Artificial intelligence rapid identification and positioning method and device for operators in runway, slideway area and airport apron of transportation airport

Abstract

The invention provides an artificial intelligence rapid identification and positioning method and device for operators in a transport airport runway, a slideway area and an airport apron, and the method comprises the steps: collecting an image through an AI camera, and transmitting the image to artificial intelligence judgment software, so as to extract the identification features of the operators in the image; and comparing and identifying with a pre-trained reflective clothing feature model of different post personnel to obtain post information of the operating personnel, generating a comprehensive report after deep data analysis and processing, and synchronously sending the report to an airport apron monitoring large screen and an airport operation control and airport ground support operation management department in real time. The method can quickly confirm the unit and the operation qualification of the airport, improves the operation safety of the airport, effectively solves the problems of flight guarantee and reasonable personnel allocation, can quickly dispatch the handling personnel for the emergencies on the runway, the slideway area and the airport apron, provides powerful support for the overall operation management of the airport, and improves the safety of the airport. The report is also used as a data basis for tracing accidents in the future.

Description

Technical Field

[0001] The present invention relates to the field of airport management technology using image recognition and artificial intelligence in the field of aviation transportation, and in particular to a method and system for rapid artificial intelligence identification and positioning of runway, taxiway area and apron operators at a transport airport. Background Art

[0002] According to the "Regulations on the Operation and Management of Aprons at Transport Airports" (Civil Aviation Regulations

[2022] No. 20), workers in the runway and taxiway areas must wear reflective clothing of different colors. In the traditional airport runway, taxiway area and apron management model, the visual observation of supervisors is mainly relied on to identify job operators and qualifications and other information. This traditional manual identification method has many limitations, such as slow speed, low efficiency, poor accuracy, and great subjective influence. It is impossible to quickly identify the attributes of operators and confirm their qualifications at a long distance, which is extremely unfavorable for the efficient operation and safety of the airport. At the same time, the runway intrusion risk caused by unrelated personnel mistakenly entering the ground protection area for aircraft landing and takeoff is difficult to effectively prevent. During the flight guarantee process, it is also impossible to quickly and accurately dispatch and arrange personnel, which is not conducive to the timely handling of emergencies on the apron. This poses a challenge to modern airport management. The Chinese patent application with publication number CN117671727A proposes a multi-feature fusion reflective vest wearing detection method. This method is to fuse the acquired pixels of the target personnel with the corresponding pixels of the reflective vest to obtain the corresponding pixels of the reflective vest area where the target personnel wears; judge whether the reflective vest is worn based on the corresponding pixels of the reflective vest area where the target personnel wears, and complete the reflective vest wearing detection with multi-feature fusion. This method only collects and identifies the image information of the reflective vest of the operator, and cannot determine the corresponding relationship between the reflective vest and the operator. Due to the multi-fusion and complexity of modern airport transportation operations, airport operations cannot be guaranteed to be carried out according to the timeline. Operators in multiple fields often work at the same time, causing the operations to overlap. Therefore, it is necessary to combine operators in different professional fields, the identity information of each operator and wearable devices for detection to meet the requirements of modern airport operations for operator identification. Obviously, this cannot be achieved through the method of publication number CN117671727A. Summary of the invention

[0003] To solve the problems in the prior art, the object of the present invention is to provide an artificial intelligence rapid identification and positioning method and device for operating personnel in the runway, taxiway area and apron of a transportation airport, aiming at the operation mode of superposition of operating personnel in multiple fields, so as to improve the recognition speed and accuracy of operating personnel in the runway, taxiway area and apron, solve the problem that the attributes of operating personnel cannot be quickly identified and the operation qualifications cannot be confirmed at a long distance in the existing airport operation, effectively eliminate the runway incursion and the risks of invading the runway, taxiway area and apron, improve the safety of airport operation, realize the reasonable scheduling and arrangement of flight guarantee personnel, and ensure the rapid and accurate recording and management of personnel information, etc. It can effectively improve the A-CDM decision-making ability of the airport.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] The object of the first aspect of the present invention is to provide an artificial intelligence rapid identification and positioning method for operating personnel in the runway, taxiway area and apron of a transportation airport, and the method includes the following steps:

