Optical cable routing identification detection method and device, and electronic device

Abstract

The application discloses a kind of optical cable route mark detection method and device, belong to communication technical field.The method includes: through the target detection of pre-trained long-distance optical cable route mark detection model to the image collected by vehicle image acquisition device in first distance range, obtains the GPS position coordinates of the optical cable route mark to be verified;According to GPS position coordinates and geographic information, obtain the second distance inspection position and position score corresponding to the optical cable route mark to be verified;According to position score, obtain the verification order of the optical cable route mark to be verified;According to the navigation of inspection vehicle according to verification order and second distance inspection position, carry out close-range identification to the optical cable route mark to be verified, obtain the first detection result of the optical cable route mark corresponding to GPS position coordinates.The application carries out two-stage inspection to optical cable route mark by inspection vehicle combined with artificial intelligence technology, realizes the cost reduction and efficiency increase of optical cable route inspection.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to methods and apparatus for detecting optical cable route identifiers, as well as electronic devices and computer-readable storage media. Background Technology

[0002] Fiber optic cables are laid outside of equipment buildings, operating in a complex and variable environment. External construction work and natural disasters constantly affect the operational quality of the fiber optic network. Fiber optic route markers are indicative signs deployed along the fiber optic cable laying lines, indicating the cable's location, operator, and route. Therefore, telecommunications operators often invest significant manpower and resources in frequent inspections of these route markers to promptly identify and eliminate potential problems, ensuring the stable operation of the fiber optic communication network. Traditionally, fiber optic route inspections can only be conducted manually. However, after years of development, fiber optic networks are now deployed nationwide, with long lines, numerous points of contact, and wide coverage. The number of route markers deployed along the cables is vast. Achieving full network coverage through daily inspections would require massive resource investment, and the inspection efficiency would be low.

[0003] It is evident that the existing optical cable route identification inspection method needs improvement. Summary of the Invention

[0004] This application provides a method, apparatus, and electronic device for detecting optical cable route identifiers, which can solve the problems of high labor costs and low inspection efficiency in manual inspection of optical cable routes, and achieve cost reduction and efficiency improvement in optical cable route inspection.

[0005] In a first aspect, embodiments of this application disclose a method for detecting optical cable route identifiers, the method comprising:

[0006] By using a pre-trained long-distance optical cable route identifier detection model, target detection is performed on images acquired by the vehicle-mounted image acquisition device within the first distance range to obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0007] Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location;

[0008] Based on the location score, the verification order of the optical cable route identifier to be verified is obtained;

[0009] Based on the verification sequence and the second distance inspection position, obtain the first vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified;

[0010] The first image acquired at the second distance inspection location during the second distance inspection along the first vehicle navigation route is used to identify the target optical cable route identifier, and the first detection result of the optical cable route identifier corresponding to the GPS location coordinates is obtained.

[0011] Optionally, the step of using a pre-trained long-distance optical cable route identifier detection model to perform target detection on images acquired by the vehicle-mounted image acquisition device within a first distance range, and obtaining the GPS location coordinates of the optical cable route identifier to be verified, includes:

[0012] Based on the GPS location coordinates of the pre-acquired optical cable route identifier, the second vehicle navigation route is obtained during the first distance inspection;

[0013] Using the aforementioned long-distance optical cable route identifier detection model, the target optical cable identifier is detected in the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and a first confidence level is obtained in which the second image includes the optical cable route identifier;

[0014] Based on the first confidence level and the second location, obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0015] Optionally, obtaining the GPS location coordinates of the optical cable route identifier to be verified based on the first confidence level and the second location includes:

[0016] In response to the first confidence level being less than a preset confidence threshold, the GPS location coordinates of the second location match are obtained as the GPS location coordinates of the optical cable route identifier to be verified.

[0017] Optionally, after using the long-distance optical cable route identifier detection model to perform target optical cable identifier detection on the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and obtaining a first confidence level that the second image includes an optical cable route identifier, the method further includes:

[0018] In response to the first confidence level being greater than or equal to the preset confidence threshold, the optical cable route identifier detected at the GPS location coordinates matched by the second location is incremented to obtain a second detection result of the optical cable route identifier corresponding to the GPS location coordinates matched by the second location.

[0019] Optionally, the step of using a pre-trained long-distance optical cable route identifier detection model to perform target detection on images acquired by the vehicle-mounted image acquisition device within a first distance range, and obtaining the GPS location coordinates of the optical cable route identifier to be verified, includes:

[0020] Using the aforementioned long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the third image acquired by the vehicle-mounted image acquisition device at the third location, and a second confidence level is obtained in which the third image includes the optical cable route identifier to be verified.

[0021] In response to the second confidence level indicating that the third image includes an optical cable route identifier to be verified, the GPS location coordinates matching the third location are obtained based on the pre-acquired GPS location coordinates of the optical cable route identifier and the third location.

[0022] In response to the fact that the image size proportion of the optical cable route identifier to be verified included in the third image is greater than or equal to a specified proportion threshold, and the distance between the GPS location coordinates of the optical cable route identifier to be verified and the third location is less than or equal to a specified distance threshold, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the optical cable route identifier to be verified.

[0023] Optionally, the specified percentage threshold is the average percentage of the image size of the optical cable route identifier in multiple images pre-collected by the vehicle-mounted image acquisition device; the specified distance threshold is the average distance between the GPS location coordinates of the optical cable route identifier in the multiple images and the acquisition location of the multiple images.

[0024] Optionally, the long-distance optical cable route identifier detection model is trained using the following method:

[0025] The fourth image, which includes the unobstructed optical cable route identifier and is acquired within the first distance range, is labeled with the optical cable route identifier to generate the first training dataset;

[0026] For the fifth image, which includes the obstructed optical cable route identifier and is acquired within the first distance range, the optical cable route identifier and obstruction type are labeled.

[0027] A second training dataset is generated based on the labeled image obtained by marking the optical cable route identification in the fifth image.

[0028] The images of the optical cable route identifiers that are pre-collected are used to perform image information enhancement processing on the images of the occlusion objects that match the occlusion type labels to generate a third training dataset;

[0029] A long-distance optical cable route identifier detection model is trained based on the first training dataset, the second training dataset, and the third training dataset.

[0030] Optionally, the image information enhancement processing includes:

[0031] The image of the occlusion that matches the occlusion type label is used to partially cover the pre-collected image of the optical cable route identifier.

[0032] Optionally, obtaining the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location based on the GPS location coordinates and geographical information includes:

[0033] Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified;

[0034] The location score of the second distance inspection position is obtained based on the distance between the second distance inspection position and the inspection start point, the distance between the second distance inspection position and the optical cable route identifier to be verified, and the angle between the second distance inspection position and the optical cable route identifier to be verified.

[0035] Secondly, embodiments of this application disclose an optical cable route identification detection device, the device comprising:

[0036] The fiber optic cable route identifier location information acquisition module is used to perform target detection on the images acquired by the vehicle-mounted image acquisition device within a first distance range using a pre-trained long-distance fiber optic cable route identifier detection model, and to obtain the GPS location coordinates of the fiber optic cable route identifier to be verified.

[0037] The location scoring module is used to obtain the second distance inspection location and the location score of the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographic information.

[0038] The sorting module is used to obtain the verification order of the optical cable route identifiers to be verified based on the location score;

[0039] The first navigation route acquisition module is used to acquire the first vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified, based on the verification order and the second distance inspection position.

[0040] The close-range detection module is used to identify the target optical cable route identifier in the first image acquired at the second distance inspection position during the second distance inspection along the first vehicle navigation route, and to obtain the first detection result of the optical cable route identifier corresponding to the GPS location coordinates.

[0041] Optionally, the optical cable route identifier location information acquisition module to be verified is further used for:

[0042] Based on the GPS location coordinates of the pre-acquired optical cable route identifier, the second vehicle navigation route is obtained during the first distance inspection;

[0043] Using the aforementioned long-distance optical cable route identifier detection model, the target optical cable identifier is detected in the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and a first confidence level is obtained in which the second image includes the optical cable route identifier;

[0044] Based on the first confidence level and the second location, obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0045] Optionally, obtaining the GPS location coordinates of the optical cable route identifier to be verified based on the first confidence level and the second location includes:

[0046] In response to the first confidence level being less than a preset confidence threshold, the GPS location coordinates of the second location match are obtained as the GPS location coordinates of the optical cable route identifier to be verified.

