A differentiated security inspection method and security inspection system for urban rail transit

Abstract

The present invention discloses a differentiated security inspection method and system for urban rail transit, and belongs to the field of security inspection technology. The method comprises the following steps: S10, collecting real-name authentication information of passengers and storing it in an identification database; S20, diverting passengers who need security inspection according to real-name authentication passengers and non-real-name authentication passengers according to whether the passenger information is in the identification database; S30, performing differentiated security inspection on diverted passengers, wherein rapid security inspection is performed on real-name authentication passengers, and conventional security inspection is performed on non-real-name authentication passengers; S40, transmitting the security inspection information of real-name authenticated passengers back to the identification database, and removing passenger information with dangerous behaviors in the security inspection information from the identification database. The present invention can effectively reduce the number of stranded passengers at the station, improve the passenger travel efficiency, shorten the waiting time of passengers in queues, and improve the service level by classifying and diverting passengers according to whether real-name authentication is performed and taking differentiated security inspection methods.

Description

Technical Field

[0001] The present invention belongs to the technical field of security inspection, and more specifically, relates to a differentiated security inspection method and a security inspection system for urban rail transit. Background Art

[0002] With the continuous acceleration of urbanization and the continuous expansion of the scale of rail transit network, urban rail transit, as an important part of the urban public transportation system, has become an important means to alleviate traffic congestion. The main function of rail transit is to provide travel services for passengers. Therefore, passengers are the primary consideration in rail transit design.

[0003] Passengers entering the station are required to purchase tickets, go through security checks, queue up, and check tickets. The level of security check services will directly affect the attractiveness of rail transit to passengers. The existing security check method requires that each passenger must pass through security checks in sequence, and full security checks will affect the passenger's travel efficiency and increase the time it takes for passengers to enter the station.

[0004] After searching, the Chinese patent publication number is CN 207867532 U; the publication date is September 14, 2018; a system and a classification security checkpoint for completing subway classification security inspection through real-name registration are disclosed, and the system includes: a real-name authentication device for assisting in the completion of real-name authentication; a Ministry of Public Security device for sending identity information and blacklist information to the security inspection operation management device, and receiving the target information sent by the security inspection operation management device in real time; a third-party device for sending third-party data to the security inspection operation management device; the security inspection operation management device is used to classify users according to the blacklist information and third-party data, and obtain classification results, so that users can enter the target security inspection channel for security inspection according to the classification results. This application case uses a system for classification security inspection to improve the efficiency of passenger passage, but the system implementation process of this application case is complicated, and the improvement of passage efficiency is limited. There is still a large burden for implementation and application under peak passenger flow. Summary of the invention

[0005] In order to solve at least one of the above technical problems, according to one aspect of the present invention, a differentiated security inspection method for urban rail transit is provided, comprising the following steps:

[0006] S10, collecting passenger real-name authentication information and storing it in an identification database;

[0007] S20. Divide passengers who need security check into real-name authenticated passengers and non-real-name authenticated passengers according to whether the passenger information is in the identification database;

[0008] S30. Carry out differentiated security checks on the diverted passengers, wherein rapid security checks are performed on real-name authenticated passengers and regular security checks are performed on non-real-name authenticated passengers.

[0009] The urban rail transit differentiated security inspection method according to an embodiment of the present invention may optionally further include:

[0010] S40: The security check information of the real-name authenticated passengers is transmitted back to the identification database, and the passenger information with dangerous behaviors in the security check information is removed from the identification database.

[0011] According to the differentiated security inspection method for urban rail transit according to an embodiment of the present invention, optionally, the rapid security inspection in step S3 includes a mode of performing random inspection on real-name authenticated passengers, not inspecting small bags, and inspecting large bags.

[0012] According to the differentiated security inspection method for urban rail transit according to an embodiment of the present invention, optionally, in step S2, facial information comparison is used to identify passenger information.

[0013] According to the differentiated security inspection method for urban rail transit according to an embodiment of the present invention, optionally, in step S3, both the rapid security inspection and the conventional security inspection use the YOLO model method to detect prohibited items.