[0006] Step 1, AI camera image acquisition and positioning: The AI camera captures images of operating personnel in the runway, taxiway area and apron in real time to obtain the image information of the operating personnel, and the image information includes the reflective clothing colors, sizes, light-emitting strips, printed words and identification features worn by operating personnel in different positions. The satellite locator is used to realize the positioning of the operating personnel;

[0007] Step 2, image transmission to the artificial intelligence judgment software: The acquired image information is synchronously transmitted to the artificial intelligence judgment software in real time through the multi-camera communication system;

[0008] Step 3, image feature extraction: The artificial intelligence judgment software extracts the reflective clothing colors, sizes, light-emitting strips, printed words and identification features of the operating personnel in the image;

[0009] Step 4, comparison and recognition with the pre-trained model: The extracted features are compared and recognized with the pre-trained reflective clothing feature models of different positions of personnel, so as to recognize the position information of the operating personnel;

[0010] Step 5, real-time synchronous transmission of the recognition result to the data analysis system: The recognition result of the position information of the operating personnel is synchronously transmitted to the data analysis system in real time for further processing;

[0011] Step 6, evaluation and optimization of the model: In order to improve the accuracy and robustness of artificial intelligence rapid identification, the model is evaluated and optimized respectively, and the model is further optimized through the evaluation results. More accurately determine the information of operating personnel in each position within a certain working area, and judge whether the affiliated unit, working position and personal information of the operating personnel are consistent with the pre-set ones;

[0012] Step 7, generate a comprehensive report: Generate a comprehensive report based on the analysis results, including the personnel distribution, operation status, and arrival status of each position.

[0013] Step 8, real-time synchronous transmission of the report through a wired or wireless remote transmission system: The comprehensive report is real-time synchronously sent to the apron monitoring large screen, as well as the airport operation control and airport ground support operation management departments through a wired or wireless remote transmission system, providing a basis for management decisions. The report will also be stored as a data basis for future accident tracing.

[0014] Furthermore, in Step 1, the image information collected by the AI camera includes the image information collected in chronological order and / or the image information collected according to the superimposed time.

[0015] Furthermore, in Step 3, the image feature extraction is obtained through the following methods: A large amount of data is collected according to different information, and collected and labeled with diverse information such as different angles and lighting conditions to clarify the position, reflective clothing color, size, luminous strips, printed text, and identification information of the operating personnel in each image. Then, data augmentation, rotation and scaling, brightness and contrast adjustment, and noise addition are performed on the images to improve the robustness of the model to noise. On this basis, the following image morphological feature extractions are realized: Shape feature extraction of the reflective clothing is obtained by using edge detection and contour extraction methods; Color feature extraction: Extract the color histogram and color moment features of the reflective clothing; Texture feature extraction: Extract the texture features of the reflective clothing through the gray-level co-occurrence matrix and Gabor filter methods; Text and identification recognition: Use OCR technology to recognize the printed text and identification and extract relevant features.

[0016] Furthermore, in Step 4, the pre-trained model is obtained through the following methods: Using the convolutional neural network CNN: Select a CNN architecture suitable for image recognition, including the following steps:

[0017] Through transfer learning: Use the pre-trained CNN model for transfer learning and fine-tune the network parameters to adapt to specific tasks;

[0018] Multi-task learning: Simultaneously train multiple related tasks, share the underlying feature representation, and improve the overall performance; Multiple related tasks include but are not limited to position recognition and color recognition;

[0019] Iterative optimization: Iteratively optimize the model according to user feedback and actual application effects.

[0020] Furthermore, in Step 6, the method for in-depth data analysis and processing includes the following steps:

[0021] Cross-validation: Use the K-fold cross-validation method to evaluate the generalization ability of the model;

[0022] Hyperparameter Tuning: Adjust the hyperparameters of the model through methods such as grid search and random search. The hyperparameters include but are not limited to the learning rate and batch size.