[0047] Optionally, after using the long-distance optical cable route identifier detection model to perform target optical cable identifier detection on the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and obtaining a first confidence level that the second image includes an optical cable route identifier, the method further includes:

[0048] In response to the first confidence level being greater than or equal to the preset confidence threshold, the optical cable route identifier detected at the GPS location coordinates matched by the second location is incremented to obtain a second detection result of the optical cable route identifier corresponding to the GPS location coordinates matched by the second location.

[0049] Optionally, the optical cable route identifier location information acquisition module to be verified is further used for:

[0050] Using the aforementioned long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the third image acquired by the vehicle-mounted image acquisition device at the third location, and a second confidence level is obtained in which the third image includes the optical cable route identifier to be verified.

[0051] In response to the second confidence level indicating that the third image includes an optical cable route identifier to be verified, the GPS location coordinates matching the third location are obtained based on the pre-acquired GPS location coordinates of the optical cable route identifier and the third location.

[0052] In response to the fact that the image size proportion of the optical cable route identifier to be verified included in the third image is greater than or equal to a specified proportion threshold, and the distance between the GPS location coordinates of the optical cable route identifier to be verified and the third location is less than or equal to a specified distance threshold, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the optical cable route identifier to be verified.

[0053] Optionally, the specified percentage threshold is the average percentage of the image size of the optical cable route identifier in multiple images pre-collected by the vehicle-mounted image acquisition device; the specified distance threshold is the average distance between the GPS location coordinates of the optical cable route identifier in the multiple images and the acquisition location of the multiple images.

[0054] Optionally, the long-distance optical cable route identifier detection model is trained using the following method:

[0055] The fourth image, which includes the unobstructed optical cable route identifier and is acquired within the first distance range, is labeled with the optical cable route identifier to generate the first training dataset;

[0056] For the fifth image, which includes the obstructed optical cable route identifier and is acquired within the first distance range, the optical cable route identifier and obstruction type are labeled.

[0057] A second training dataset is generated based on the labeled image obtained by marking the optical cable route identification in the fifth image.

[0058] The images of the optical cable route identifiers that are pre-collected are used to perform image information enhancement processing on the images of the occlusion objects that match the occlusion type labels to generate a third training dataset;

[0059] A long-distance optical cable route identifier detection model is trained based on the first training dataset, the second training dataset, and the third training dataset.

[0060] Optionally, the image information enhancement processing includes:

[0061] The image of the occlusion that matches the occlusion type label is used to partially cover the pre-collected image of the optical cable route identifier.

[0062] Optionally, obtaining the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location based on the GPS location coordinates and geographical information includes:

[0063] Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified;

[0064] The location score of the second distance inspection position is obtained based on the distance between the second distance inspection position and the inspection start point, the distance between the second distance inspection position and the optical cable route identifier to be verified, and the angle between the second distance inspection position and the optical cable route identifier to be verified.

[0065] Thirdly, embodiments of this application also disclose an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the optical cable route identifier detection method described in embodiments of this application.

[0066] Fourthly, embodiments of this application disclose a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, performs the steps of the optical cable route identifier detection method disclosed in embodiments of this application.

[0067] The optical cable route identifier detection method disclosed in this application uses a pre-trained long-distance optical cable route identifier detection model to perform target detection on images acquired by a vehicle-mounted image acquisition device within a first distance range. This eliminates some long-distance targets, allowing for accurate identification of the optical cable route identifiers. For the remaining optical cable route identifiers to be verified, close-range verification is performed. Specifically, based on the GPS location coordinates and geographic information, a second distance inspection position and a location score corresponding to the optical cable route identifier to be verified are obtained; based on the location score, the verification order of the optical cable route identifiers to be verified is obtained; and based on the verification order and the second distance inspection position... The system detects the location and obtains the first vehicle navigation route when performing a second distance inspection on the optical cable route identifier to be verified. It then performs target optical cable route identifier identification on the first image captured at the second distance inspection location during the second distance inspection along the first vehicle navigation route, obtaining the first detection result of the optical cable route identifier corresponding to the GPS location coordinates. By combining inspection vehicles with artificial intelligence technology to perform a two-stage inspection of the optical cable route identifier, the system solves the problems of high labor costs and low inspection efficiency in existing manual inspections of optical cable routes, achieving cost reduction and efficiency improvement in optical cable route inspection, and also improving the detection accuracy of optical cable route identifiers.

[0068] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0069] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0070] Figure 1 This is a flowchart of the optical cable route identifier detection method disclosed in the embodiments of this application;

[0071] Figure 2 This is one of the flowcharts of the method for obtaining the GPS location coordinates of the optical cable route identifier to be verified in the embodiments of this application;

[0072] Figure 3 This is the second flowchart of the method for obtaining the GPS location coordinates of the optical cable route identifier to be verified in this application embodiment;

[0073] Figure 4 This is a schematic diagram of the optical cable route identification detection device disclosed in an embodiment of this application;

[0074] Figure 5 A block diagram schematically illustrates an electronic device for performing the method according to this application; and

[0075] Figure 6 A storage unit for holding or carrying program code implementing the method according to this application is illustrated schematically. Detailed Implementation

[0076] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0077] like Figure 1 As shown in the embodiment of this application, a method for detecting optical cable route identifiers includes steps 110 to 150.

[0078] Step 110: Using a pre-trained long-distance optical cable route identifier detection model, target detection is performed on the images acquired by the vehicle-mounted image acquisition device within the first distance range to obtain the GPS (Global Positioning System) location coordinates of the optical cable route identifier to be verified.

[0079] In this embodiment of the application, the first distance range is a distance range greater than or equal to a first distance, and the first distance is greater than a second distance. For example, the first distance can be 30 meters or 50 meters, and the second distance can be 15 meters.

[0080] The optical cable route identifiers described in this application embodiment include, but are not limited to, route identifiers such as markers, utility poles, and billboards.

[0081] In the embodiments of this application, a long-distance optical cable route identifier detection model is first trained based on an image including a target optical cable route identifier (e.g., a telecommunications optical cable route identifier). This model is used to perform target detection on images of optical cable route identifiers collected over long distances, detect whether there is a target optical cable route identifier in the image, and obtain a first confidence level that the image includes the target optical cable route identifier.

[0082] During the inspection of optical cables, images captured by the vehicle-mounted image acquisition device within a first distance range are input to the long-distance optical cable route identifier detection model. The long-distance optical cable route identifier detection model can output a first confidence level that the image includes the target optical cable route identifier. Further, based on the first confidence level, it can be determined that the image includes the target optical cable route identifier, is suspected of including the optical cable route identifier, or does not include the target optical cable route identifier. For images that include or do not include the target optical cable route identifier, in some embodiments of this application, the detection result of the current image can be directly output; for images suspected of including the optical cable route identifier, further close-range photographic verification is required.

[0083] In the embodiments of this application, the acquisition location of an image suspected of containing an optical cable route identifier can be used as the GPS location coordinates of the optical cable route identifier to be verified. Alternatively, optical cable route identifiers near the acquisition location can be found in the operation and maintenance system, and the GPS location coordinates of the found optical cable route identifiers can be used as the GPS location coordinates of the optical cable route identifier to be verified to guide subsequent steps for further verification.

[0084] Step 120: Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location.

[0085] For multiple fiber optic cable route markers to be verified discovered during the first-distance inspection, the GPS location coordinates of each marker can be used to query the geographic information system to obtain information on terrain, traffic obstacles, and other transportation conditions near each marker. Then, based on the obtained transportation condition information, a second-distance inspection location reachable from the markers to be verified can be determined.

[0086] Furthermore, based on the distance, angle, and orientation between the second distance inspection location and the corresponding optical cable route identifier to be verified, a location score is calculated for the second distance inspection location. Optionally, the location score indicates the accuracy of collecting the corresponding optical cable route identifier to be verified at the second distance inspection location for identification and verification. The smaller the angle of deviation between the second distance inspection location and the optical cable route identifier to be verified, the lower the location score. For example, the smaller the distance between the second distance inspection location and the optical cable route identifier to be verified, the lower the location score.

[0087] Step 130: Based on the location score, obtain the verification order of the optical cable route identifier to be verified.

[0088] Next, the optical cable route identifiers to be verified can be sorted from front to back according to the location score from low to high to obtain the verification order of the optical cable route identifiers to be verified.

[0089] In some embodiments of this application, optical cable route identifiers with location scores lower than a preset score threshold can be selected for close-range verification.

