Integrated boarding island device and method for improving passenger passing efficiency

Abstract

The invention discloses an integrated boarding island device and method for improving passenger passing efficiency. The passenger passing efficiency is improved to the maximum extent through a multi-time pre-shooting technology. The integrated boarding island device comprises a gate machine, a flexible door wing, a portal frame and a snapshot stand column, and the gate machine is combined with the flexible door wing, can be kept in an open state under the normal condition and is only closed under the abnormal condition. A hidden multi-angle camera is arranged on the portal frame, high-quality face snapshot can be carried out on teams meeting requirements at the same time, and optimal face images are screened out through an algorithm for comparison. The snapshot stand column works independently and is responsible for identifying special passengers and guiding the special passengers to carry out manual verification. The snapshot stand column extends out of the gate and can be planned and extended into a linear or L-shaped passage area according to a field area, so that the passage order of passengers is effectively standardized, and meanwhile, the subsequent snapshot quality of multiple cameras is improved. The traffic efficiency of passengers can be improved to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of turnstiles, and particularly to an integrated boarding island device and method that can maximize the passenger passing efficiency. Background Art

[0002] With the continuous growth of air transportation demand, the improvement of passenger passing efficiency has become an important issue in the aviation industry. Traditional boarding methods often lead to passengers staying in front of the boarding gate, resulting in low passing efficiency and poor passenger travel experience. There are problems such as recognition delay and long queuing time in the existing technology during the passenger passing process.

[0003] In view of this, the present invention deeply conceives in response to the many deficiencies and inconveniences caused by the imperfect structural design of the existing tea-making furnace, and actively researches, improves and tries to develop the present invention. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the existing technology, and provide an integrated boarding island device and method that can maximize the passenger passing efficiency, improve the passenger passing efficiency and passing safety, and enhance the turnstile security.

[0005] To achieve the above purpose, the solution of the present invention is: An integrated boarding island device for improving passenger passing efficiency, which includes a turnstile, a flexible door wing, a gantry and a capture column; The flexible door wing is arranged at the tail of the turnstile, and the flexible door wing and the turnstile remain in an open state under normal circumstances; The gantry is installed on the turnstile, and multiple concealed cameras are arranged on the gantry to capture the faces of passengers from different angles. Multiple cameras capture at the same time, and the best-quality face image is intercepted by using a multi-image fusion algorithm to identify the identity of the passengers; The capture column is arranged in front of the turnstile, and extends with the turnstile to form a straight-line or L-shaped passing area. A camera and an interactive screen are installed on the capture column. The image recognition algorithm and the special passenger database built in the capture column are used to identify special passengers, and the special passengers are guided to the manual verification channel through the interactive screen.

[0006] Furthermore, the capture column is connected to the passenger identity verification system and the manual verification channel management system of the turnstile for data interaction and collaborative work. The passenger image data captured by the capture column is transmitted to the identity verification system for simultaneous processing with the multiple face images captured by the cameras on the gantry; when the special passenger information identified by the capture column is transmitted to the manual verification channel management system.

[0007] Furthermore, the steps of intercepting the best-quality face image by the multi-image fusion algorithm are: Step S1: When a passenger appears within the shooting range of the gantry cameras, multiple cameras work simultaneously to capture the passenger's face images from different angles, obtaining multiple sets of image data of the same face from different perspectives. Step S2: Preprocess each captured face image. Step S3: Extract features from each of the preprocessed face images respectively. Step S4: Conduct quality assessment on each face image according to the extracted features, and assign weights to each face image based on the quality assessment results of the face images. Step S5: Perform face image fusion according to the assigned weights. Step S6: Conduct quality assessment on the fused face image to determine whether the fused face image meets the best quality standard. If the quality of the fused face image does not meet the requirements, re - conduct the fusion process. Step S7: According to the quality assessment results in Step S6, intercept the part of the face image with the best quality from the fused face image or the original multiple face images. Determine the position and size of the face in the image according to the face detection algorithm, and then completely intercept the face area, removing irrelevant background information to obtain the best - quality image containing only the face.

[0008] Furthermore, in Step S2, the image preprocessing includes grayscale conversion, noise reduction, and normalization operations. The grayscale conversion converts the color face image into a grayscale image; the noise reduction process uses a combination of mean filtering and median filtering methods to remove the noise interference in the face image and improve the image quality; the normalization adjusts the pixel values of the face image to a specific range.

[0009] Furthermore, in Step S3, the feature extraction uses a convolutional neural network based on deep learning, and the extracted features are used to reflect the key information of the face.