[0023] Loss Function Optimization: Select the cross-entropy loss or Focal Loss function and optimize it to improve the model performance;

[0024] Adopt the method of ensemble learning;

[0025] Utilize model fusion: Integrate multiple models with different architectures or different training strategies to improve the overall prediction accuracy;

[0026] Continuously perform continuous learning and updating;

[0027] Online Learning: As new data is continuously collected, adopt the online learning strategy to continuously update the model;

[0028] Regular Retraining: Regularly retrain the model with the latest data to adapt to the changes in airport operators;

[0029] Implement a feedback loop through usage;

[0030] User Feedback: Establish a user feedback mechanism to collect error cases and improvement suggestions in actual use.

[0031] The purpose of the second aspect of the present invention is to provide an artificial intelligence rapid identification device for airport runway, taxiway area and apron operators to implement the method of the present invention, including the following components:

[0032] An image information acquisition system for obtaining the image information of operators by capturing images of the runway, taxiway area and apron operators;

[0033] An image feature extraction system for extracting the image features of the captured operators;

[0034] An image feature recognition system for comparing and recognizing the extracted features with the pre-trained reflective clothing feature models of operators in different positions, so as to identify the position information of the operators;

[0035] A data analysis and processing system: For in-depth analysis of the recognition result data, determine the affiliated units of operators in each position within a certain working area, and judge whether the affiliated units, working positions and personal information of the operators are consistent with the pre-set ones;

[0036] A comprehensive report formation system for generating a comprehensive report based on the analysis results of the data analysis and processing system;

[0037] A wired or wireless remote transmission system for real-time synchronous remote transmission of the generated comprehensive report;

[0038] A memory for storing data of the device.

[0039] Furthermore, the image information acquisition system of the present invention includes:

[0040] AI cameras: responsible for capturing images of the runway, taxiway areas, and apron operating personnel, distributed in various key areas of the apron, including but not limited to positions, areas around the maneuvering area, and passenger service areas;

[0041] Pan-tilt heads: connected to the AI cameras, capable of adjusting the camera angles to ensure full coverage of the apron area;

[0042] Auto-zoom lenses: adapted to the AI cameras, capable of clearly capturing operating personnel at different distances;

[0043] Multi-camera communication systems: aggregating the image data captured by each AI camera, and achieving stable data transmission by combining wired and wireless methods;

[0044] A satellite locator for determining the location distribution of operating personnel.

[0045] Furthermore, the present invention further includes receiving ends such as the airport operation control center, etc., for receiving the information processed by the wired or wireless remote transmission system. The receiving ends such as the airport operation control center include but are not limited to the terminal devices of the airport command department, ground support operation management department, and safety management department.

[0046] The beneficial effects of the present invention are as follows: By collecting and identifying the information of different professional operating personnel in multiple fields working simultaneously, the present invention solves the problem of quickly identifying superimposed operating personnel, and at least includes the following technical effects.

[0047] Improve supervision efficiency: Apron supervisors can quickly confirm their affiliated units and operation qualifications through external images, which brings convenience to supervising whether operating personnel operate according to procedures and specifications.

[0048] Enhance safety guarantee: Effectively eliminate the runway incursion risk of unauthorized personnel entering the ground protection areas for aircraft landing and takeoff, and improve the safety of airport operations.

[0049] Optimize personnel scheduling: The airport operation control department, airport ground support operation management department, and apron supervisors can schedule and arrange flight support personnel according to the real-time operating personnel information provided by the system, effectively solve the problems of flight support and reasonable personnel allocation, and at the same time can quickly mobilize the personnel for handling emergencies on the apron.

[0050] Improve decision-making ability: Through image AI recognition, improve the decision-making ability of airport A-CDM, provide strong support for the overall operation management of the airport, and the relevant information reports will also be used as the data basis for tracing accidents in the future. Brief Description of the Drawings

[0051] Figure 1 It is a flowchart of the method of the present invention.