[0090] Step 140: Based on the verification sequence and the second distance inspection position, obtain the first vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified.

[0091] In the embodiments of this application, the second distance inspection refers to close-range inspection. For example, inspection within a distance of 15 meters from the target.

[0092] Next, following the verification order obtained in the previous step and the second distance inspection position corresponding to each optical cable route identifier to be verified, the second distance inspection is performed on each optical cable route identifier to be verified in sequence to verify whether the optical cable route identifier to be verified is the target optical cable route identifier, and whether there are any phenomena such as obstruction or damage.

[0093] In the embodiments of this application, a second distance inspection is performed on the optical cable route identifier to be verified by an inspection vehicle. Therefore, it is first necessary to obtain the vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified. For example, the second distance inspection location of the optical cable route identifier S1 to be verified can be used as the destination, and the first vehicle navigation route for performing the second distance inspection on the optical cable route identifier S1 to be verified can be obtained by calling a navigation application.

[0094] Step 150: For the first image collected at the second distance inspection position during the second distance inspection along the first vehicle navigation route, target optical cable route identifier identification is performed to obtain the first detection result of the optical cable route identifier corresponding to the GPS location coordinates.

[0095] The first location is a location within a preset distance range from the optical cable route identifier to be verified. For example, it can be a location less than or equal to 30 meters from the optical cable route identifier to be verified.

[0096] For example, when the inspection vehicle is traveling along the first vehicle navigation route to perform close-range verification of the optical cable route identifier S1 to be verified, the vehicle-mounted image acquisition device can be turned on when the inspection vehicle travels along the first vehicle navigation route to a distance of less than 30 meters from the optical cable route identifier S1 to be verified, and the image of the optical cable route identifier S1 to be verified can be acquired (referred to as "first image" in this application embodiment).

[0097] Next, the first image is subjected to target optical cable route identification to obtain the identification result of whether the first image contains the target optical cable route identifier. For example, whether the first image contains the optical cable route identifier of a telecommunications company.

[0098] In some embodiments of this application, a pre-trained target detection model can be used to detect whether the first image contains a target optical cable route identifier, such as whether it contains a telecommunications icon.

[0099] In other embodiments of this application, optical character recognition algorithms (such as the PaddleOCR algorithm, Baidu's open-source optical character recognition framework) can also be used to identify whether the target optical cable route identifier in the first image contains the word "telecom", thereby determining whether the first image contains the target optical cable route identifier.

[0100] When the first image contains a target optical cable route identifier, the optical cable route identifier to be verified corresponding to the GPS location coordinates can be considered as the target optical cable route identifier, that is, the target optical cable route identifier is detected at the GPS location coordinates. Conversely, if none of the multiple first images collected at multiple first locations near the GPS location coordinates (i.e., multiple images collected when the inspection vehicle travels along the first vehicle navigation route to the vicinity of the GPS location coordinates) contain a target optical cable route identifier, then the detection result is that no target optical cable route identifier was detected at the GPS location coordinates.

[0101] In some embodiments of this application, if the target optical cable route identifier is not detected at the GPS location coordinates corresponding to the optical cable route identifier to be verified, an alarm message indicating the optical cable route identifier is missing can be output to the operation and maintenance system to facilitate manual verification by operation and maintenance personnel.

[0102] This application also discloses a method for detecting optical cable route identifiers. Using a pre-trained long-distance optical cable route identifier detection model, it performs target detection on images acquired by a vehicle-mounted image acquisition device within a first distance range. This eliminates some long-distance targets, allowing for accurate identification of the optical cable route identifiers. For the remaining optical cable route identifiers to be verified, it performs near-distance verification. Specifically, based on the GPS location coordinates and geographic information, it obtains the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location; based on the location score, it obtains the verification order of the optical cable route identifier to be verified; and based on the verification order and the second distance... The inspection location is used to obtain the first vehicle navigation route when performing a second distance inspection on the optical cable route identifier to be verified; the first image collected at the second distance inspection location during the second distance inspection along the first vehicle navigation route is used to identify the target optical cable route identifier, and the first detection result of the optical cable route identifier corresponding to the GPS location coordinates is obtained. By combining inspection vehicles with artificial intelligence technology to perform two-stage inspection of optical cable route identifiers, the problems of high labor costs and low inspection efficiency in the existing technology of manual inspection of optical cable routes are solved, thereby achieving cost reduction and efficiency improvement in optical cable route inspection and improving the detection accuracy of optical cable route identifiers.

[0103] To make the optical cable route identification detection method disclosed in the embodiments of this application clearer, the specific implementation methods of the aforementioned steps are further illustrated below.

[0104] In step 110 above, when obtaining the GPS location coordinates of the optical cable route identifier to be verified, different methods are used to obtain the GPS location coordinates of the optical cable route identifier to be verified for different images acquired within the first distance range. Below, we will use two different image acquisition methods as examples to illustrate the specific implementation methods for obtaining the GPS location coordinates of the optical cable route identifier to be verified.

[0105] The first method involves using inspection vehicles to collect images of fiber optic cable route markers from a distance.

[0106] like Figure 2 As shown, when using a professional inspection vehicle to inspect optical cable route markers, the step of using a pre-trained long-distance optical cable route marker detection model to perform target detection on the images acquired by the vehicle-mounted image acquisition device within a first distance range and obtain the GPS location coordinates of the optical cable route marker to be verified includes: sub-steps 1101, 1102, and 1103.

[0107] Sub-step 1101: Based on the GPS location coordinates of the pre-acquired optical cable route identifier, obtain the second vehicle navigation route during the first distance inspection.

[0108] In the embodiments of this application, the first distance inspection refers to long-distance inspection, for example, performing an inspection on the target at a location more than 30 meters away from the target.

[0109] In some embodiments of this application, before detecting optical cable route markers, the deployment geographical location information of the optical cable route markers, i.e., the GPS location coordinates of the optical cable route markers, can first be obtained through the optical cable operation and maintenance system. Then, using the GPS location coordinates of the optical cable route markers as the destination, a second vehicle navigation route for inspecting each optical cable route marker at a first distance is obtained through a navigation application. In the embodiments of this application, the first distance inspection can be understood as inspecting optical cable route markers along the main route. Since main optical cables are usually laid along highways or trunk roads, inspection vehicles can achieve long-distance inspection of optical cable route markers by traveling along highways or trunk roads.

[0110] In embodiments of this application, a long-distance threshold can be set. When the distance between the inspection vehicle and the optical cable route marker to be inspected meets the long-distance threshold, the vehicle-mounted image acquisition device of the inspection vehicle begins to acquire images of the optical cable route marker to perform a first-distance inspection of the optical cable route marker. The long-distance threshold can be set, for example, to 30 meters.

[0111] Sub-step 1102: Using the long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and a first confidence level is obtained in which the second image includes the optical cable route identifier.

[0112] In some embodiments of this application, when the inspection vehicle travels along the second vehicle navigation route and performs a first-distance inspection of the optical cable route marker, in order to reduce the computational load of image processing, the on-board image acquisition device can be activated to acquire the image of the optical cable route marker to be inspected only when the inspection vehicle reaches the designated location. For example, when the distance between the inspection vehicle and the optical cable route marker to be inspected is less than or equal to a preset long-distance threshold, the on-board image acquisition device is activated to acquire the image of the optical cable route marker to be inspected. In the embodiments of this application, the image of the optical cable route marker to be inspected acquired by the on-board image acquisition device on the inspection vehicle is recorded as the "second image," and the position where the on-board image acquisition device acquires the second image is recorded as the "second position."

[0113] In some embodiments of this application, the inspection vehicle can associate a second image acquired by the vehicle-mounted image acquisition device with a second location where the second image was acquired, and then send the association to a background detection data processing system for subsequent detection data processing. The background detection data processing system calls a pre-trained long-distance optical cable route identifier detection model to perform target detection on the second image and outputs a first confidence level that the second image includes the target optical cable route identifier.

[0114] In other embodiments of this application, the inspection vehicle can perform detection data processing on the second image acquired by the vehicle-mounted image acquisition device locally, call a pre-trained long-distance optical cable route identifier detection model to perform target detection on the second image, and obtain a first confidence score that the second image includes the optical cable route identifier. Then, the first confidence score and the second location are correlated and sent to the background detection data processing system.