[0010] Further, in step S4, the image quality is evaluated based on the clarity of the face image combined with the face posture. The clarity of the face image is analyzed using any one of the algorithms of the gradient value method, the Laplace operator method or the spectrum analysis method. The gradient value method calculates the gradient amplitude of the face image by finding the gradient of the face image in the horizontal and vertical directions and then calculating its amplitude. The larger the gradient amplitude, the richer the edges and details of the face image and the higher the clarity. The Laplace operator method uses the Laplace operator to perform a convolution operation on the face image and calculate the second-order derivative of the face image. The larger the absolute value of the result, the more drastic the change in the details of the face image and the higher the clarity. The spectrum analysis method performs a Fourier transform on the face image to convert the face image from the spatial domain to the frequency domain, and the high-frequency components The more, the richer the details and edge information of the face image, and the higher the clarity; the spectrum analysis method evaluates the image clarity by analyzing the energy proportion of the high-frequency part in the frequency domain; the face posture assessment adopts the face key point detection method or the posture estimation algorithm. The face key point detection method uses the face key point detection algorithm to locate the key feature points of the face, and judge whether the face posture is normal according to the position and distribution of these key points; the posture estimation algorithm adopts a model-based posture estimation algorithm, which matches the detected facial features with the 3D face model based on the 3D face model, and calculates the posture parameters of the face rotation angle, pitch angle and yaw angle. The face image quality is relatively good when the posture parameters are within a certain range. If they are out of the range, the face recognition effect is affected and the quality is relatively low.

[0011] Furthermore, step S5 fuses the multiple face images using weighted fusion. Weighted fusion is to multiply the pixel value of each face image by its corresponding weight, and then add the weighted pixel values ​​to obtain a fused face image.

[0012] A method for improving passenger passage efficiency using the integrated boarding island device for improving passenger passage efficiency comprises the following steps: Step A: When the passenger approaches the integrated boarding island device, he / she enters the queuing area; When a passenger enters the capture area of ​​the capture column, the interactive screen of the capture column automatically lights up, displays a welcome message and guides the passenger to face the screen through voice, text or graphics; Step B: The capture column interacts with the passenger through the interactive screen and uses the built-in camera to identify the passenger; If a special passenger is identified, he / she will be guided to the manual verification channel; Step C: The gantry's multi-angle camera captures the team members who meet the requirements and obtains multiple facial images; Step D: All captured images are processed through built-in algorithms to screen out the face images of the best quality and send them to the comparison system for identity authentication; Step E: If the identity verification is successful, the turnstile and the flexible door wing remain open, and passengers can pass through smoothly; If the identity verification fails, the flexible door wing automatically closes, and the passenger is prompted through the interactive screen of the capture column to go to the manual channel for identity confirmation; In case of an abnormality, the flexible door wing automatically closes.

[0013] After adopting the above solution, the integrated boarding island device and method for improving the passenger passing efficiency of the present invention have the following advantages compared with the existing conventional turnstiles: Through multiple pre-shootings by the multi-angle cameras on the gantry of the present invention, passengers do not need to stay in front of the turnstile, significantly reducing the passing time and improving the passing efficiency.

[0014] The turnstile and the flexible door wing of the present invention remain open under normal conditions, providing a non-sensing passing experience for passengers, reducing the waiting time for passengers to open the turnstile, and improving the overall travel experience; while in case of an abnormality, the flexible door wing can automatically close to ensure the safety of passengers passing through and improve the safety of the turnstile.

[0015] Through the concealed camera and advanced algorithms of the present invention, passengers can be identified efficiently and accurately, reducing the misidentification rate.

[0016] The interactive screen of the capture column of the present invention can effectively attract the attention of passengers, improve the recognition efficiency, and the capture column can form a linear or L-shaped passing area according to the on-site area planning, improving the passing order, and thus improving the capture quality. The capture column can also identify special passengers in advance and guide special passengers to go through the manual channel through the interactive screen. Description of the Drawings

[0017] Figure 1 It is a schematic structural diagram of the device of the present invention.

[0018] Figure 2 It is a schematic diagram of another implementation state of the device of the present invention.

[0019] Figure 3 It is a schematic diagram of the gantry of the device of the present invention, and the cameras on it are actually concealed.

[0020] Figure 4 It is a schematic diagram of the rear door wing of the turnstile of the present invention being closed. Detailed Embodiments

[0021] In order to further explain the technical solution of the present invention, the present invention will be elaborated in detail below through specific embodiments.

[0022] AsFigures 1 to 4 As shown, the present invention discloses an integrated boarding island device for improving the passenger passing efficiency, which includes a turnstile 1, a flexible door wing 2, a gantry 3 and a capture column 4.