[0052] Figure 2 It is a block diagram of the composition of the device of the present invention. Detailed Embodiments

[0053] Referring to the attached Figure 1 drawings, the main processes of the method of the present invention include:

[0054] 1. AI camera image acquisition and positioning: The AI camera captures images of the runway, taxiway area and apron operating personnel in real time. Obtain the image information of the operating personnel, which includes the reflective clothing colors, sizes, light-emitting strips, printed texts and identification features worn by operating personnel in different positions, and use a satellite locator to achieve the positioning of the operating personnel.

[0055] 2. Image transmission to the artificial intelligence judgment software: The captured image information is synchronously transmitted to the artificial intelligence judgment software in real time through a multi-camera communication system.

[0056] 3. Extraction of image features such as color, size, and identification: The artificial intelligence judgment software extracts features such as the reflective clothing color, size, light-emitting strip, printed text and identification worn by the operating personnel in the image.

[0057] 4. Comparison and recognition with the pre-trained model: Compare and recognize the extracted features with the pre-trained reflective clothing feature models of different position personnel, so as to recognize information such as the positions of the operating personnel.

[0058] 5. Real-time synchronous transmission of the recognition result to the data analysis system: The recognition result of the operating personnel position information is synchronously transmitted to the data analysis system in real time for further processing.

[0059] 6. Evaluation and optimization of the model: In order to improve the accuracy and robustness of the artificial intelligence's rapid identification, the data analysis system deeply analyzes the recognition result data, evaluates and optimizes the model respectively, and further optimizes the model through the evaluation results. More accurately determine the information of operating personnel in each position within a certain working area, and judge whether the affiliated unit, working position, and personal information of the operating personnel are consistent with the pre-set ones.

[0060] 7. Generate a comprehensive report: Generate a comprehensive report according to the analysis results, and the content may include the distribution of personnel in each position, the operation situation, and the arrival situation.

[0061] 8. The report is sent through a wired or wireless remote transmission system: The comprehensive report is synchronously sent to the apron monitoring large screen, the airport operation control department, the airport ground support operation management department and other relevant departments in real time through the wired or wireless remote transmission system, providing a basis for management decision-making.

[0062] In step 1 of the present invention, the image information collected by the AI camera includes the image information collected in chronological order and / or the image information collected according to the superimposed time. Due to the tight time requirements for airport operations, the operating personnel of each process and position may work simultaneously. The image information of the operating personnel in the same process position is collected in chronological order, and the image information of the operating personnel in different process positions shows a time superposition situation after collection. In subsequent image recognition and processing, real-time synchronous processing can be performed on the image information collected in chronological order and / or the image information collected according to the superimposed time, so that the image processing result matches the actual working conditions of the airport, and the obtained result is more accurate.

[0063] Based on the above method, the present invention uses computer vision technology to collect digital images of the support operating personnel in the airport working area through a video digital camera. According to multiple information such as the different colors, sizes, reflective strips, printed words and logos of the reflective clothing worn by the operating personnel, the personnel in positions serving passengers, positions operating in the aircraft position, regulatory positions in the runway and taxiway areas, positions working in the maneuvering area and its surrounding areas, etc. are identified in real time and quickly. The general recognition function is completed by using software algorithms such as convolutional neural networks in deep learning or by calling the API of the AI big data model, and accurate judgment is achieved by adopting a training optimization strategy. For example, by extracting the color features and shape features of the reflective clothing and combining the model trained by machine learning algorithms, the personnel in different positions can be accurately judged. For printed words and logos, OCR (Optical Character Recognition) technology is used for recognition. The affiliated unit and personnel information are further determined by using the time process coordinates and multiple information such as the arrival time requirements of the staff.

[0064] In an embodiment of the present invention, in step 3, the image feature extraction is obtained as follows: A large amount of data is collected according to different information, and the data is collected and labeled with diverse information such as different angles and lighting conditions. The position, reflective clothing color, size, reflective strips, printed text, and identification information of the operators in each image are clarified. Then, data augmentation, rotation and scaling, brightness and contrast adjustment, and noise addition are performed on the images to improve the model's robustness to noise. On this basis, the following image morphological feature extractions are realized: Shape feature extraction of the reflective clothing is obtained by using methods such as edge detection and contour extraction; Color feature extraction: The color histogram and color moment features of the reflective clothing are extracted; Texture feature extraction: The texture features of the reflective clothing are extracted by using methods such as gray-level co-occurrence matrix and Gabor filter; Text and identification recognition: OCR technology is used to recognize the printed text and identification, and relevant features are extracted.

