through contour detection, the significant shapes or region contours in the binary image can be identified and extracted. These contours represent possible obstacles. By only focusing on the external shape of the target region and without dealing with the internal details, it helps to more clearly identify the obstacles.

[0079] The method provided by the present invention can effectively achieve the automatic detection of obstacles in civil aviation radio stations by setting the minimum contour area threshold and performing contour detection. By filtering small-area change regions and extracting the contours of significant shapes or regions, it is possible to focus on detecting large-area obstacles that may interfere with radio signal transmission, thereby promptly stopping relevant construction behaviors and avoiding greater losses after construction.

[0080] In addition, the present invention provides strong support for the automatic detection of obstacles in civil aviation radio stations through contour drawing, visually displaying the difference regions, assisting in analysis and verification, and filtering small-region noise. By saving the reference image and the current image with red boxes, the system can record the historical record of changes and provide a visual reference for subsequent analysis and processing. This helps the civil aviation department promptly stop construction behaviors that violate regulations and ensure compliance with the environment around radio stations.

[0081] Furthermore, the method for detecting obstacles in civil aviation radio stations provided by the embodiments of the present invention further includes: statistically analyzing the frequency and / or duration of the newly added foreign object contours in the time series; when the frequency and / or duration exceed the preset frequency and / or preset time, it is determined that the newly added foreign object contour is a long-term obstacle.

[0082] Specifically, since temporary obstacles, such as temporarily erected buildings or vehicles, usually appear at a low frequency or for a short duration in the time series. In contrast, long-term obstacles, such as permanent buildings or facilities, will continuously appear in the time series with a high frequency and a long duration. Existing methods often cannot accurately distinguish between temporary and long-term obstacles, resulting in a high false alarm rate. Moreover, for temporary obstacles with a long duration, if their frequency and duration exceed the preset threshold, they can be correctly identified as long-term obstacles, thereby reducing missed detections.

[0083] The method provided by the present invention can more quickly identify long-term obstacles through automatic statistical analysis of frequency and duration, reduce the need for manual verification, and improve the detection efficiency. Moreover, by adjusting the thresholds of the preset frequency and preset time, it is possible to adapt to the foreign object detection requirements in different environments and scenarios, thereby improving the flexibility of detection.

[0084] Furthermore, the method for detecting obstacles in civil aviation radio stations provided by the embodiments of the present invention introduces how to compare multiple real-time images with the reference image respectively to identify the newly added foreign object contours in each real-time image compared with the reference image. Figure 4 It is a schematic flowchart of the method for identifying foreign object contours in the embodiments of the present invention, as Figure 4 shown. The method specifically includes the following steps:

[0085] Step 401: Obtain multiple real-time images and a reference image;

[0086] Step 402: Perform preprocessing on the multiple real-time images and the reference image respectively to obtain the processed multiple real-time images and the reference image;

[0087] Step 403: Perform a pixel-level subtraction operation on each processed real-time image and the processed reference image to correspondingly obtain multiple difference images;

[0088] Step 404: Perform binarization processing on each difference image with a preset threshold to correspondingly obtain multiple foreground image masks;

[0089] Step 405: According to the multiple foreground image masks, correspondingly obtain multiple background image masks of the real-time image;

[0090] Step 406: Process the multiple foreground image masks and the multiple background image masks respectively to obtain the processed multiple foreground image masks and the multiple background image masks;

[0091] Step 407: Use the processed multiple foreground image masks and the multiple background image masks as weights to perform weighted synthesis on the corresponding foreground image and background image, so as to obtain the synthesized image of each real-time image;

[0092] Step 408: Compare the synthesized image of each real-time image with the processed reference image respectively to identify the new foreign object contours in each real-time image compared with the reference image, and mark the new foreign object contours. The marking content includes: the timestamp when the foreign object contour appears, the number of foreign object contours, the position and the area.

[0093] Specifically, the present invention first performs a pixel difference operation on the reference image and the real-time image through the absolute difference calculation method. The calculation formula is as follows:

[0094] D(x,y) = |I baseline (x,y) - I current (x,y)| (2)

[0095] Wherein, D(x,y) is the difference image, I baseline (x,y) is the pixel corresponding to the reference image; I current (x,y) is the pixel corresponding to the real-time image.