[0014] According to the urban rail transit differentiated security inspection method according to an embodiment of the present invention, optionally, the YOLO model method includes:

[0015] S31, preprocessing, set the image containing the passenger's belongings to 448×448 size, use the CSPDarknet53 backbone feature extraction network to extract the deep features of the carried items, send the feature vector to SPP and PANet for predicting output features, and predict and output features at three scales;

[0016] S32, detection, the preprocessed image is divided into S×S grids, the grid carrying the center coordinates of the object in the image is responsible for detection, the output dimension of the detection network is S×S×[B×(5+C)], where B is the number of three predicted bounding boxes, 5 is the position parameter and confidence of each bounding box, C is the category probability, and the predicted category confidence Conf(object) is obtained.

[0017]

[0018] Among them, Pr(object) indicates whether there are contraband items in the predicted border. If there are, Pr(object) = 1, if not, Pr(object) = 0. Characterize the accuracy of the bounding box;

[0019] S33, prediction result processing, using the NMS algorithm to process the prediction results of each scale, find the best target bounding box, eliminate redundant bounding boxes, and obtain the location information of prohibited items in the image.

[0020] According to the differentiated security inspection method for urban rail transit according to an embodiment of the present invention, optionally, in step S3, when the time is in peak hours, a quick security inspection is performed on the real-name authenticated passengers, and when the time is in off-peak hours, a regular security inspection is performed on the real-name authenticated passengers.

[0021] According to another aspect of the present invention, there is provided a differentiated security inspection system for urban rail transit, comprising:

[0022] Real-name authentication channel, ordinary channel, X-ray security equipment and detection door;

[0023] The real-name authentication channel and the common channel are arranged parallel to each other, the real-name authentication channel is used to pass the real-name authentication passengers, and the common channel is used to pass the non-real-name authentication passengers;

[0024] The entrance of the real-name authentication channel is equipped with an identification section and face recognition system to identify whether the passenger is in the identification database;

[0025] The X-ray security inspection equipment is used to detect whether there are prohibited items in the belongings of passengers in the real-name authentication channel and the ordinary channel;

[0026] There are two detection gates, which are respectively arranged in the real-name authentication channel and the ordinary channel, and are used to detect whether the passengers entering the channel carry metal objects and count the number of people passing through.

[0027] According to the differentiated security inspection system for urban rail transit according to an embodiment of the present invention, optionally, the real-name authentication channel and the ordinary channel share a set of X-ray security inspection equipment.

[0028] According to the differentiated security inspection system for urban rail transit according to an embodiment of the present invention, optionally, when the time is in off-peak hours, the real-name authentication channel can allow non-real-name authenticated passengers to enter and perform routine security inspections, and the face recognition system continuously collects face information of passengers in the channel for improving the recognition database.

[0029] Beneficial Effects

[0030] Compared with the prior art, the present invention has at least the following beneficial effects:

[0031] (1) Compared with the existing single passenger security inspection channel, the present invention adds a real-name authentication fast security inspection channel to divert passengers, which can significantly improve the security inspection speed and achieve a safe passing rate of more than 1,800 people per hour during peak hours at the station;

[0032] (2) The present invention implements the fast security inspection standard for real-name authenticated passengers, which not only facilitates the fast passage of real-name authenticated passengers, but also helps with passenger information management, and effectively alleviates the "capacity-based" contradiction caused by insufficient security personnel and equipment capacity during large-scale passenger flow;

[0033] (3) The present invention adopts the optimized YOLO model for intelligent detection of prohibited items, which can solve the problems of insufficient security personnel, overwork and the resulting false detection and missed detection, improve the recognition rate of prohibited items, and enhance the working efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly introduced below. Obviously, the drawings in the following description only relate to some embodiments of the present invention, but are not intended to limit the present invention.

[0035] Figure 1 The flowchart of the differentiated security inspection method for urban rail transit of the present invention is shown;

[0036] Figure 2 A schematic diagram of the differentiated security inspection system for urban rail transit of the present invention is shown. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] Unless otherwise defined, technical or scientific terms used herein shall have the same meaning as commonly understood by one of ordinary skill in the art to which the invention belongs.

[0039] The existing security inspection method requires that each passenger must pass the security inspection in order, and all the passengers' large and small bags must be checked by X-ray security inspection equipment. During peak hours with large passenger flow, this full-staff security inspection method will affect the passenger's travel efficiency, increase the time for passengers to enter the station, and cause crowd congestion. Based on this, the present invention designs a differentiated security inspection method and security inspection system for urban rail transit.