[0065] In an embodiment of the present invention, in step 4, the pre-trained model is obtained as follows: Using a convolutional neural network (CNN): Select a CNN architecture suitable for image recognition, such as ResNet, VGG, etc. It includes the following steps:

[0066] Through transfer learning: Use a pre-trained CNN model (such as a model pre-trained on ImageNet) for transfer learning, and fine-tune the network parameters to adapt to specific tasks.

[0067] Multi-task learning: Simultaneously train multiple related tasks (such as position recognition, color recognition, etc.), share the underlying feature representation, and improve the overall performance.

[0068] Iterative optimization: Iteratively optimize the model according to user feedback and actual application effects.

[0069] Through the above pre-training method, the accuracy and robustness of the artificial intelligence judgment system can be significantly improved, ensuring efficient and accurate operator recognition in a complex and changing airport environment.

[0070] In an embodiment of the present invention, in step 6, the method for in-depth data analysis and processing includes the following steps;

[0071] Cross-validation: Use methods such as K-fold cross-validation to evaluate the generalization ability of the model.

[0072] Hyperparameter tuning: Adjust the hyperparameters of the model, such as learning rate, batch size, etc., through methods such as grid search and random search.

[0073] Loss function optimization: Select a suitable loss function (such as cross-entropy loss, Focal Loss, etc.), and optimize it to improve the model performance.

[0074] Adopt the method of ensemble learning;

[0075] Utilize model fusion: Integrate multiple models with different architectures or different training strategies, such as voting, weighted averaging, etc., to improve the overall prediction accuracy.

[0076] Continuously perform continuous learning and updating;

[0077] Online learning: As new data is continuously collected, adopt an online learning strategy to continuously update the model.

[0078] Regular retraining: Regularly retrain the model with the latest data to adapt to changes in airport operators.

[0079] Implement a feedback loop by using

[0080] User feedback: Establish a user feedback mechanism to collect error cases and improvement suggestions in actual use.

[0081] Based on the method for rapid identification and positioning of airport runway, taxiway area and apron operators in the embodiments of the present invention, the present invention also provides a device for implementing this method, including the following components:

[0082] An image information acquisition system for obtaining the image information of operators by capturing images of the runway, taxiway area and apron operators;

[0083] An image feature extraction system for extracting the image features of the captured operators;

[0084] An image feature recognition system for comparing and recognizing the extracted features with the pre-trained reflective clothing feature models of operators in different positions, so as to identify the position information of the operators;

[0085] A data analysis and processing system: used for in-depth analysis of the recognition result data to determine the affiliated units of operators in a certain working area, and to judge whether the affiliated units, working positions and personal information of the operators are consistent with the pre-set ones;

[0086] A comprehensive report formation system for generating a comprehensive report based on the analysis results of the data analysis and processing system;

[0087] A wired or wireless remote transmission system for real-time synchronously remotely transmitting the generated comprehensive report;

[0088] A memory for storing the data of the device.

[0089] Based on the above method, the present invention realizes the safety management of operating personnel by collecting data of airport operating personnel and further forming a comprehensive report after analysis and processing. At the same time, the present invention can also use a satellite locator to locate operating personnel, and through the combination of a satellite positioning system and an image information collection system, it can more accurately realize the rapid identification and positioning of operating personnel.

[0090] See the appendix Figure 2 , the main components of the device of the present invention include:

[0091] 1. AI cameras (A, E, H, etc.): Responsible for taking pictures of the runway, taxiway area and apron operating personnel images, distributed in various key areas of the apron, such as aircraft positions, around the maneuvering area, passenger service areas, etc.