[0096] Then, perform binarization processing on the difference image to highlight the significant difference area. The calculation formula is as follows:

[0097]

[0098] Wherein, T diffis the difference threshold, which is dynamically adjusted according to the ambient light and noise level.

[0099] When monitoring the surrounding of the station, dynamic interference factors (such as the swaying of trees and grass, temporarily erected antennas, parking lots, people walking, passing vehicles, etc.) are not obstacles. Before identifying the outline of newly added foreign objects, it is necessary to remove them as interference. At the same time, static interference factors (such as the sky, ground, etc.) around the station, due to light changes, etc., are regarded as background noise and cannot be effectively removed, thus interfering with the detection of obstacles. For this reason, the embodiment of the present invention provides a method capable of effectively removing interference factors, that is, hierarchical fuzzy. Through hierarchical fuzzy, the outline and changes of dynamic interference factors can be clearly detected; at the same time, the noise caused by static interference factors can also be reduced, thereby improving the detection accuracy and reducing the false alarm rate. Specifically, by respectively obtaining the foreground image mask and the background image mask of the real-time image, and then processing the foreground image mask to blur the dynamic interference factors; processing the background image mask to remove the background noise, making the background cleaner and reducing false alarms caused by small noises.

[0100] The civil aviation radio station obstacle detection method provided by the present invention, on the one hand, by respectively obtaining the foreground image mask and the background image mask of the real-time image and performing corresponding processing on them, the outline and changes of dynamic interference factors can be clearly detected, thereby effectively distinguishing these interferences from real obstacles; thus effectively improving the detection accuracy and reducing the false alarm rate. On the other hand, whether it is dynamic interference or static interference, effective processing can be carried out, so as to be able to cope with various complex monitoring environments, making the detection method provided by the present invention adaptable to the monitoring requirements at different time periods and different weather conditions.

[0101] Furthermore, the following introduces how to preprocess multiple real-time images and the reference image to obtain the processed real-time images and reference images. Figure 5 is a schematic flow chart of the method for preprocessing the real-time image and the reference image in the embodiment of the present invention, as Figure 5 shown, the method specifically includes the following steps:

[0102] Step 501: Perform gray-scale processing on multiple real-time images and the reference image respectively to obtain multiple gray-scale real-time images and a gray-scale reference image correspondingly;

[0103] Step 502: Divide the multiple gray-scale real-time images and the gray-scale reference image respectively according to at least three different sizes of grids to obtain different sizes of grids for each gray-scale real-time image and different sizes of grids for the gray-scale reference image correspondingly;

[0104] Step 503: Using the contrast limit threshold, perform CLAHE enhancement on each size of grid one by one, and correspondingly obtain local enhanced image patches corresponding to each size of grid;

[0105] Step 504: Fuse the local enhanced image patches corresponding to each size of grid according to their corresponding weights, and correspondingly obtain multiple fused real-time images and a fused reference image;

[0106] Step 505: Perform row normalization on each of the fused real-time images and the fused reference image to obtain multiple normalized real-time images and a normalized reference image;

[0107] Step 506: Perform Gamma correction on each of the normalized multiple real-time images and the normalized reference image, and correspondingly obtain the processed multiple real-time images and the reference image.

[0108] Specifically, first, uniformly convert each real-time image and the reference image into grayscale images; then, use large, medium, and small grids to perform CLAHE enhancement on the grayscale images respectively, so as to obtain 3 different grid division results for the grayscale images. In the embodiment of the present invention, small grids (8×8), medium grids (16×16), and large grids (32×32) are used to perform CLAHE enhancement on the grayscale images. Among them, small grids are used to enhance the detail contrast. Medium grids are used to balance detail and global enhancement. Large grids are used to improve the overall contrast. During the process of performing CLAHE enhancement, by setting the contrast limit threshold (clip_limit = 2.2), it is possible to prevent artifacts or noise amplification caused by over-enhancement. Finally, by defining the weight ratio of the processing results of each grid, the results of different scales are fused. In the embodiment of the present invention, the weights corresponding to the grid sizes (such as 8, 16, 32) are: 0.4, 0.4, 0.2. That is, the images processed by each grid size are accumulated into the fusion result array according to the weights, and the final image calculation formula is as shown in formula (4), and the pixel values in the fusion result array are limited between 0 and 255.