[0040] Example 1

[0041] The differentiated security inspection method for urban rail transit of the present invention is as follows: Figure 1 As shown, the following steps are included:

[0042] S10, collecting passenger real-name authentication information and storing it in an identification database;

[0043] Passengers are evaluated and authenticated in the database before entering the station, and their identity information and corresponding facial information are entered and stored in the recognition database.

[0044] S20. Divide passengers who need security check into real-name authenticated passengers and non-real-name authenticated passengers according to whether the passenger information is in the identification database;

[0045] When passengers go through security check and enter the station, they are diverted according to whether they have undergone real-name authentication. Facial information comparison is used to determine whether the passenger is in the identification database. The passenger facial information features collected in real time are compared with the information in the identification database. The comparison database uses 1:N 10,000 face recognition, and the recognition response time is ≤0.5s. Passengers who are in the identification database after identification enter the real-name authentication passenger flow. Passengers who are not in the identification database after identification can be guided by staff into the non-real-name authentication passenger flow.

[0046] S30, performing differentiated security checks on the diverted passengers, wherein fast security checks are performed on real-name authenticated passengers, and regular security checks are performed on non-real-name authenticated passengers;

[0047] The fast security check for real-name authenticated passengers means that passengers in the real-name authenticated passenger flow will be randomly checked, and small bags will not be checked, but large bags will be checked;

[0048] Carrying out routine security checks on non-real-name authenticated passengers means conducting security checks on all passengers in the non-real-name authenticated passenger flow, and checking both small and large bags;

[0049] The above inspection refers to checking whether the items carried by passengers contain prohibited items. Specifically, the YOLO model method is used to detect prohibited items. The YOLO model method includes:

[0050] S31, preprocessing, set the image containing the passenger's belongings to 448×448 size, use the CSPDarknet53 backbone feature extraction network to extract the deep features of the carried items, send the feature vector to SPP and PANet for predicting output features, and predict and output features at three scales;

[0051] The features of the three-scale prediction output can detect object images of different sizes in complex scenes, thereby improving the accuracy of the model;

[0052] S32, detection, the preprocessed image is divided into S×S grids, the grid carrying the center coordinates of the object in the image is responsible for detection, the output dimension of the detection network is S×S×[B×(5+C)], where B is the number of three predicted bounding boxes, 5 is the position parameter and confidence of each bounding box, C is the category probability, and the predicted category confidence Conf(object) is obtained.

[0053]

[0054] Among them, Pr(object) indicates whether there are contraband items in the predicted border. If there are, Pr(object) = 1, if not, Pr(object) = 0. The accuracy of the bounding box is an important indicator for measuring prediction accuracy;

[0055] S33, prediction result processing, using the NMS algorithm to process the prediction results of each scale, find the best target bounding box, eliminate redundant bounding boxes, and obtain the location information of prohibited items in the image.

[0056] S40, transmitting the security check information of the real-name authenticated passengers back to the identification database, and removing the passenger information with dangerous behaviors in the security check information from the identification database;

[0057] When random checks are conducted on passengers who have undergone real-name authentication, if the passenger is found to be carrying prohibited items, a secondary security check will be required, and the passenger information will be marked and deleted from the identification database. Subsequently, the passenger will need to be re-evaluated and re-authenticated to restore his or her real-name authentication qualifications.

[0058] The differentiated security inspection method for urban rail transit of this embodiment classifies and diverts incoming passengers according to whether they have real-name authentication, and performs differentiated security inspections on different diversions, which can greatly improve the efficiency of security inspection entry, effectively alleviate the "capacity" contradiction caused by insufficient security inspection personnel and equipment capabilities during large-scale passenger flows, and significantly improve the security inspection speed. Furthermore, the optimized YOLO model method of this embodiment is used to detect contraband, which has a higher recognition rate for contraband, reduces the false detection and missed detection rates, and improves the safety of incoming passengers.

[0059] Example 2

[0060] The urban rail transit differentiated security inspection system of this embodiment is as follows: Figure 2 As shown, including:

[0061] Real-name authentication channel, ordinary channel, X-ray security equipment and detection door.

[0062] The real-name authentication channel and the ordinary channel are arranged parallel to each other. The real-name authentication channel is used to pass real-name authenticated passengers and implement the standards of random inspection of personnel, no inspection of small bags and inspection of large bags. The ordinary channel is used to pass non-real-name authenticated passengers and implement the standards of full inspection of personnel and inspection of both large and small bags.