[0092] 2. Pan-tilt heads (C, F, I, etc.): Connected to the AI cameras, can adjust the camera angle to ensure full coverage of the apron area.

[0093] 3. Auto-focus lenses (D, G, J, etc.): Adapted to the AI cameras, can clearly take pictures of operating personnel at different distances.

[0094] 4. Multi-camera communication system (B): Aggregates the image data taken by each AI camera and transmits it to the subsequent system, using a combination of wired and wireless methods to ensure stable data transmission.

[0095] 5. Apron monitoring large screen (K): Installed in a prominent position in the airport supervision center, used to display operating personnel identification information, etc., for the convenience of supervisors to view.

[0096] 6. Image feature extraction system, used to extract the image features of operating personnel collected.

[0097] 7. Image feature recognition system, used to compare and recognize the extracted features with the pre-trained reflective clothing feature models of different positions of personnel, so as to identify the position information of operating personnel.

[0098] 8. Data analysis and processing system, used to deeply analyze the recognition result data, determine the affiliated units of operating personnel in each position within a certain working area, and judge whether the affiliated units, working positions, and personal information of operating personnel are consistent with the pre-set ones.

[0099] 9. Comprehensive report formation system, used to generate a comprehensive report based on the analysis results of the data analysis and processing system.

[0100] 10. Wired or wireless remote transmission system (L): The processed information is transmitted in real time and synchronously to the receiving end (M) such as the airport operation control center, such as the airport command department, ground support operation management department and other relevant management departments, to achieve information sharing. The receiving end (M) of the airport operation control center includes but is not limited to the terminal equipment of the airport command department, ground support operation management department and security management department.

[0101] In addition to the above main components, the device of the present invention is also provided with a power supply, artificial intelligence rapid recognition software, a memory, etc.

[0102] Based on the above system, the AI camera of the present invention is equipped with an automatic zoom lens to shoot different areas of the apron, and the pan / tilt can adjust the camera angle to ensure full coverage. The captured image is transmitted to the artificial intelligence judgment software through a multi-camera communication system. The software analyzes and identifies the image and transmits the identification result to the data analysis system. The data analysis system further processes the data, generates a comprehensive report, and then transmits it to the apron monitoring screen and relevant departments such as the airport operation control department and the airport ground support operation management through a wireless remote transmission system. The comprehensive report generated at the same time is stored in the memory for data preservation and future use in tracing accidents.

[0103] Based on the above implementation methods, the main innovations of the present invention are: 1. Computer images are targeted to achieve image enhancement and other processing according to the airport working environment, lighting, and sky color, so as to minimize interference information, extract relevant information of ground support operations, and improve signal-to-noise ratio and image quality. 2. Multi-information fusion technology is used to improve image accuracy recognition. Multi-information fusion such as color spectrum, wearing shape, reflective tape, posture, LOGO pattern, text, time series, scene superposition, and multi-camera information sharing is used to realize staff recognition. 3. Artificial intelligence algorithm is iteratively optimized. Artificial intelligence training is performed on the image recognition algorithm model through big data, machine learning, manual training and other technologies to iteratively optimize the software algorithm model. 4. Big data analysis and scientific management strategies are established, a set of data analysis systems, comprehensive report formation systems and operation management strategies are developed, and wired or wireless remote transmission is carried out to provide accurate real-time data and records for management personnel such as towers, operation command departments, and airport ground support operation management departments, so as to realize the reasonable scheduling and scientific management of support workers and provide strong support for the improvement of airport A-CDM decision-making capabilities. 5. The automatic controllable hardware system provides multi-directional and effective image acquisition. The automatic zoom lens can track images at a long distance and adjust the visual range of the workplace. The controllable pan / tilt can capture images from multiple angles, and the camera positions and numbers are reasonably arranged to achieve full-field scanning.