[0109]

[0110] Among them, i is the number of grid sizes. In the embodiment of the present invention, i is 3, that is, corresponding to small grids (8×8), medium grids (16×16), and large grids (32×32).

[0111] Since the fused image may contain pixel values outside the normal pixel value range (0 - 255), because the enhancement effects of different grids may be superimposed on each other during fusion, resulting in some pixel values being too high or too low. To avoid this situation, it is necessary to normalize the fused image to ensure that all pixel values fall within the legal range (0 - 255). The specific method of normalization is usually to linearly map the pixel values of the fused image to the range of 0 - 255. In this way, the processed image not only retains the effect of multi-scale CLAHE enhancement but also ensures the legality of pixel values.

[0112] It should be noted that: the weights of the present invention can be adjusted according to actual needs, such as [0.4, 0.4, 0.2]; different grid sizes can also be adjusted according to the actual situation.

[0113] The present invention uses multi-scale CLAHE to process graphics and has the following advantages:

[0114] On the one hand, by using multi-scale CLAHE, which combines various grid sizes, different-sized grids can perform more detailed and comprehensive local histogram equalization processing on the image. Thus, while retaining the local details of the image, it can better balance the global contrast and avoid the problems of local detail loss or global contrast imbalance that may be caused by a single grid size. On the other hand, due to a single grid, especially a small grid, it is easy to over-enhance local details when processing an image, resulting in the amplification of image artifacts or noise. However, in this application, by assigning weights to the processing results of each grid through multi-scale CLAHE and integrating the results of small grids and large grids, this method can fuse the enhancement effects under different grid sizes, thereby effectively reducing the amplification of artifacts and noise. On the other hand, since ordinary CLAHE uses the same enhancement parameters throughout the image and cannot be adaptively adjusted according to the characteristics of different regions of the image. However, in this application, by using multi-scale CLAHE, the image can be processed under different grid sizes, thereby dynamically adapting to the enhancement requirements of different regions of the image. This method can more accurately identify and enhance the key information in the image while avoiding over-processing of unimportant regions; on the other hand, in regions with extremely low contrast, the processing method of ordinary single-grid CLAHE may over-enhance the contrast, resulting in the amplification of artifacts and noise. However, in this application, the clipLimit parameter is set to limit the over-enhancement of local contrast. At the same time, the pixel values are normalized in the final result to ensure that the overall brightness and contrast of the image are maintained within a reasonable range. This limitation and normalization processing help prevent over-enhancement and distortion of the image.

[0115] By processing the image using multi-scale CLAHE, the contrast between obstacles and the background can be enhanced, especially the subtle changes in low-contrast areas, making the obstacle contours more clearly distinguishable. Moreover, the recognition accuracy for fixed and growing obstacles is improved, and the false alarm rate is reduced; furthermore, by comprehensively considering image features at different scales, multi-scale CLAHE can reduce the artifacts and noise amplification problems that may occur in the traditional CLAHE algorithm, making the enhanced image clearer and more natural, and further improving the recognition accuracy for obstacles.

[0116] For the civil aviation radio station obstacle detection method provided by the present invention, on the one hand, by uniformly converting the real-time image and the reference image into grayscale images, the computational complexity is reduced, only the luminance information is retained, and the efficiency of subsequent difference detection is improved; on the other hand, by applying contrast-limited adaptive histogram equalization (CLAHE) to the grayscale image, the contrast of local regions is enhanced, the visibility of image details is improved, especially the subtle changes in low-contrast regions, thereby improving the recognition accuracy of foreign object contours. On the other hand, by adopting the multi-scale CLAHE enhancement technology, the visual effect and detail retention ability of the image can be significantly improved, while reducing the amplification of artifacts and noise, further improving the recognition accuracy for obstacles and reducing the false alarm rate.