[0115] In this way, the background detection data processing system can obtain multiple sets of inspection data based on the data sent by the inspection vehicle. Each set of inspection data includes a second location and a first confidence level in the second image collected at the second location, which includes the optical cable route identifier.

[0116] Sub-step 1103: Based on the first confidence level and the second location, obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0117] In some embodiments of this application, obtaining the GPS location coordinates of the optical cable route identifier to be verified based on the first confidence level and the second location includes: in response to the first confidence level being less than a preset confidence threshold, obtaining the GPS location coordinates that match the second location as the GPS location coordinates of the optical cable route identifier to be verified.

[0118] As can be seen from the aforementioned method for acquiring the second image, the second location is the real-time location of the inspection vehicle. The inspection vehicle acquires images of fiber optic cable route identifiers within a preset distance range (such as a preset long-distance threshold) from the second location. Therefore, it is necessary to further obtain the location information of the fiber optic cable route identifiers that may be included in the second image based on the second location. In some embodiments of this application, the GPS location coordinates of the fiber optic cable route identifiers can be retrieved in the operation and maintenance system based on the second location to obtain each fiber optic cable route identifier near the second location (such as within a preset long-distance threshold), as the fiber optic cable route identifiers that may be included in the second image, and the GPS location coordinates of the fiber optic cable route identifiers that may be included in the second image are recorded as the GPS location coordinates matched with the second location.

[0119] During the initial distance inspection, if the fiber optic route identifier is obscured or located far from the inspection route, the confidence score (i.e., the first confidence score) obtained by calling the long-distance fiber optic route identifier detection model to perform target detection on the second image will be low. When the first confidence score is less than the confidence score threshold, it is considered that the long-distance fiber optic route identifier detection model cannot accurately identify whether the second image includes the fiber optic route identifier to be detected, and further close-range detection is required. In the embodiments of this application, the fiber optic route identifier that needs to be further close-range detected is considered to be the fiber optic route identifier to be verified, and the GPS location coordinates of the fiber optic route identifier to be verified are obtained.

[0120] Conversely, if the fiber optic cable route marker is not obstructed, or is relatively close to the inspection route, the fiber optic cable route marker can be clearly displayed in the second image collected by the inspection vehicle. In this case, the confidence level (i.e., the first confidence level) obtained by calling the long-distance fiber optic cable route marker detection model to perform target detection on the second image will be higher.

[0121] In some embodiments of this application, after using the long-distance optical cable route identifier detection model to detect target optical cable identifiers in the second image acquired by the vehicle-mounted image acquisition device at a second location on the second vehicle navigation route, and obtaining a first confidence level that the second image includes optical cable route identifiers, the method further includes: in response to the first confidence level being greater than or equal to the preset confidence threshold, incrementing the count of optical cable route identifiers detected at the GPS location coordinates matched at the second location, to obtain a second detection result of the optical cable route identifiers corresponding to the GPS location coordinates matched at the second location. For example, when the first confidence level is greater than or equal to the confidence threshold, it is considered that the long-distance optical cable route identifier detection model has accurately identified that the second image includes the optical cable route identifier to be detected (such as a telecommunications optical cable route identifier), and no further close-range detection is required.

[0122] In some embodiments of this application, the number of optical cable route identifiers detected at the GPS location coordinates matching the second location is incremented to obtain the number of optical cable route identifiers detected at the second location. Further, the number of GPS location coordinates matching the second location is obtained through the operation and maintenance system, i.e., the number of optical cable route identifiers deployed near the second location. If the number of optical cable route identifiers detected at the second location is less than the number of optical cable route identifiers deployed near the second location, the detection is considered incomplete, and a second detection result indicating that repeated inspection is needed can be output. If the number of optical cable route identifiers detected at the second location is greater than the number of optical cable route identifiers deployed near the second location, it is considered that the detected optical cable route identifiers to be verified include signposts and optical cable route identifiers not to be inspected (such as optical cable route identifiers from other companies), requiring further close-range verification.

[0123] The second method involves collecting images of fiber optic cable route identifiers from a distance using non-inspection vehicles.

[0124] like Figure 3 As shown, when using non-professional inspection vehicles (such as buses, passenger dedicated line vehicles, etc.) to collect images of optical cable route identifiers, the step of using a pre-trained long-distance optical cable route identifier detection model to perform target detection on the images collected by the vehicle-mounted image acquisition device within a first distance range and obtain the GPS location coordinates of the optical cable route identifier to be verified includes: sub-steps 1105, 1106, and 1107.

[0125] Sub-step 1105: Using the long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the third image acquired by the vehicle-mounted image acquisition device at the third location, and a second confidence level is obtained in the third image that includes the optical cable route identifier to be verified.

[0126] Since fiber optic cables are typically laid along main highways, vehicles that travel on fixed routes, such as buses and dedicated passenger transport vehicles, can also be used to collect fiber optic cable route information. For example, by installing onboard image acquisition devices with positioning mechanisms on buses and dedicated passenger transport vehicles that travel on fixed routes, the onboard image acquisition devices can collect images of both sides of the road along the route while the vehicle is in motion, and at the same time, the positioning devices can collect the real-time location of the vehicle.

[0127] In the embodiments of this application, the images collected by the above-mentioned buses, passenger dedicated line vehicles and other vehicles traveling along fixed routes are recorded as "third images", and the real-time position of the vehicle when the third image is collected (i.e. the real-time position of the vehicle-mounted image acquisition device) is recorded as "third position".

[0128] The vehicle-mounted image acquisition device can associate the real-time acquired third image with the third location and send it to the background detection data processing system for subsequent detection data processing. The background detection data processing system calls a pre-trained long-distance optical cable route identifier detection model to perform target detection on the third image and outputs the confidence score (i.e., the second confidence score) that the third image contains the target optical cable route identifier.

[0129] In other embodiments of this application, the vehicle-mounted image acquisition device can also perform detection data processing on the acquired third image locally, calling a pre-trained long-distance optical cable route identifier detection model to perform target detection on the third image, and obtain a second confidence score that the third image includes the optical cable route identifier. Then, the second confidence score and the third location are associated and sent to the background detection data processing system.

[0130] In this way, the background detection data processing system can obtain multiple sets of inspection data based on the data sent by the vehicle-mounted image acquisition device. Each set of inspection data includes a third location and a second confidence level in the third image acquired at that third location, which includes the optical cable route identifier.

[0131] Sub-step 1106: In response to the second confidence level indicating that the third image includes a suspected optical cable route identifier, the GPS location coordinates matching the third location are obtained based on the pre-acquired GPS location coordinates of the optical cable route identifier and the third location.

[0132] In some embodiments of this application, when the second confidence level obtained in the aforementioned steps is greater than or equal to a specified confidence threshold, it can be assumed that the third image includes the optical cable route identifier to be verified. In this case, the GPS location coordinates of the optical cable route identifier can be retrieved in the operation and maintenance system based on the third location to obtain each optical cable route identifier near the third location (e.g., less than a preset distance threshold), which can be used as the optical cable route identifier that may be included in the third image, i.e., the GPS location coordinates matched by the third location.

[0133] Sub-step 1107: In response to the fact that the image size proportion of the optical cable route identifier to be verified included in the third image is greater than or equal to a specified proportion threshold, and the distance between the GPS location coordinates of the optical cable route identifier to be verified and the third location is less than or equal to a specified distance threshold, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the optical cable route identifier to be verified.

[0134] Since the vehicles equipped with the vehicle-mounted image acquisition devices are not professional inspection vehicles, and the distance between their driving routes and the fiber optic cable deployment locations is uncertain, the target objects in the third image acquired by the vehicle-mounted image acquisition device may be fiber optic cable route markers, or they may be road signs or other markers. Determining whether the third image contains a fiber optic cable route marker or a fiber optic cable route marker to be verified solely based on the confidence level output by the long-distance fiber optic cable route marker detection model will result in decreased accuracy. In the embodiments of this application, after the confidence level output by the long-distance fiber optic cable route marker detection model indicates that the third image contains a fiber optic cable route marker to be verified, the accuracy is further determined based on the image size proportion of the target object (i.e., the fiber optic cable route marker to be verified) in the third image, and the distance between the third location and the location of the fiber optic cable route marker actually deployed in the operation and maintenance system, to determine whether the third image contains a fiber optic cable route marker that needs to be verified at close range.

[0135] In some embodiments of this application, the specified percentage threshold is the average of the image size percentage of the optical cable route identifier in multiple images pre-collected by the vehicle-mounted image acquisition device; the specified distance threshold is the average of the distance between the GPS location coordinates of the optical cable route identifier in the multiple images and the acquisition location of the multiple images.