[0104] The application of the present invention in practical scenarios is:

[0105] Hardware Installation: Install AI cameras reasonably at key positions in the airport, such as runways, taxiways, various aircraft stands, around the maneuvering area, and passenger service areas. Connect and fix the pan-tilt head to the camera to ensure that the camera can adjust the angle flexibly. Install an auto-zoom lens and make it compatible with the camera to ensure that it can clearly capture operators at different distances. The multi-camera communication system uses a combination of wired and wireless methods to ensure stable data transmission. The apron monitoring large screen is installed in the airport supervision center or the ground support operation management department for convenient viewing by supervisors, and the stored reports will also be used as data basis for tracing accidents in the future.

[0106] Software Settings: Initialize the artificial intelligence judgment software and import the pre-trained models of the reflective clothing characteristics of personnel in different positions, including characteristic parameters such as color, size, and text markings. Associate the data analysis system with the artificial intelligence judgment software and set up the data transmission interface and analysis rules. The comprehensive report formation system sets up the report template and generation rules according to the airport management requirements. The wired or wireless remote transmission system sets up the transmission frequency and the receiving end address to ensure that data can be transmitted to relevant departments such as the airport operation control center in real-time and accurately.

[0107] System Operation: When the system starts, the AI camera begins to capture real-time images of the runway, taxiway area, and apron operators. After being optimized by the auto-zoom lens, the images are transmitted to the artificial intelligence judgment software. The software analyzes and identifies the images, and transmits the identification results to the data analysis system in real-time. The data analysis system processes them according to the set rules, generates a comprehensive report, and sends it to the apron monitoring large screen, the airport operation control department, and the ground support operation management department through the wired or wireless remote transmission system. Apron supervisors and airport operation control center staff perform corresponding management operations based on the received information.

[0108] It should be further noted that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. The technical features involved, including the composition structure and implementation methods, can be selected and replaced on the basis of the understanding of those skilled in the art in accordance with the technical concept of the present invention, as long as they can solve the technical problems of the present invention and achieve the technical effects of the present invention, which are in line with the purpose of the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall also be included within the protection scope of the present invention.

Claims

1. An artificial intelligence rapid identification and positioning method for operating personnel in the runway, taxiway area and apron of a transportation airport, characterized in that It includes the following steps: Step 1, AI camera image acquisition and positioning: The AI camera captures images of the runway taxiway area and apron operating personnel in real time to obtain the image information of the operating personnel. This image information includes the reflective clothing colors, sizes, light-emitting strips, printed texts, and identification features worn by operating personnel in different positions. The satellite locator is used to achieve the positioning of the operating personnel; Step 2, Image transmission to the artificial intelligence judgment software: The captured image information is synchronously transmitted to the artificial intelligence judgment software in real time through the multi-camera communication system; Step 3, Image feature extraction: The artificial intelligence judgment software extracts the reflective clothing colors, sizes, light-emitting strips, printed texts, and identification features of the operating personnel in the image; Step 4, Comparison and recognition with the pre-trained model: The extracted features are compared and recognized with the pre-trained reflective clothing feature models of different position personnel, so as to recognize the position information of the operating personnel; Step 5, Real-time synchronous transmission of the recognition result to the data analysis system: The recognition result of the operating personnel position information is synchronously transmitted to the data analysis system in real time for further processing; Step 6, Evaluation and optimization of the model: The data analysis system conducts in-depth analysis on the recognition result data, evaluates and optimizes the model respectively, more accurately determines the information of operating personnel in each position within a certain working area, and judges whether the affiliated unit, working position, and personal information of the operating personnel are consistent with the pre-set ones; Step 7, Generate a comprehensive report: Generate a comprehensive report according to the analysis results, and the content includes the distribution of personnel in each position, operation status, and arrival status; Step 8, The report is sent through the wired or wireless remote transmission system: The comprehensive report is synchronously sent to the runway taxiway area and apron monitoring large screen, as well as the airport operation control and airport ground support operation management departments in real time through the wired or wireless remote transmission system, providing a basis for management decision-making.

2. The method according to claim 1, wherein: In Step 1, the image information collected by the AI camera includes the image information collected in chronological order and / or the image information collected according to the superimposed time.