[0117] Furthermore, the following will explain in detail how to perform Gamma correction on the normalized multiple real-time images and the normalized reference image respectively. Figure 6 It is a schematic flowchart of the method for performing Gamma correction on the real-time image and the reference image in the embodiment of the present invention, as Figure 6 shown, which specifically includes the following steps:

[0118] Step 601: Divide each normalized real-time image and reference image into grid regions of a predetermined size, respectively obtaining multiple grid regions of each real-time image and multiple grid regions of the reference image;

[0119] Step 602: Determine the average luminance value of each grid region, and dynamically calculate the Gamma value applicable to each grid region according to the average luminance value of each grid region;

[0120] Step 603: Generate a corresponding Gamma lookup table for each such grid region, and use the generated Gamma lookup table to perform Gamma correction on the pixel values of each grid region, thereby obtaining the corrected grid region;

[0121] Step 604: Process the adjacent corrected grid regions with smooth transition to finally obtain the processed multiple real-time images and reference images.

[0122] Specifically, first, the normalized image is divided into multiple small regions according to the set grid size. Assuming the image size is W×H and the grid size is gW×gH, then it can be divided into small regions. Calculate the average brightness value for each small region, that is: sum the brightness values of all pixels in each small region and then divide by the number of pixels. Then, according to the average brightness value of each region, dynamically determine the appropriate Gamma value using formula (5). For example: if the average brightness is low (dark part), set the Gamma value greater than 1. If the average brightness is high (highlight), set the Gamma value less than 1. Next, generate an independent Gamma correction lookup table for each region. The size of the lookup table can be selected as needed, and each item k of the lookup table can be calculated according to the Gamma value, as shown in formula (6) specifically. The current grid generates a lookup table containing 256 elements to map the original pixel values to the corrected pixel values. Then, use formula (7) to dynamically calculate the Gamma value of each corrected grid region to obtain the dynamically adjusted grid region. Finally, perform a smooth transition on the adjacent grid regions after dynamic adjustment to finally obtain multiple processed real-time images and a reference image.

[0123]

[0124] where γ is the Gamma value.

[0125]

[0126] γ = gamma_base - (mean_brightness / 255 × empirical weight) (7)

[0127] where gamma_base is the base Gamma value (in the implementation of the present invention, it is 2.2), mean_brightness is the average brightness of the current local region; the empirical weight controls the adjustment range of the Gamma value (in the implementation of the present invention, it is 0.8).

[0128] By adopting dynamic Gamma correction, the present invention can adaptively adjust the Gamma value according to the average brightness of each grid (or local area). This means that regions with different brightness levels will be processed differently, thus avoiding the problems of insufficient processing in dark areas or over - processing in bright areas caused by a fixed Gamma value. In addition, since the Gamma value is dynamically calculated based on the local brightness, the details in the dark areas can be better preserved, and at the same time, the bright areas will not lose information due to over - exposure. By dividing the image into small grids, the brightness distribution within each grid is relatively uniform, enabling more accurate calculation of the Gamma value for each grid. This helps to achieve a better correction effect in scenarios with uneven brightness distribution. Moreover, the division of the grids can be adjusted according to actual needs. Smaller grids can provide more refined correction, while larger grids may reduce the computational amount and improve the processing speed. Independent processing of each grid means that the correction strategy can flexibly adapt to different brightness distributions. In scenarios with drastic brightness changes, this strategy can significantly reduce the problems of insufficient or excessive processing. Since each grid is optimized according to its brightness distribution, the visual effect of the entire image will be more natural and uniform.

[0129] The civil aviation radio station obstacle detection method provided by the present invention enables real - time adjustment of image parameters through dynamic Gamma correction, thereby accurately capturing minute changes in obstacles (such as the start of construction of a new building). Moreover, by updating the Gamma value of the image in real - time, high - sensitivity monitoring of obstacles can be maintained, and potential problems can be detected and reported in a timely manner. Therefore, by adjusting the image brightness and contrast through dynamic Gamma correction, the present invention can more accurately distinguish obstacles from the background, thereby reducing false alarms and missed detections.

[0130] The processing methods for the foreground image mask and multiple background image masks are introduced in detail below, specifically including: performing an erosion operation followed by a dilation operation on each foreground image mask to obtain multiple processed foreground image masks; performing a bilateral filtering operation on each background image mask to obtain multiple processed background image masks.

[0131] Specifically, the processing methods include the following steps:

[0132] Step 1: Read the reference image and the current real - time image, that is, the real - time image and the reference image after Gamma correction.

[0133] Step 2: Perform a pixel - level subtraction operation on the reference image and the current real - time image to obtain a difference image.