[0023] As a basic passing facility, the turnstile 1 remains open with the flexible door wing 2 under normal circumstances, providing a touchless passing experience.

[0024] The flexible door wing 2 is arranged at the tail of the turnstile 1 and can automatically close in case of abnormality to ensure safety.

[0025] The gantry 3 is installed on the turnstile 1, and multiple hidden cameras 3 at different angles are arranged on the gantry 3 (such as Figure 3 ), which capture the faces of passengers from different angles. The installation of the cameras minimizes the blind spots as much as possible to ensure that the face can be presented completely and clearly in the field of view of each camera. Multiple cameras can simultaneously capture the queue of people meeting the requirements, and use a multi-image fusion algorithm to intercept the face image with the best quality. The cameras on the gantry 3 are installed in a concealed manner to avoid causing pressure to the crowd.

[0026] The steps of intercepting the face image with the best quality by the multi-image fusion algorithm are as follows: Step S1: When a passenger appears within the shooting range of the camera, multiple cameras work simultaneously to capture the face images of the passenger from different angles, and obtain multiple groups of image data of different perspectives containing the same face. Step S2: Perform preprocessing on each captured face image, including operations such as grayscale conversion, noise reduction, and normalization. Grayscale conversion can convert a color image into a grayscale image, reducing the amount of data and facilitating subsequent processing; the noise reduction process uses a combination of mean filtering and median filtering methods to remove noise interference in the image and improve the image quality; normalization adjusts the pixel values of the image to a specific range for unified processing by the algorithm.

[0027] Step S3: Extract features from the preprocessed multiple face images respectively. Feature extraction uses a convolutional neural network (CNN) based on deep learning, and the extracted features can reflect the key information of the face, such as facial contours, positions of facial features, etc.

[0028] Step S4, assigning weights to each image according to the image quality assessment results. The image quality assessment is based on the image clarity combined with the facial posture. The image clarity is analyzed using any one of the algorithms in the gradient value method, the Laplace operator method or the spectrum analysis method. The gradient value method calculates the gradient amplitude of the image by finding the gradient of the image in the horizontal and vertical directions and then calculating its amplitude. The larger the gradient amplitude, the richer the edges and details of the image and the higher the clarity. The Laplace operator method uses the Laplace operator to perform a convolution operation on the image and calculate the second-order derivative of the image. The larger the absolute value of the result, the more drastic the changes in the details of the image and the higher the clarity. The spectrum analysis method performs a Fourier transform on the image and converts the image from the spatial domain to the frequency domain. The more high-frequency components there are, the richer the details and edge information of the image and the higher the clarity. The image clarity can be evaluated by analyzing the energy proportion of the high-frequency part in the frequency domain. Face posture assessment uses face key point detection method or posture estimation algorithm. Face key point detection method uses face key point detection algorithm to locate key feature points of face, such as eyes, nose, mouth, eyebrows, etc., and judge whether the face posture is normal according to the position and distribution of these key points. The posture estimation algorithm uses a model-based posture estimation algorithm. Based on the 3D face model, the detected facial features are matched with the model to calculate the rotation angle, pitch angle and yaw angle of the face. The image quality is relatively good when the posture parameters are within a certain range. If the range is exceeded, it may affect the face recognition effect and the quality is relatively low.

[0029] Step S5, image fusion is performed according to the assigned weights, and multiple images are fused using weighted fusion or other fusion methods. Weighted fusion is to multiply the pixel value of each image by its corresponding weight, and then add the weighted pixel values ​​to obtain a fused image. In addition to weighted fusion, pyramid fusion, wavelet fusion and other methods can also be used to fuse images according to different frequency components to better retain the detailed information of the image.

[0030] Step S6: Perform a quality assessment on the fused face image. The clarity assessment described in step S4 can be used, or evaluation indicators such as noise or color can be used to determine whether the fused image meets the best quality standard. If the quality of the fused image does not meet the requirements, the fusion process is repeated.

[0031] Step S7, based on the quality evaluation result of step S6, the best quality face image portion is captured from the fused image or the original multiple images, the position and size of the face in the image is determined according to the face detection algorithm, and then the face area is completely captured, and irrelevant background information is removed to obtain the best quality image containing only the face.