3. The method according to claim 1, wherein: In Step 3, the image feature extraction is obtained through the following methods: A large amount of data is collected according to different information, and the data is collected and labeled with diverse information in different angles and lighting conditions to clarify the positions, reflective clothing colors, sizes, light-emitting strips, printed texts, and identification information of the operating personnel in each image. Then, data augmentation, rotation and scaling, brightness and contrast adjustment, and noise addition are performed on the images to improve the robustness of the model to noise. On this basis, the following image morphological feature extraction is realized: The shape features of the reflective clothing are obtained by using the methods of edge detection and contour extraction; Color feature extraction: Extract the color histogram and color moment features of the reflective clothing; Texture Feature extraction: The texture features of the reflective clothing are extracted through the gray-level co-occurrence matrix and Gabor filter methods; Text and identification recognition: The OCR technology is used to recognize the printed texts and identifications and extract relevant features.

4. The method according to claim 1, wherein: In Step 4, the pre-trained model is obtained through the following methods: Using the convolutional neural network CNN: Select a CNN architecture suitable for image recognition, including the following steps: Through transfer learning: Utilize a pre-trained CNN model for transfer learning and fine-tune network parameters to adapt to specific tasks; Multi-task learning: Simultaneously train multiple related tasks, share underlying feature representations, and improve overall performance; The multiple related tasks include, but are not limited to, job identification and color identification; Iterative optimization: Iteratively optimize the model based on user feedback and actual application effects.

5. The method according to claim 1, characterized in that: In step 6, the method for in-depth data analysis and processing includes the following steps: Cross-validation: Use the K-fold cross-validation method to evaluate the generalization ability of the model; Hyperparameter tuning: Adjust the hyperparameters of the model through methods such as grid search and random search. The hyperparameters include, but are not limited to, the learning rate and batch size. Loss function optimization: Select the cross-entropy loss or FocalLoss function and optimize it to improve the model performance; Adopt the method of ensemble learning; Utilize model fusion: Fuse multiple models with different architectures or different training strategies to improve the overall prediction accuracy; Continuously perform continuous learning and update; Online learning: As new data is continuously collected, adopt an online learning strategy to continuously update the model; Regular retraining: Regularly retrain the model with the latest data to adapt to changes in airport operators; Implement a feedback loop by using; User feedback: Establish a user feedback mechanism to collect error cases and improvement suggestions in actual use.

6. An artificial intelligence rapid identification and positioning device for operating personnel in the runway, taxiway area and apron of an airport, which implements the method according to any one of claims 1-5, characterized in that, Include the following components: An image information acquisition system for obtaining image information of operators in the runway taxiway area and apron by capturing images of the runway taxiway area and apron operators; An image feature extraction system for extracting the image features of the captured operators; An image feature recognition system for comparing and recognizing the extracted features with the pre-trained reflective clothing feature models of operators in different positions to identify the position information of the operators; A data analysis and processing system for deeply analyzing the recognition result data to determine the affiliated units of operators in each position within a certain working area and judge whether the affiliated units, working positions, and personal information of the operators are consistent with the pre-set ones; A comprehensive report generation system for generating a comprehensive report based on the analysis results of the data analysis and processing system; A wired or wireless remote transmission system for real-time synchronously remotely transmitting the generated comprehensive report; A memory for storing the generated comprehensive report.

7. The device according to claim 6, characterized in that, The image information acquisition system includes: AI cameras: Responsible for capturing images of the runway taxiway area and apron operators, distributed in various key areas of the runway taxiway area and apron, including but not limited to aircraft positions, around the maneuvering area, and passenger service areas; Pan-tilt heads: Connected to the AI cameras, they can adjust the camera angles to ensure full coverage of the runway taxiway area and apron area; Auto-zoom lenses: Adapted to the AI cameras, they can clearly capture operators at different distances; A multi-camera communication system: Aggregates the image data captured by each AI camera and realizes stable data transmission by combining wired and wireless methods.

8. The device according to claim 6, characterized in that: It further includes an airport operation control center receiving end for receiving the information processed by the wired or wireless remote transmission system, and the airport operation control center receiving end includes, but is not limited to, the terminal devices of the airport command department, the ground support operation management department, and the security management department.

Images