[0134] Step 3: Apply binarization to the difference image (set the threshold to 40), set the pixels with a difference greater than 40 as the foreground (white, pixel value 255), and the rest as the background (black, pixel value 0), so as to obtain the foreground image mask.

[0135] Step 4: Apply morphological opening (erosion first and then dilation) to the foreground image mask, and use a 3×3 rectangular kernel to eliminate small-area noise points;

[0136] Step 5: Perform an inversion operation on the foreground image mask to obtain the background image mask.

[0137] Step 6: Use the foreground image mask obtained in Step 3 to extract the foreground part from the current image in Step 1 to obtain the foreground image;

[0138] Step 7: Apply bilateral filtering to the extracted foreground part to smooth the interior of the region while maintaining the edges.

[0139] Step 8: Use the background image mask obtained in Step 5 to extract the background part from the current image in Step 1.

[0140] Step 9: Apply median filtering to the extracted background part to remove unnecessary noise and obtain the background image;

[0141] Step 10: Use the foreground image mask obtained in Step 3 and the background image mask obtained in Step 5 as weights to perform weighted synthesis on the foreground image obtained in Step 6 and the background image obtained in Step 9 to obtain the synthesized image of the real-time image.

[0142] Through the application of morphological opening, the present invention can effectively remove some noise in the foreground image mask, such as: the influence caused by the movement of grass in the wind, changes in light, weather effects (such as raindrops, snowflakes), and noise generated during camera shooting. Applying bilateral filtering to the extracted foreground part makes the pixel value changes in the background image mask smoother, reducing false alarms or missed detections caused by environmental changes.

[0143] The civil aviation radio station obstacle detection method provided by the present invention significantly improves the recognition accuracy of obstacles and reduces the false alarm rate by performing erosion operation followed by dilation operation on the foreground image mask and bilateral filtering operation on the background image mask.

[0144] Further, the method provided by the present invention further includes: regularly updating the reference image, and detecting civil aviation radio station obstacles based on the updated reference image.

[0145] Specifically, since the environment around civil aviation radio stations is dynamically changing, such as the construction of new buildings, the growth of trees, and minor changes in topography. These changes may cause the environment around the station that originally met the radio emission requirements to no longer meet the standards, which may interfere with the transmission of radio signals. By regularly updating the reference image in this application, when new obstacles appear, they can be detected in a timely manner and an alarm can be issued to prevent greater losses after construction. If the reference image is outdated or inaccurate, false alarms or missed alarms may occur during the differential detection. Regularly updating the reference image can ensure that the image data on which the system is based is accurate, thereby improving the accuracy of monitoring. Therefore, by comparing the updated reference image with the real-time acquired image in this application, newly added or changed obstacles can be identified more accurately, reducing the occurrence of false alarms and missed alarms.

[0146] Furthermore, information is marked on the newly added foreign object contour, and based on the marked information and the persistence of this information in the time series, it is determined whether the entity corresponding to the newly added foreign object contour is a long-term obstacle; the marked information includes: the time, location, height, growth situation, and the impact on radio wave propagation when the foreign object contour appears.

[0147] Specifically, the method provided by the present invention can automatically record the time, location of each monitoring, and the change information detected, and archive these information together with the corresponding image data. This structured data management method makes subsequent analysis efficient and reliable. Managers can quickly locate the monitoring records for a specific time period or location, providing support for decision-making. When an obstacle problem is found, managers can retrieve historical monitoring data, analyze the growth trend and potential impact of the obstacle, and thus take more precise countermeasures.

[0148] In summary, the method provided by the present invention effectively solves the problems of large errors and low efficiency existing in the traditional manual monitoring mode, and also provides a new idea and practical example for the intelligent upgrade in the field of civil aviation radio station management. Specifically:

[0149] 1. It realizes automatic obstacle detection, greatly improving the monitoring efficiency and accuracy.

[0150] 2. By combining intelligent archiving and traceability functions, it provides powerful data support and analysis tools for managers, helping to achieve more intelligent management decisions.

[0151] 3. It can operate stably in various complex scenarios, providing a strong guarantee for the safe operation of civil aviation radio stations.

[0152] The civil aviation radio station obstacle detection device provided by the present invention will be described below. The civil aviation radio station obstacle detection device described below can be correspondingly referred to the civil aviation radio station obstacle detection method described above.