[0136] For example, when inspection personnel are exploring routes by themselves during inspections of fiber optic cable route markers, the route information and videos they capture during the journey can be sent to a relevant application client on their smart terminals. The personnel can then annotate the video time point where the target fiber optic cable route marker was found, along with the image size percentage of the target fiber optic cable route marker within the video image at that time point, and record the distance between the vehicle's geographical location and the found target fiber optic cable route marker at that time point. After recording the distance multiple times, the average distance is taken as a specified distance threshold corresponding to the road segment at that video time point. Furthermore, based on the image size percentages of the target fiber optic cable route markers recorded multiple times, the average image size percentage is calculated and used as a specified percentage threshold.

[0137] During the inspection, if the image size of the fiber optic route identifier to be verified included in the third image is greater than or equal to a specified percentage threshold, and the distance between the GPS location coordinates of the fiber optic route identifier to be verified and the third location is less than or equal to a specified distance threshold, it indicates that the vehicle is relatively close to the fiber optic route identifier to be verified, the acquired image has high reliability, and the long-distance target detection result of the third image can be considered accurate. In this case, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the fiber optic route identifier to be verified.

[0138] In some embodiments of this application, the long-distance optical cable route identifier detection model can be a target detection model based on a Faster R-CNN (a convolutional neural network) network structure, or it can be a target detection model based on other network structures. In these embodiments, the specific structure of the long-distance optical cable route identifier detection model is not limited.

[0139] Optionally, the long-distance optical cable route identifier detection model can be trained based on several labeled images containing target optical cable route identifiers. For example, images of optical cable route identifiers can be collected within a first distance range. Then, the collected images are manually labeled with the target objects, marking the position coordinates of the optical cable route identifiers in each image. Afterward, the labeled images containing optical cable route identifiers are used as a training dataset to train the long-distance optical cable route identifier detection model.

[0140] The long-distance optical cable route identification detection model mainly consists of a backbone convolutional neural network, a feature pyramid structure, a Region Proposal Network (RPN) fully convolutional neural network, and a Region of Interest (ROI) classifier. After data preprocessing, the image is input into the backbone convolutional neural network and the feature pyramid structure to extract features. Then, the feature map output by the feature pyramid is subjected to a 3×3 convolution to increase the receptive field of the network, followed by 1×1 convolution for dimensionality reduction, finally generating proposal candidate boxes. After that, the ROI classifier performs bilinear interpolation to obtain the mapped feature map, and finally connects to a fully connected layer to output regression and classification to obtain positive candidate boxes.

[0141] In the embodiments of this application, during the training phase, the long-distance optical cable route identification detection model preprocesses and augments the images, then sends them to the backbone network for downsampling to extract features, and then sends them to the feature pyramid for upsampling. During this process, features from different levels are fused, and then the images are sent to the RPN fully convolutional neural network to generate proposals. After that, the feature map is ROI-aligned to a size of 7×7, and then classification and regression are performed to obtain the optical cable route identification detection results.

[0142] For example, for the dataset of obtained telecommunications fiber optic cable route identifiers, data preprocessing is performed first. For instance, the width and height of the images in the dataset are restricted, and the input images are uniformly restricted to a preset size. In the input network, features are fused through 4 downsampling, 3 upsampling, and 1 max pooling operation using a 152-layer residual neural network and a feature pyramid structure, ensuring that each fusion is divisible by 32.

[0143] By saving the trained model and loading the parameters of the trained model, it is possible to directly predict the telecommunications fiber optic cable route identifiers and locations in an image of any size end-to-end.

[0144] In other embodiments of this application, given the complexity of the fiber optic route identifier deployment environment and the possibility of obstruction or damage, in order to improve the accuracy of long-distance fiber optic route identifier detection, an image information enhancement processing method is used to expand the training dataset to improve the generalization ability of the long-distance fiber optic route identifier detection model.

[0145] The long-distance optical cable route identifier detection model is trained using the following method: A fourth image, including unobstructed optical cable route identifiers, acquired within a first distance range, is labeled with optical cable route identifiers to generate a first training dataset; a fifth image, including obstructed optical cable route identifiers, acquired within the first distance range, is labeled with both optical cable route identifiers and obstruction type; a second training dataset is generated based on the labeled image obtained from the optical cable route identifier labeling of the fifth image; image information enhancement processing is performed on the pre-acquired images of the optical cable route identifiers using an image of an obstruction that matches the obstruction type label, generating a third training dataset; and the long-distance optical cable route identifier detection model is trained based on the first training dataset, the second training dataset, and the third training dataset.

[0146] For example, part of the training dataset consists of unobstructed images of the optical cable route markers collected within a first distance range, denoted as "Fourth Image". The Fourth Image is manually labeled to indicate the location of the optical cable route markers in the Fourth Image.

[0147] For example, another part of the training dataset consists of images of the obscured fiber optic route markers collected within a first distance range, denoted as "the fifth image." The fifth image is manually labeled, indicating the location of the fiber optic route markers within it. In this fifth image, the fiber optic route markers are partially obscured.

[0148] For example, image enhancement processing can be performed on the fourth or fifth image to obtain a richer training dataset. Image enhancement processing includes, but is not limited to, image enhancement processing based on geometric transformations and color transformations. Geometric transformation-based image enhancement processing includes, but is not limited to, random cropping, random expansion, random horizontal flipping, and random scaling; color transformation-based image enhancement processing includes, but is not limited to, color dithering and Fancy PCA (Principal Component Analysis).

[0149] For example, based on the type of occlusion appearing in the fifth image, image processing techniques can be used to generate images of other occlusion scenarios to enhance the image information. Optionally, the image information enhancement processing includes: using occlusion images that match the occlusion type labels to partially cover the pre-collected image of the optical cable route identifier. Specifically, the types of occlusions included in the fifth image can be counted, and occlusion images can be extracted. Then, the extracted occlusion images are used to occlude the image of the optical cable route identifier to simulate the optical cable route identifier being occluded by different types of occlusions, and the occlusion of different positions of the optical cable route identifier, thereby expanding the training dataset of the long-distance optical cable route identifier detection model.

[0150] In some embodiments of this application, the fifth image can be acquired using the following method: When the inspection vehicle travels along the second vehicle navigation route, if it reaches the vicinity of the GPS location coordinates of a suspected telecommunications fiber optic cable route marker to be inspected, and cannot detect the marker from a distance, the GPS location coordinates can be considered incorrect. The inspection personnel can then drive themselves to explore the surrounding area for telecommunications fiber optic cable route markers. After the inspection personnel discover the marker, an image of the marker, i.e., the fifth image, is acquired, and the location coordinates of the discovered marker, as well as the location and type of obstructions in the fifth image, are marked. This operation uncovers more image material of obstructed or hidden telecommunications fiber optic cable route markers, enhancing the model's generalization ability.

[0151] In one application scenario, when a user drives according to the navigation-recommended route, but upon reaching the GPS location point indicated by the GPS coordinates for a suspected fiber optic cable route marker, the system fails to detect the marker at a distance. This indicates the coordinates of the fiber optic cable route marker are incorrect, and the user is instructed to drive independently and explore the surrounding area for the suspected marker. If the suspected marker is found, the exploration route is recorded, and the distance between the actual location where the suspected marker was found and the GPS-located marker is determined. If the distance is below a threshold, it may be due to obstruction, preventing the marker from being detected. Video footage of the recorded exploration route is sent to the user's mobile app, allowing the user to annotate the video timestamps of the suspected marker discovery. The timetamps of the discovered markers and the timestamps of obstructed markers are then used for target detection annotation, resulting in a fiber optic cable route marker obstruction dataset, which is added to the overall fiber optic cable route marker dataset. The system analyzes the types of route identifiers that are suspected to be blocked by telecommunications fiber optic cables; it also analyzes the exploration routes driven by users in this section of road, and performs data enhancement on specific types of blocked route identifiers in similar road sections.

[0152] The types of obstructed route identifiers that are suspected to be obstructed by telecommunications optical cable routes include: trees, weeds, stone blocks, etc. These types of obstructed route identifiers will provide image enhancement support for subsequent training of a better optical cable route identifier target detection model.