[0032] The capture column 4 is set in front of the turnstile 1. It can be extended into a straight or L-shaped passage area according to the on-site area planning, effectively regulating the passage order of passengers and improving the capture quality of multiple cameras on the subsequent gantry 3. A camera and an interactive screen are installed on the capture column 4. The capture column 4 can work independently, responsible for identifying special passengers and guiding special passengers to the manual verification channel through the interactive screen. Through the built-in image recognition algorithm and the special passenger database in the capture column 4, using the images captured by the camera, it can quickly and accurately identify special passengers. When special passengers such as passengers who need manual verification and passengers with limited mobility are identified, the capture column 4 immediately starts the guidance program and displays it on the interactive screen. Clear guidance information is displayed on the interactive screen, guiding special passengers to the manual verification channel in the way of arrow indication and text prompt. At the same time, the interactive screen also plays voice prompts to ensure that special passengers can accurately understand the guidance information.

[0033] The capture column 4 can be connected to the passenger identity verification system and the manual verification channel management system of the turnstile 1 for data interaction and collaborative work, and transmit the information of special passengers identified by the capture column 4 to the manual verification channel management system in time so that the staff can make reception preparations. At the same time, the passenger image data captured by the capture column 4 can be transmitted to the identity verification system and processed simultaneously with multiple images captured by the cameras on the gantry 3 to provide further support for identity verification.

[0034] In the present invention, a capture column 4 is provided at the front end of the turnstile 1, forming a guidance channel for passengers to queue or guide passengers into the turnstile 1 between the turnstile 1 and the capture column 4, making advance planning for the passage route of passengers, effectively alleviating traffic congestion, and improving the passage efficiency and experience of passengers; in addition, it can also create favorable conditions in advance for the capture of multiple cameras on the gantry 3, thereby improving the capture quality of the cameras on the gantry 3 and providing more reliable image support for passenger identity verification. The capture column 4 can also realize the intelligent guidance of special passengers through the interactive screen, greatly reducing manual intervention and improving the efficiency and accuracy of special passenger management.

[0035] The present invention also discloses a method for improving the passage efficiency of passengers using an integrated boarding island device, which includes the following steps: Step A: When a passenger approaches the integrated boarding island device, enter the queuing area; When the passenger enters the capture area of the capture column 4, the interactive screen of the capture column 4 automatically lights up, displays a welcome message and guides the passenger to face the screen by voice and text; Step B: The capture column 4 interacts with the passenger through the interactive screen and uses the built-in camera to identify the passenger's identity; If a special passenger (such as a passenger who needs manual verification) is identified, guide him to the manual verification channel; Step C: The multi-angle camera of the gantry 3 captures multiple face images of the eligible team members. Step D: All the captured images are processed by the built-in algorithm to select the face image with the best quality and send it to the comparison system for identity authentication. Step E: If the identity authentication is successful, the turnstile 1 and the flexible door wing 2 remain open, and the passengers can pass through smoothly. If the identity authentication fails, the flexible door wing 2 automatically closes, and the interactive screen of the capture column 4 prompts the passengers to go to the manual channel for identity confirmation. As Figure 4 shown, in case of abnormal situations (such as system failures, identity authentication failures, etc.), the flexible door wing 2 automatically closes to ensure passage safety.

[0036] The above embodiments and diagrams do not limit the product form and style of the present invention. Any appropriate changes or modifications made by those of ordinary skill in the relevant technical field shall be regarded as not departing from the patent scope of the present invention.

Claims

1. An integrated boarding island device for improving the passenger passage efficiency, characterized in that: It includes a turnstile, a flexible door wing, a gantry and a capture column; The flexible door wing is arranged at the tail of the turnstile, and the flexible door wing and the turnstile remain in an open state under normal circumstances; The gantry is installed on the turnstile. Multiple concealed cameras are arranged on the gantry to capture the faces of passengers from different angles. Multiple cameras capture at the same time, and the multi-image fusion algorithm is used to intercept the face image of the best quality for passenger identity recognition; The capture column is arranged in front of the turnstile, and extends with the turnstile to form a straight or L-shaped passage area. A camera and an interactive screen are installed on the capture column. The image recognition algorithm and the special passenger database built in the capture column are used to identify special passengers and guide special passengers to the manual verification channel through the interactive screen.

2. The integrated boarding island device for improving the passenger passage efficiency according to claim 1, wherein: The capture column is connected to the passenger identity verification system and the manual verification channel management system of the turnstile for data interaction and collaborative work. The passenger image data captured by the capture column is transmitted to the identity verification system for simultaneous processing with the multiple face images captured by the cameras on the gantry; when the special passenger information identified by the capture column is transmitted to the manual verification channel management system.