[0153] Figure 7 It is a structural block diagram of the civil aviation radio station obstacle detection device provided by an embodiment of the present invention. As Figure 7 shown, the device includes:

[0154] An acquisition unit 701, configured to acquire a reference image, where the reference image is an image of the monitoring area without obstacles or in a state with known static obstacles;

[0155] The acquisition unit 702 is further configured to continuously acquire multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the change of the monitoring area over time;

[0156] A comparison unit 703, configured to compare each of the multiple real-time images with the reference image;

[0157] An identification unit 704, configured to identify the contour of a foreign object newly added in each real-time image compared with the reference image;

[0158] A determination unit 705, configured to determine the persistence of the newly added foreign object contour in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle.

[0159] Figure 8 It exemplifies a schematic diagram of the physical structure of an electronic device. As Figure 8 shown, the electronic device may include: a processor 810, a communication interface 520, a memory 830, and a communication bus 840. Among them, the processor 810, the communication interface 820, and the memory 830 complete communication with each other through the communication bus 840. The processor 810 can call the logical instructions in the memory 830 to execute the civil aviation radio station obstacle detection method, and the method includes:

[0160] Acquire a reference image, where the reference image is an image of the monitoring area without obstacles or in a state with known static obstacles;

[0161] Continuously acquire multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the change of the monitoring area over time;

[0162] Compare each of the multiple real-time images with the reference image to identify the contour of a foreign object newly added in each real-time image compared with the reference image;

[0163] Based on the newly added foreign object contour, determine its persistence in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle.

[0164] In addition, when the logical instructions in the above-mentioned memory 830 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0165] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the civil aviation radio station obstacle detection method, and the method includes:

[0166] Obtain a reference image, where the reference image is an image of the monitoring area without obstacles or in a state with known static obstacles;

[0167] Continuously obtain multiple real-time images according to the time series. The multiple real-time images are multiple images obtained by reflecting the change of the monitoring area over time;

[0168] Compare each of the multiple real-time images with the reference image to identify the newly added foreign object contours in each real-time image compared with the reference image;

[0169] Based on the newly added foreign object contour, determine its persistence in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle.

[0170] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the civil aviation radio station obstacle detection method, and the method includes:

[0171] Obtain a reference image, where the reference image is an image of the monitoring area without obstacles or in a state with known static obstacles;

[0172] Continuously obtain multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the changes in the monitored area over time;

[0173] Compare each of the multiple real-time images with the reference image to identify the outlines of foreign objects newly added in each real-time image compared with the reference image;

[0174] Based on the outlines of the newly added foreign objects, determine their persistence in the time series to determine whether the entity corresponding to the outlines of the newly added foreign objects is a long-term obstacle.

[0175] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0176] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute each embodiment or some parts of the method of the embodiment.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting obstacles of a civil aviation radio station, characterized in that Including: Obtain a reference image, which is an image of the monitored area without obstacles or with known static obstacle states; Continuously obtain multiple real-time images in time series, where the multiple real-time images are multiple images obtained by reflecting the changes of the monitored area over time; Compare the multiple real-time images with the reference image respectively to identify the foreign object contours newly added in each real-time image compared with the reference image; According to the newly added foreign object contours, determine their persistence in the time series to determine whether the entities corresponding to the newly added foreign object contours are long-term obstacles.

2. The civil aviation radio station obstacle detection method according to claim 1, wherein The analyzing the persistence of the newly added foreign object contours in the time series to determine whether there are obstacles that exist in the monitored area for a long time includes: Count the frequency and / or duration of the newly added foreign object contours appearing in the time series; When the frequency and / or duration exceed the preset frequency and / or preset time, determine that the newly added foreign object contour is a long-term obstacle.