[0153] Optionally, specific types of occlusion route identifier data augmentation can be performed on similar road segments, including: recording GPS images of the road segment, recording the user's own driving exploration route and the types of occlusion route identifiers encountered on the road segment, and when performing image augmentation on the training set for subsequent target detection, similar road segments with a similarity higher than a threshold to the GPS images of the road segment can be obtained through image similarity. When performing image augmentation on the telecommunications fiber optic cable target route identifiers of the road segment, the same occlusion route identifier type can be used for data augmentation.

[0154] During image enhancement, a cropping method is primarily used to cut out occlusion types and integrate them into images of telecommunications fiber optic cable route markers. This aims to improve the model's generalization effect on occluded telecommunications fiber optic cable route markers. Furthermore, since the acquired dataset is based on user-explored road segments, and different users encounter different occlusions during their explorations, the occlusion cropping process is as follows: First, specific occlusions are manually selected and cropped, then cropped and integrated into images with similar road conditions for occlusion data enhancement. In subsequent image enhancement processes, random cropping is performed for different road conditions. The cropped content is set according to different road conditions and includes stone blocks, trees, weeds, etc. Road conditions can be obtained through GPS map images using road condition similarity calculations.

[0155] Optionally, image enhancement can be performed on images near the passing position, mainly by using random horizontal flipping to introduce more angles, so as to achieve the model generalization effect on telecom optical cable target objects from multiple angles.

[0156] By performing image information enhancement processing on the collected optical cable route identification images, the training dataset of the long-distance optical cable route identification detection model is expanded. This effectively overcomes the impact of problems such as the optical cable route identification angle being deflected, obstructed, or small in size in the images collected by the inspection vehicle during the inspection process on target detection, and effectively improves the accuracy of the long-distance optical cable route identification detection model in identifying the optical cable route identification to be verified in the images collected within the first distance range.

[0157] In some embodiments of this application, step 120 above, obtaining the second distance inspection location and the location score of the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographical information, includes: obtaining the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographical information; obtaining the location score of the second distance inspection location based on the distance between the second distance inspection location and the inspection starting point, the distance between the second distance inspection location and the optical cable route identifier to be verified, and the angle between the second distance inspection location and the optical cable route identifier to be verified.

[0158] Optionally, obtaining the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographic information includes: obtaining candidate arrival locations within a preset distance range of the optical cable route identifier to be verified based on the GPS location coordinates; calculating the selection score of each candidate arrival location based on the geographic information between each candidate arrival location and the optical cable route identifier to be verified; and selecting the candidate arrival location as the second distance inspection location corresponding to the optical cable route identifier to be verified based on the selection score.

[0159] Optionally, the geographic information includes one or more of the following: distance, orientation, obstacles, etc.

[0160] In some embodiments of this application, after determining the optical cable route identifier to be verified at close range, a navigation application can be invoked to obtain all routes that the vehicle can take to reach the optical cable route identifier, and the location on each route that is closest to the optical cable route identifier can be determined. Then, a location whose distance from the optical cable route identifier meets a preset distance range is selected as a candidate arrival location. For example, a location less than or equal to 15 meters from the optical cable route identifier is selected as a candidate arrival location.

[0161] Next, geographic information such as the distance between each candidate arrival location and the corresponding optical cable route identifier to be verified, the azimuth angle between the candidate arrival location and the front of the corresponding optical cable route identifier to be verified, and the obstacle information between the candidate arrival location and the corresponding optical cable route identifier to be verified can be obtained through the geographic information system in the prior art. Based on the obtained geographic information, the selection score of each candidate arrival location is calculated.

[0162] In some embodiments of this application, panoramic segmentation of the real-world map directly in front of the optical cable route identifier to be verified can be performed to determine the obstacles between the candidate arrival location and the optical cable route identifier to be verified. The obstacles may include: stone blocks, stagnant water, trees, etc.

[0163] In some embodiments of this application, the maximum azimuth angle and the maximum distance that can accurately identify and verify whether it is a target optical cable route identifier can be determined empirically. When the maximum angle is, for example, 60 degrees and the maximum distance is, for example, 15 meters, the accuracy rate of verifying the target optical cable route identifier is higher than the threshold of 95%. However, when the angle is 70 degrees, the verification accuracy rate can only reach 80%. Therefore, the maximum azimuth angle can be set to 60 degrees and the maximum distance can be set to 15 meters. Then, based on the ratio between the distance between the candidate arrival location and the optical cable route identifier to be verified and the set maximum distance, and the ratio between the azimuth angle between the candidate arrival location and the optical cable route identifier to be verified and the preset maximum azimuth angle, a first position score of the candidate arrival location is calculated. Then, based on whether there are obstacles between the candidate arrival location and the optical cable route identifier to be verified, a second position score of the candidate arrival location is calculated. Finally, based on the first position score and the second position score, the selection score of the candidate arrival location is calculated.

[0164] The azimuth and distance between the candidate arrival location and the optical cable route identifier to be verified are used to negatively correlate the first location score. The second location score is higher when there are no obstacles between the candidate arrival location and the optical cable route identifier to be verified than when there are obstacles.

[0165] Then, the candidate arrival location with the highest point score among the candidate arrival locations corresponding to each optical cable route identifier to be verified can be used as the second distance inspection location corresponding to that optical cable route identifier to be verified.

[0166] In some embodiments of this application, a location score for the second distance inspection location is obtained based on the distance between the second distance inspection location and the inspection start point, the distance between the second distance inspection location and the optical cable route identifier to be verified, and the angle between the second distance inspection location and the optical cable route identifier to be verified. This includes: performing a weighted summation of the distance between the second distance inspection location and the inspection start point, the distance between the second distance inspection location and the optical cable route identifier to be verified, and the angle between the second distance inspection location and the optical cable route identifier to be verified, with preset weights; and using the summation obtained by the weighted summation as the location score for the second distance inspection location.

[0167] For example, the location score of the second distance inspection position can be calculated using the formula Al lw = Wd*d + Wa*a + Wj*j, where Al lw is the weighted sum; d is the distance between the second distance inspection position and the optical cable route identifier to be verified (e.g., the distance between the suspected telecommunications identifier and the optimal position), Wd is the corresponding distance weight value; a is the angle between the second distance inspection position and the optical cable route identifier to be verified (e.g., the angle between the suspected telecommunications identifier and the optimal position), Wa is the angle weight value; j is the distance between the second distance inspection position and the inspection starting point (e.g., the distance the inspection vehicle is currently traveling to the optimal position), Wj is the distance weight of the inspection vehicle currently traveling to the second distance inspection position.

[0168] As mentioned earlier, the second distance inspection position is determined by panoramic segmentation to determine whether there is an obstacle directly in front of the suspected optical cable route sign, and to determine the optimal position that the vehicle can drive to at the closest and most direct angle due to the obstacle.

[0169] In actual inspections, adjusting the camera's focal length allows for larger and more detailed images. Distance becomes easier to determine, while angle is more difficult to obtain. Therefore, different weights can be assigned to distance and angle based on the specific requirements. For example, the distance weight Wd can be set to 1, and the angle weight to 5. Specifically, when calculating the location score for the second distance inspection position corresponding to each fiber optic cable route identifier (such as a telecom fiber optic cable route identifier), the score can be calculated as: 1 meter distance * weight 1 + 5 degrees angle * weight 5. Larger angles are more difficult to identify, and greater distances are also more difficult to identify.

[0170] On the other hand, since verification is performed in a sorted manner, it's necessary to consider that inspection vehicles should ideally start verifying from the nearest location and then move to the more distant ones. Therefore, the distance from the current inspection vehicle to the second-distance inspection location is taken into account, with closer locations being inspected first. For example, if the current distance from the inspection vehicle to the second-distance inspection location is calculated as 1 meter * weight 3, then the inspection weight of the second-distance inspection location corresponding to the optical cable route identifier to be verified is obtained. The higher the weight, the more difficult it is to identify, and therefore, the later it is sorted.

[0171] As can be seen from the calculation method of the location score of the second distance inspection location, the higher the location score, the greater the verification difficulty. When the location score is higher than a certain value, it means that the optical cable route identifier to be verified cannot be accurately verified at the second distance inspection location.

[0172] In some embodiments of this application, when obtaining the verification order of the optical cable route identifier to be verified based on the location score, the second distance inspection position with the location score higher than the preset score threshold can be regarded as a failing position and not sorted.