3. The integrated boarding island device for improving the passenger passage efficiency according to claim 1, characterized in that: The steps of the multi-image fusion algorithm to intercept the face image of the best quality are as follows: Step S1: When a passenger appears within the shooting range of the cameras on the gantry, multiple cameras work simultaneously to capture the face images of the passenger from different angles, and obtain multiple groups of image data of different perspectives containing the same face; Step S2: Preprocess each captured face image; Step S3: Extract features from each of the preprocessed multiple face images; Step S4: Evaluate the quality of each face image according to the extracted features, and assign weights to each face image according to the quality evaluation results of the face images; Step S5: Perform face image fusion according to the assigned weights; Step S6: Evaluate the quality of the fused face image, and judge whether the fused face image meets the best quality standard. If the quality of the fused face image does not meet the requirements, perform the fusion process again; Step S7: According to the quality evaluation results of Step S6, intercept the part of the face image with the best quality from the fused face image or the original multiple face images, determine the position and size of the face in the image according to the face detection algorithm, and then completely intercept the face area to remove irrelevant background information, and obtain the best quality image containing only the face.

4. The integrated boarding island device for improving the passenger passage efficiency according to claim 3, characterized in that: In Step S2, the image preprocessing includes grayscale conversion, noise reduction and normalization operations. The grayscale conversion converts the color face image into a grayscale image; the noise reduction process uses a combination of mean filtering and median filtering to remove the noise interference in the face image and improve the image quality; the normalization adjusts the pixel values of the face image to a specific range.

5. The integrated boarding island device for improving the passenger passage efficiency according to claim 3, characterized in that: In Step S3, the feature extraction uses a convolutional neural network based on deep learning, and the extracted features are used to reflect the key information of the face.

6. The integrated boarding island device for improving the passenger passage efficiency according to claim 3, wherein: In step S4, the image quality is evaluated based on the clarity of the face image combined with the face posture. The clarity of the face image is analyzed using any one of the algorithms of the gradient value method, the Laplace operator method or the spectrum analysis method. The gradient value method calculates the gradient amplitude of the face image by finding the gradient of the face image in the horizontal and vertical directions and then calculating its amplitude. The larger the gradient amplitude, the richer the edge and details of the face image and the higher the clarity. The Laplace operator method uses the Laplace operator to perform a convolution operation on the face image and calculate the second-order derivative of the face image. The larger the absolute value of the result, the more drastic the change in the details of the face image and the higher the clarity. The spectrum analysis method performs a Fourier transform on the face image and converts the face image from the spatial domain to the frequency domain. The more high-frequency components there are, the richer the details and edge information of the face image and the higher the clarity. The spectrum analysis method evaluates the image clarity by analyzing the energy proportion of the high-frequency part in the frequency domain. Face posture assessment uses face key point detection method or posture estimation algorithm. Face key point detection method uses face key point detection algorithm to locate the key feature points of the face, and judges whether the face posture is normal according to the position and distribution of these key points. The posture estimation algorithm adopts a model-based posture estimation algorithm. Based on the 3D face model, the detected facial features are matched with the 3D face model to calculate the posture parameters of the face's rotation angle, pitch angle and yaw angle. The face image quality is relatively good when the posture parameters are within a certain range. If they are out of the range, the face recognition effect is affected and the quality is relatively low.

7. The integrated boarding island device for improving the passenger passage efficiency according to claim 3, wherein: Step S5 uses weighted fusion to fuse multiple face images. Weighted fusion is to multiply the pixel value of each face image by its corresponding weight, and then add the weighted pixel values ​​to obtain a fused face image.

8. A method for improving the passenger passage efficiency by using the integrated boarding island device for improving the passenger passage efficiency according to any one of claims 1 to 7, characterized in that, The following steps are involved: Step A: When the passenger approaches the integrated boarding island device, he / she enters the queuing area; When a passenger enters the capture area of ​​the capture column, the interactive screen of the capture column automatically lights up, displays a welcome message and guides the passenger to face the screen through voice, text or graphics; Step B: The capture column interacts with the passenger through the interactive screen and uses the built-in camera to identify the passenger; If a special passenger is identified, he / she will be guided to the manual verification channel; Step C: The multi-angle camera of the gantry captures the crowd of people who meet the requirements and obtains multiple facial images; Step D: All captured images are processed by built-in algorithms to select the best quality face images and send them to the comparison system for identity authentication; Step E: If the identity verification is successful, the gate and the flexible door wing remain open and the passenger can pass smoothly; If identity verification fails, the flexible door wing will automatically close, and the passenger will be prompted to go to the manual channel for identity confirmation through the interactive screen on the capture column; In abnormal situations, the flexible door wings close automatically.

Images