3. The civil aviation radio station obstacle detection method according to claim 1 or 2, characterized in that, The comparing the multiple real-time images with the reference image respectively to identify the foreign object contours newly added in each real-time image compared with the reference image includes: Obtain the multiple real-time images and the reference image; Preprocess the multiple real-time images and the reference image respectively to obtain the processed multiple real-time images and reference image; Perform a pixel-level subtraction operation on each processed real-time image and the processed reference image to correspondingly obtain multiple difference images; Perform binary processing on each difference image with a preset threshold to correspondingly obtain multiple foreground image masks; According to the multiple foreground image masks, correspondingly obtain multiple background image masks of the real-time images; Process the multiple foreground image masks and multiple background image masks respectively to obtain the processed multiple foreground image masks and multiple background image masks; Use the processed multiple foreground image masks and multiple background image masks as weights to perform weighted synthesis on the corresponding foreground images and background images, so as to obtain the synthesized image of each real-time image; Compare the synthesized image of each real-time image with the processed reference image respectively to identify the foreign object contours newly added in each real-time image compared with the reference image, and mark the newly added foreign object contours, and the marking content includes: the timestamp when the foreign object contour appears, the number, position and area of the foreign object contour.

4. The civil aviation radio station obstacle detection method according to claim 3, wherein The preprocessing the multiple real-time images and the reference image respectively to obtain the processed multiple real-time images and reference image includes: Perform gray-scale processing on the multiple real-time images and the reference image respectively to correspondingly obtain multiple gray-scale real-time images and gray-scale reference images; Divide the multiple gray-scale real-time images and the gray-scale reference image into grids of at least three different sizes respectively to correspondingly obtain different-size grids of each gray-scale real-time image and different-size grids of the gray-scale reference image; Use the contrast-limiting threshold to perform CLAHE enhancement on each size grid one by one to correspondingly obtain local enhanced image blocks corresponding to each size grid; Fuse the local enhanced image patches corresponding to each size grid according to their corresponding weights to obtain multiple fused real-time images and a fused reference image respectively; Perform row normalization on each of the fused real-time images and the fused reference image to obtain multiple normalized real-time images and a normalized reference image; Perform Gamma correction on the multiple normalized real-time images and the normalized reference image respectively to obtain the processed multiple real-time images and the reference image; 5. The civil aviation radio station obstacle detection method according to claim 4, wherein The performing Gamma correction on the multiple normalized real-time images and the normalized reference image respectively includes: Divide each of the normalized real-time images and the reference image into grid regions of a predetermined size to obtain multiple grid regions of each real-time image and multiple grid regions of the reference image; Determine the average brightness value of each grid region, and dynamically calculate the Gamma value applicable to each grid region according to the average brightness value of each grid region; Generate a corresponding Gamma lookup table for each of the grid regions, and use the generated Gamma lookup table to perform Gamma correction on the pixel values of each grid region, so as to obtain the corrected grid region; Process the adjacent corrected grid regions with smooth transition to finally obtain the processed multiple real-time images and the reference image; 6. The civil aviation radio station obstacle detection method according to claim 3, wherein The processing the multiple foreground image masks and the multiple background image masks respectively to obtain the processed multiple foreground image masks and the multiple background image masks includes: Perform an erosion operation and then a dilation operation on each of the foreground image masks to obtain multiple processed foreground image masks; Perform a bilateral filtering operation on each of the background image masks to obtain multiple processed background image masks; 7. The civil aviation radio station obstacle detection method according to claim 3, characterized in that, further comprising: regularly updating the reference image, and detecting civil aviation radio station obstacles according to the updated reference image; The determining the persistence of the newly added foreign object contour in the time series to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle includes: Perform information marking on the newly added foreign object contour, and determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle according to the marked information and the persistence of the information in the time series; The marked information includes: the time, position, height, growth condition, and influence on radio wave propagation when the foreign object contour appears; 8. An obstacle detection device for a civil aviation radio station, characterized in that, Comprising: An acquisition unit for acquiring a reference image, where the reference image is an image of the monitoring area in a state without obstacles or with known static obstacles; The acquisition unit is further configured to continuously acquire multiple real-time images according to a time series, where the multiple real-time images are multiple images obtained by reflecting the change of the monitoring area over time; A comparison unit for comparing each of the multiple real-time images with the reference image; An identification unit for identifying newly added foreign object contours in each real-time image compared with the reference image; A determination unit, configured to determine the persistence of the newly added foreign object contour in a time series according to the newly added foreign object contour, so as to determine whether the entity corresponding to the newly added foreign object contour is a long-term obstacle.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, When the processor executes the program, the civil aviation radio station obstacle detection method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the civil aviation radio station obstacle detection method according to any one of claims 1 to 7 is implemented.

Images