[0173] After completing the sequential verification of the fiber optic route identifiers to be verified, for those failing the verification at their corresponding locations, an elimination method can be used to determine whether the unreachable location warrants a short-distance route navigation detection to identify the target fiber optic route identifier. For example, during the sequential verification process, it is continuously checked whether the number of target fiber optic route identifiers near the coordinates of the stored fiber optic route identifiers in the operation and maintenance system matches the number of such identifiers. If only one is needed to match, and only one fiber optic route identifier remains to be verified, a sufficiently confident judgment result cannot be obtained because the unreachable location prevents a clear confirmation. Therefore, it is determined whether the coordinates of the stored fiber optic route identifiers in the operation and maintenance system and the coordinates of the last unverified fiber optic route identifier are less than a set threshold. If they are less, the unverified fiber optic route identifier is identified as the target fiber optic route identifier.

[0174] The optical cable route identification detection method disclosed in this application first performs preliminary target detection and judgment on suspected target optical cable route identifications under the condition of multiple obstructions of small targets at a distance. After excluding some target optical cable route identifications that can be accurately identified at a distance, the remaining suspected target optical cable route identifications that need to be verified at close range and from multiple angles are further verified. Priority is given to verifying the optical cable route identifications that can reach the optimal position, thereby obtaining efficient and accurate target optical cable route identification detection results. By replacing manual inspection with vehicle inspection and detecting potential hazards, the inspection efficiency of optical cable route identification is effectively improved.

[0175] On the other hand, by combining long-distance target detection with short-distance target recognition, problems with optical cable routing markings can be more accurately identified, avoiding failures caused by subjective reasons during manual inspections failing to detect potential hazards in time.

[0176] Furthermore, by continuously exploring the obstruction status of optical cable route markers during the inspection process, and expanding the training dataset through image information enhancement processing, the long-distance optical cable route marker detection model is updated and trained to continuously improve the model's generalization ability and enhance the accuracy of identifying optical cable route markers to be verified in images acquired at long distances.

[0177] Accordingly, embodiments of this application also disclose an optical cable route identification detection device, such as... Figure 4 As shown, the device includes:

[0178] The optical cable route identifier location information acquisition module 410 is used to perform target detection on the image acquired by the vehicle-mounted image acquisition device within a first distance range through a pre-trained long-distance optical cable route identifier detection model, and obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0179] The location scoring module 420 is used to obtain the second distance inspection location and the location score of the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographical information.

[0180] The sorting module 430 is used to obtain the verification order of the optical cable route identifier to be verified based on the location score;

[0181] The first navigation route acquisition module 440 is used to acquire the first vehicle navigation route when performing a second distance inspection on the optical cable route identifier to be verified, based on the verification sequence and the second distance inspection position.

[0182] The near-field detection module 450 is used to identify the target optical cable route identifier in the first image acquired at the second distance inspection position during the second distance inspection along the first vehicle navigation route, and to obtain the first detection result of the optical cable route identifier corresponding to the GPS position coordinates.

[0183] In some embodiments of this application, the optical cable route identifier location information acquisition module 410 is further configured to:

[0184] Based on the GPS location coordinates of the pre-acquired optical cable route identifier, the second vehicle navigation route is obtained during the first distance inspection;

[0185] Using the aforementioned long-distance optical cable route identifier detection model, the target optical cable identifier is detected in the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and a first confidence level is obtained in which the second image includes the optical cable route identifier;

[0186] Based on the first confidence level and the second location, obtain the GPS location coordinates of the optical cable route identifier to be verified.

[0187] In some embodiments of this application, obtaining the GPS location coordinates of the optical cable route identifier to be verified based on the first confidence level and the second location includes:

[0188] In response to the first confidence level being less than a preset confidence threshold, the GPS location coordinates of the second location match are obtained as the GPS location coordinates of the optical cable route identifier to be verified.

[0189] In some embodiments of this application, after using the long-distance optical cable route identifier detection model to perform target optical cable identifier detection on the second image acquired by the vehicle-mounted image acquisition device at a second location on the second vehicle navigation route, and obtaining a first confidence level that the second image includes an optical cable route identifier, the method further includes:

[0190] In response to the first confidence level being greater than or equal to the preset confidence threshold, the optical cable route identifier detected at the GPS location coordinates matched by the second location is incremented to obtain a second detection result of the optical cable route identifier corresponding to the GPS location coordinates matched by the second location.

[0191] In some embodiments of this application, the optical cable route identifier location information acquisition module 410 is further configured to:

[0192] Using the aforementioned long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the third image acquired by the vehicle-mounted image acquisition device at the third location, and a second confidence level is obtained in which the third image includes the optical cable route identifier to be verified.

[0193] In response to the second confidence level indicating that the third image includes an optical cable route identifier to be verified, the GPS location coordinates matching the third location are obtained based on the pre-acquired GPS location coordinates of the optical cable route identifier and the third location.

[0194] In response to the fact that the image size proportion of the optical cable route identifier to be verified included in the third image is greater than or equal to a specified proportion threshold, and the distance between the GPS location coordinates of the optical cable route identifier to be verified and the third location is less than or equal to a specified distance threshold, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the optical cable route identifier to be verified.

[0195] In some embodiments of this application, the specified percentage threshold is the average of the image size percentage of the optical cable route identifier in multiple images pre-collected by the vehicle-mounted image acquisition device; the specified distance threshold is the average of the distance between the GPS location coordinates of the optical cable route identifier in the multiple images and the acquisition location of the multiple images.

[0196] In some embodiments of this application, the long-distance optical cable route identifier detection model is trained using the following method:

[0197] The fourth image, which includes the unobstructed optical cable route identifier and is acquired within the first distance range, is labeled with the optical cable route identifier to generate the first training dataset;

[0198] For the fifth image, which includes the obstructed optical cable route identifier and is acquired within the first distance range, the optical cable route identifier and obstruction type are labeled.

[0199] A second training dataset is generated based on the labeled image obtained by marking the optical cable route identification in the fifth image.

[0200] The images of the optical cable route identifiers that are pre-collected are used to perform image information enhancement processing on the images of the occlusion objects that match the occlusion type labels to generate a third training dataset;

[0201] A long-distance optical cable route identifier detection model is trained based on the first training dataset, the second training dataset, and the third training dataset.

[0202] In some embodiments of this application, the image information enhancement processing includes:

[0203] The image of the occlusion that matches the occlusion type label is used to partially cover the pre-collected image of the optical cable route identifier.

[0204] In some embodiments of this application, obtaining the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location based on the GPS location coordinates and geographic information includes:

[0205] Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified;

[0206] The location score of the second distance inspection position is obtained based on the distance between the second distance inspection position and the inspection start point, the distance between the second distance inspection position and the optical cable route identifier to be verified, and the angle between the second distance inspection position and the optical cable route identifier to be verified.

[0207] The optical cable route identification detection device disclosed in this application is used to implement the optical cable route identification detection method described in this application. The specific implementation methods of each module of the device will not be repeated here, but can be found in the specific implementation methods of the corresponding steps in the method embodiments.

[0208] This application discloses a fiber optic cable route identifier detection device. Using a pre-trained long-distance fiber optic cable route identifier detection model, it performs target detection on images acquired by a vehicle-mounted image acquisition device within a first distance range. This eliminates some long-distance targets, allowing for accurate identification of the fiber optic cable route identifiers. For the remaining fiber optic cable route identifiers to be verified, it performs near-distance verification. Specifically, based on the GPS location coordinates and geographic information, it obtains the second distance inspection location corresponding to the fiber optic cable route identifier to be verified and its location score; based on the location score, it obtains the verification order of the fiber optic cable route identifiers to be verified; and based on the verification order and the second distance... The inspection location is used to obtain the first vehicle navigation route when performing a second distance inspection on the optical cable route identifier to be verified; the first image collected at the second distance inspection location during the second distance inspection along the first vehicle navigation route is used to identify the target optical cable route identifier, and the first detection result of the optical cable route identifier corresponding to the GPS location coordinates is obtained. By combining inspection vehicles with artificial intelligence technology to perform two-stage inspection of optical cable route identifiers, the problems of high labor costs and low inspection efficiency in the existing technology of manual inspection of optical cable routes are solved, thereby achieving cost reduction and efficiency improvement in optical cable route inspection and improving the detection accuracy of optical cable route identifiers.

[0209] Furthermore, by using image information enhancement processing methods to expand the training dataset, the generalization ability of the long-distance optical cable route identifier detection model is improved, thereby further enhancing the accuracy of long-distance optical cable route identifier detection.