[0063] The entrance of the real-name authentication channel is equipped with an identification section and face recognition system to identify whether the passenger is in the identification database;

[0064] Passengers enter the identification section of the real-name authentication channel entrance through the guide sign. The face recognition system performs face detection, face recognition, and face tracking functions. The camera captures the face of the passenger currently entering the channel, and generates a face pattern code from the current face image file. The current face pattern code is compared with the identification database inventory. The comparison database adopts 1:N 10,000 face recognition. The system recognition response time is ≤0.5s. A special channel door is set at the end of the identification section. When the recognition meets the implementation standards, the special channel door opens and prompts the passenger to pass through in the form of voice or light. When the recognition fails, the special channel door closes, and the security staff guides the passenger to leave the identification section and enter the station through the ordinary channel security check;

[0065] Furthermore, the length of the recognition section must provide sufficient recognition time for the face recognition system.

[0066] The X-ray security inspection equipment is used to detect whether there are prohibited items in the belongings of passengers in the real-name authentication channel and the ordinary channel;

[0067] In this embodiment, the ordinary channel and the real-name authentication channel are set in parallel at the entrance. The two channels can be separated by half-height railings, or order can be maintained by using warning tapes, etc., and they share a set of X-ray security inspection equipment. The security inspection system of the X-ray security inspection equipment adopts the YOLO model method to detect prohibited items, and has the function of intelligent image recognition and judgment. It automatically collects images of items in the X-ray security inspection equipment, completes image cleaning, filtering, labeling and other processing, and detects and alarms prohibited items in the image at the same time, and marks the location and type of prohibited items in the image. The security inspection intelligent system conducts deep learning on the collected data to realize image acquisition, image preprocessing, target detection, prohibited item warning and other functions for detecting prohibited items in passengers' belongings, thereby reducing space occupancy and improving inspection efficiency.

[0068] There are two detection gates, which are respectively arranged at the entrance of the real-name authentication channel and the ordinary channel, and are used to detect whether the passengers entering the channel carry metal objects and count the number of people passing through.

[0069] Furthermore, the real-name authentication channel is mainly used during peak hours in the morning and evening to alleviate the pressure of entering the station. During off-peak periods, the door of the dedicated channel is opened and can be used as a channel for ordinary passengers. At the same time, the facial recognition system continues to work to collect information of passengers entering the station.

[0070] The urban rail transit differentiated security inspection system of this embodiment is specifically implemented as follows:

[0071] When passengers arrive at the concourse level, they follow the guidance signs and choose the corresponding channel for security check according to whether they have undergone real-name authentication. Passengers who choose the real-name authentication channel must undergo verification by the identification section and face recognition system when entering. After passing the verification, the dedicated channel door opens, and the system prompts passengers to enter the real-name authentication channel to perform fast security check standards in the form of voice or light. If the verification fails, the staff will guide the passengers to enter the ordinary channel;

[0072] Passengers in the general channel will be subject to regular security inspection standards. During the security inspection, passengers enter the channel in order and place their luggage on the conveyor belt of the X-ray security inspection equipment in accordance with the requirements. The check-in staff will identify the shape and type of items in the image of the checked luggage on the monitor of the security inspection equipment to check whether the passengers are carrying prohibited items. Then, the detection gate will be used to detect whether the passengers are carrying metal objects and count the number of people passing through and the number of people who have been alarmed. If the detection gate alarms, the manual inspector will conduct a second inspection on the passengers who have passed the inspection.

[0073] Passengers in the real-name authentication channel will be subject to fast security inspection standards. For real-name authenticated passengers who pass through the dedicated channel door, random inspection will be carried out, and small bags will not be inspected, but large bags will be inspected. Passengers who are not selected or are not carrying luggage or small bags can continue to pass through and swipe their cards to enter the station after passing through the detection gate. For passengers who are selected, they will put their luggage on the conveyor belt of the X-ray security inspection equipment as required. The check-in staff will identify the shape and type of items in the image of the inspected luggage on the monitor of the security inspection equipment to check whether they are carrying prohibited items. Then, the detection gate will be used to detect whether the passengers are carrying metal objects, and the number of people passing through and the number of people who are alarmed will be counted. If the detection gate The alarm is triggered, and the manual inspector then conducts a second check on the passengers who have passed the check. Passengers found carrying prohibited items during security checks in the real-name authentication channel are required to undergo a second security check and be handled by security personnel. The passenger's information will be marked and deleted from the identification database, and re-evaluation and re-authentication are required to restore the real-name authentication qualification. The passenger will not be able to enter the station through the real-name authentication channel before re-authentication. When the time is in the off-peak period, the real-name authentication channel also implements regular security inspection standards and can allow passengers who have not been authenticated to enter. At this time, the face recognition system remains on to collect facial information of passengers in the channel for the purpose of improving the identification database.