[0210] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus embodiments, since they are substantially similar to the method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0211] The above provides a detailed description of the optical cable route identification detection method and device provided in this application. Specific examples have been used to illustrate the principle and implementation of this application. The description of the above embodiments is only for the purpose of helping to understand the method of this application and its core idea. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0212] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0213] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components in the electronic device according to the embodiments of this application. This application can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such a program implementing this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0214] For example, Figure 5 An electronic device is shown that can implement the methods according to this application. The electronic device may be a PC, mobile terminal, personal digital assistant, tablet computer, etc. The electronic device conventionally includes a processor 510 and a memory 520, and program code 530 stored in the memory 520 and executable on the processor 510, which, when executing the program code 530, implements the methods described in the above embodiments. The memory 520 may be a computer program product or a computer-readable medium. The memory 520 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. The memory 520 has a storage space 5201 for the program code 530 of a computer program for performing any of the method steps described above. For example, the storage space 5201 for the program code 530 may include various computer programs for implementing the various steps in the above methods. The program code 530 is computer-readable code. These computer programs can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, CDs, memory cards, or floppy disks. The computer program includes computer-readable code that, when executed on an electronic device, causes the electronic device to perform the method according to the above embodiments.

[0215] This application also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the optical cable route identifier detection method as described in this application.

[0216] Such a computer program product can be a computer-readable storage medium, which can have the same characteristics as... Figure 5 The memory 520 in the illustrated electronic device is similarly arranged with storage segments, storage spaces, etc. Program code can be stored, for example, in a compressed form on the computer-readable storage medium. The computer-readable storage medium is typically as shown in the reference. Figure 6 The portable or fixed storage unit is described above. Typically, the storage unit includes computer-readable code 530', which is code read by a processor and, when executed by the processor, implements the various steps of the method described above.

[0217] The terms "an embodiment," "embodiment," or "one or more embodiments" as used herein mean that a particular feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of this application. Furthermore, please note that the examples of the phrase "in one embodiment" do not necessarily all refer to the same embodiment.

[0218] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0219] In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

[0220] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for detecting optical cable route identifiers, characterized in that, The method includes: By using a pre-trained long-distance optical cable route identifier detection model, target detection is performed on images acquired by the vehicle-mounted image acquisition device within the first distance range to obtain the GPS location coordinates of the optical cable route identifier to be verified. Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified and the location score of the second distance inspection location; Based on the location score, the verification order of the optical cable route identifier to be verified is obtained; Based on the verification sequence and the second distance inspection position, obtain the first vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified; The first distance is greater than the second distance; For the first image collected at the second distance inspection position during the second distance inspection along the first vehicle navigation route, target optical cable route identification is performed to obtain the first detection result of the optical cable route identification corresponding to the GPS location coordinates; The step of obtaining the second distance inspection location and the location score of the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographical information includes: Based on the GPS location coordinates and geographic information, obtain the second distance inspection location corresponding to the optical cable route identifier to be verified; The location score of the second distance inspection position is obtained based on the distance between the second distance inspection position and the inspection start point, the distance between the second distance inspection position and the optical cable route identifier to be verified, and the angle between the second distance inspection position and the optical cable route identifier to be verified.

2. The method according to claim 1, characterized in that, The method of using a pre-trained long-distance optical cable route identifier detection model to perform target detection on images acquired by the vehicle-mounted image acquisition device within a first distance range, and obtaining the GPS location coordinates of the optical cable route identifier to be verified, includes: Based on the GPS location coordinates of the pre-acquired optical cable route identifier, the second vehicle navigation route is obtained during the first distance inspection; Using the aforementioned long-distance optical cable route identifier detection model, the target optical cable identifier is detected in the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and a first confidence level is obtained in which the second image includes the optical cable route identifier; Based on the first confidence level and the second location, obtain the GPS location coordinates of the optical cable route identifier to be verified.

3. The method according to claim 2, characterized in that, The step of obtaining the GPS location coordinates of the optical cable route identifier to be verified based on the first confidence level and the second location includes: In response to the first confidence level being less than a preset confidence threshold, the GPS location coordinates of the second location match are obtained as the GPS location coordinates of the optical cable route identifier to be verified.

4. The method according to claim 3, characterized in that, The step of using the long-distance optical cable route identifier detection model to perform target optical cable identifier detection on the second image acquired by the vehicle-mounted image acquisition device at the second location on the second vehicle navigation route, and obtaining a first confidence level that the second image includes an optical cable route identifier, further includes: In response to the first confidence level being greater than or equal to the preset confidence threshold, the optical cable route identifier detected at the GPS location coordinates matched by the second location is incremented to obtain a second detection result of the optical cable route identifier corresponding to the GPS location coordinates matched by the second location.

5. The method according to claim 1, characterized in that, The method of using a pre-trained long-distance optical cable route identifier detection model to perform target detection on images acquired by the vehicle-mounted image acquisition device within a first distance range, and obtaining the GPS location coordinates of the optical cable route identifier to be verified, includes: Using the aforementioned long-distance optical cable route identifier detection model, target optical cable identifier detection is performed on the third image acquired by the vehicle-mounted image acquisition device at the third location, and a second confidence level is obtained in which the third image includes the optical cable route identifier to be verified. In response to the second confidence level indicating that the third image includes an optical cable route identifier to be verified, the GPS location coordinates matching the third location are obtained based on the pre-acquired GPS location coordinates of the optical cable route identifier and the third location. In response to the fact that the image size proportion of the optical cable route identifier to be verified included in the third image is greater than or equal to a specified proportion threshold, and the distance between the GPS location coordinates of the optical cable route identifier to be verified and the third location is less than or equal to a specified distance threshold, the GPS location coordinates matched in the third image are used as the GPS location coordinates of the optical cable route identifier to be verified.

6. The method according to claim 5, characterized in that, The specified percentage threshold is the average percentage of the image size of the optical cable route identifier in multiple images pre-collected by the vehicle-mounted image acquisition device; the specified distance threshold is the average distance between the GPS location coordinates of the optical cable route identifier in the multiple images and the acquisition location of the multiple images.

7. The method according to claim 2 or 5, characterized in that, The long-distance optical cable route identifier detection model was trained using the following method: The fourth image, which includes the unobstructed optical cable route identifier and is acquired within the first distance range, is labeled with the optical cable route identifier to generate the first training dataset; For the fifth image, which includes the obstructed optical cable route identifier and is acquired within the first distance range, the optical cable route identifier and obstruction type are labeled. A second training dataset is generated based on the labeled image obtained by marking the optical cable route identification in the fifth image. The images of the optical cable route identifiers that are pre-collected are used to perform image information enhancement processing on the images of the occlusion objects that match the occlusion type labels to generate a third training dataset; A long-distance optical cable route identifier detection model is trained based on the first training dataset, the second training dataset, and the third training dataset.

8. A fiber optic cable route identification detection device, characterized in that, The device includes: The fiber optic cable route identifier location information acquisition module is used to perform target detection on the images acquired by the vehicle-mounted image acquisition device within a first distance range using a pre-trained long-distance fiber optic cable route identifier detection model, and to obtain the GPS location coordinates of the fiber optic cable route identifier to be verified. The location scoring module is used to obtain the second distance inspection location and the location score of the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographic information. The sorting module is used to obtain the verification order of the optical cable route identifiers to be verified based on the location score; The first navigation route acquisition module is used to acquire the first vehicle navigation route when performing the second distance inspection on the optical cable route identifier to be verified, based on the verification order and the second distance inspection position. The first distance is greater than the second distance; The close-range detection module is used to identify the target optical cable route identifier in the first image collected at the second distance inspection position during the second distance inspection along the first vehicle navigation route, and obtain the first detection result of the optical cable route identifier corresponding to the GPS position coordinates; The location scoring module is also used to obtain the second distance inspection location corresponding to the optical cable route identifier to be verified based on the GPS location coordinates and geographic information. The location score of the second distance inspection position is obtained based on the distance between the second distance inspection position and the inspection start point, the distance between the second distance inspection position and the optical cable route identifier to be verified, and the angle between the second distance inspection position and the optical cable route identifier to be verified.

9. An electronic device, comprising a memory, a processor, and program code stored in the memory and executable on the processor, characterized in that, When the processor executes the program code, it implements the optical cable route identifier detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium having program code stored thereon, characterized in that, When the program code is executed by the processor, it implements the steps of the optical cable route identification detection method according to any one of claims 1 to 7.

Images