[0074] The examples described in the present invention are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the concept and scope of the present invention. Without departing from the design concept of the present invention, various modifications and improvements made to the technical solutions of the present invention by engineers and technicians in this field should all fall within the protection scope of the present invention.

Claims

1. A differentiated security inspection method for urban rail transit, characterized in that: The steps include: S10, collecting passenger real-name authentication information and storing it in an identification database; S20. Based on the result of determining whether the real-name authentication information of the passenger is in the identification database, passengers who need security check are divided into real-name authenticated passengers and non-real-name authenticated passengers; S30, performing differentiated security checks on the diverted passengers, wherein rapid security checks are performed on real-name authenticated passengers, and regular security checks are performed on non-real-name authenticated passengers; both rapid security checks and regular security checks use the YOLO model method to detect prohibited items, and the YOLO model method includes: S31, preprocessing, set the image containing the passenger's belongings to 448×448 size, use the CSPDarknet53 backbone feature extraction network to extract the deep features of the carried items, send the feature vector to SPP and PANet for predicting output features and outputting features at three scales; S32, detection, the preprocessed image is divided into S×S grids, the grid carrying the center coordinates of the object in the image is responsible for detection, the output dimension of the detection network is S×S×[B×(5+C)], where B is the number of three predicted bounding boxes, 5 is the position parameter and confidence of each bounding box, C is the category probability, and the predicted category confidence Conf(object) is obtained. Among them, Pr(object) indicates whether there are contraband items in the predicted border. If there are, Pr(object) = 1, if not, Pr(object) = 0. Characterize the accuracy of the bounding box; S33, prediction result processing, using the NMS algorithm to process the prediction results of each scale, find the best target bounding box, eliminate redundant bounding boxes, and obtain the location information of prohibited items in the image.

2. The differentiated security inspection method for urban rail transit according to claim 1 is characterized in that: After step S30, the method further includes: S40: The security check information of the real-name authenticated passengers is transmitted back to the identification database, and the passenger information with dangerous behaviors in the security check information is removed from the identification database.

3. A differentiated security inspection method for urban rail transit according to claim 1 or 2, characterized in that: In step S30, the rapid security check includes a mode of performing random inspection on passengers with real-name authentication and inspecting large bags while omitting small bags.

4. A differentiated security inspection method for urban rail transit according to claim 1 or 2, characterized in that: In step S20, the passenger information is identified by comparing the facial information.

5. The differentiated security inspection method for urban rail transit according to claim 2 is characterized by: In step S30, when the time is in the peak period, the real-name authenticated passengers are subject to a quick security check, and when the time is in the off-peak period, the real-name authenticated passengers are subject to a regular security check.

6. A differentiated security inspection system for urban rail transit, characterized in that: The differentiated security inspection method for urban rail transit according to any one of claims 1 to 5 comprises: Real-name authentication channel, ordinary channel, X-ray security equipment and detection door; The real-name authentication channel and the common channel are arranged parallel to each other, the real-name authentication channel is used to pass the real-name authentication passengers, and the common channel is used to pass the non-real-name authentication passengers; The entrance of the real-name authentication channel is equipped with an identification section and face recognition system to identify whether the passenger is in the identification database; The X-ray security inspection equipment is used to detect whether there are prohibited items in the belongings of passengers in the real-name authentication channel and the ordinary channel; There are two detection gates, which are respectively arranged in the real-name authentication channel and the ordinary channel, and are used to detect whether the passengers entering the channel carry metal objects and count the number of people passing through.

7. The differentiated security inspection system for urban rail transit according to claim 6 is characterized by: The real-name authentication channel and the ordinary channel share a set of X-ray security inspection equipment.

8. The differentiated security inspection system for urban rail transit according to claim 6 is characterized by: During off-peak hours, the real-name authentication channel can allow non-real-name authenticated passengers to enter and undergo routine security checks, and the facial recognition system continuously collects facial information of passengers in the channel to improve the recognition database.

Images