Cross-border port non-inductive passing method based on front and rear cameras

Abstract

The invention discloses a cross-border port non-inductive passing method based on front and rear cameras, and particularly relates to the technical field of cross-border port security inspection.According to the system, the front camera is installed in a cross-border entrance area, the rear camera is arranged in an exit area, and non-inductive passing control is achieved; the front camera is responsible for capturing a passenger face image, and after image data is subjected to median filtering and denoising processing, facial features are extracted and sent to the central processing system; the center system quickly compares the features with identity information pre-recorded in a database through a face recognition algorithm so as to verify the identity of the passenger; the rear camera is responsible for acquiring images of luggage articles and identifying the types and the quantity of the luggage by using an image processing technology; the system compares the luggage identification result with the passenger declaration information to confirm the legality of the luggage; finally, the center system combines the identity of the passenger and a luggage verification result to realize rapid and seamless non-inductive customs clearance.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-border port security inspection, and more specifically, to a cross-border port contactless passage method based on front and rear cameras. Background Art

[0002] With the acceleration of globalization and the increasing frequency of cross-border trade and tourism, port traffic efficiency and service quality have become important indicators to measure the degree of openness and governance capabilities of a country or region.

[0003] With the rapid development of technologies such as artificial intelligence, big data, and the Internet of Things, the construction of smart ports has become a new trend in global port management. By integrating advanced technologies, port management departments can achieve efficient and accurate management of personnel and goods, thereby improving customs clearance efficiency, improving passenger experience, and reducing operating costs. Among them, the cross-border port non-sensing passage method based on front and rear cameras is an important innovative practice under this trend.

[0004] However, the traditional port access method still has some shortcomings in actual use. For example, the traditional port access method often relies on manual inspection, which is not only inefficient but also prone to errors, causing inconvenience to passengers and goods. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for seamless passage at a cross-border port based on front and rear cameras to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] Step A1: Install a front camera at the cross-border entry area and a rear camera at the cross-border exit area;

[0008] Step A2: The facial image information of the passenger at the entrance is collected by the front camera, and the real-time collected facial image data is transmitted to the central processing system through the network. The median filter denoising processing method is used to denoise the image, and the facial feature information is extracted from the denoised image;

[0009] Step A3: The central processing system quickly compares the extracted facial features with the pre-entered passenger identity information database and uses the face recognition algorithm to verify the identity;

[0010] Step A4: The rear camera collects images of the luggage items carried by the passenger and identifies the type and quantity characteristics of the luggage items through image processing technology;

[0011] Step A5: Compare the baggage identification results with the passenger's declared information to confirm the legitimacy of the baggage; and achieve seamless passage through passenger identity verification and baggage verification.

[0012] Preferably, in step A1, both the front camera and the rear camera use high-definition cameras with a resolution of 1920x1080, and have the conditions for night shooting, ensuring that the target is clearly photographed at night and under complex lighting conditions; support remote control of the camera's angle and focal length parameters, so as to flexibly adjust the monitoring range according to actual needs; at the same time, support viewing of real-time monitoring images through terminals such as computers or mobile phones, which is convenient for remote monitoring and command.

[0013] Preferably, in step A2, the method of performing median filtering to denoise the image is specifically as follows:

[0014] Step B1: Select a 5x5 filter window size and apply the median filter algorithm to each pixel in the image;

[0015] Step B2: For the pixel currently being processed, the grayscale values ​​of all pixels in the surrounding filtering window are collected; this includes the current pixel itself and the pixels at other positions in the window;

[0016] Step B3: Sort the collected pixel values; calculate the median according to the sorted pixel value sequence; if the number of pixels in the window is an odd number, the median is the pixel value in the middle position after sorting; if the number of pixels in the window is an even number, the median is the average of the two middle pixel values;

[0017] Step B4: Using the calculated median as the new grayscale value of the current pixel to replace the original pixel value, and completing the median filtering process on the current pixel;

[0018] Step B5: Repeat the above steps to perform median filtering on each pixel in the image until all pixels are processed to obtain an image denoised by median filtering.

[0019] Preferably, in step A3, the standardized value of the face-eye distance calculated in step A2 is matched with the passenger identity information database stored in the central processing system for the first face recognition. During the matching process, the system applies the similarity calculation technology in the face recognition algorithm for matching; when matching, the system sets a similarity threshold to determine whether the two belong to the same identity; if the similarity between the two exceeds the preset threshold, the system will determine that the identity authentication is passed; if the similarity between the two does not exceed the preset threshold, the first face recognition fails, and further processing is performed.

[0020] Preferably, in step A4, the collected baggage image is preprocessed, and the preprocessing includes contrast enhancement, brightness adjustment and color balance steps to improve image quality and make subsequent feature extraction more accurate; at the same time, image cropping or scaling may also be required to adapt to baggage images of different sizes and angles;

[0021] Using edge detection, texture analysis, and shape recognition image processing techniques, significant features of luggage items are extracted from pre-processed images; features include the outline, color distribution, texture pattern, and size ratio of the items; the extracted features are matched with a pre-trained classification model, which is trained based on a large amount of labeled data and can recognize a variety of common types of luggage items, such as electronic products, liquid containers, clothing, books, and food; the classification algorithm uses support vector machine technology to achieve high-precision recognition.

[0022] Preferably, in step A5, the system will compare the type and quantity of luggage items obtained by the image recognition technology with the list declared by the passenger one by one. In this process, it will not only check whether the basic attributes of the items are consistent, but also pay attention to whether the difference in quantity is within a reasonable range and whether there are any undeclared items;

[0023] After the comparison is completed, the system will evaluate the legality of the luggage items according to the preset rules and algorithms. These rules include the limited quantity of different types of items, the carrying regulations of special items, and the specific safety requirements of the country and region;

[0024] If the passenger's identity verification and baggage inspection are both passed, the passenger will be automatically released. For passengers who have passed all inspections, the system will realize contactless passage through linkage with other airport facilities (such as gates, electronic display screens, etc.). Passengers do not need to stop and wait for manual inspection, but only need to keep moving forward to quickly pass through the security inspection area; if any link is not passed, the system will immediately sound an alarm and notify the passenger and security personnel through voice broadcast; depending on the severity of the violation, the system may also trigger further inspection procedures, such as requiring passengers to open their bags for inspection, conduct stricter identity verification, or contact relevant departments for assistance; for passengers who fail the inspection, the system will record the processing process and results in detail, including specific information about the illegal items, the passenger's response, and the final solution.

[0025] Technical effects and advantages of the present invention:

[0026] The present invention can automatically capture and evaluate image quality, avoiding the need for passengers to stop and retake photos, reducing the need for passengers to cooperate with facial recognition, achieving a nearly seamless passage experience, improving efficiency, and reducing error rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments 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.

[0029] See also Figure 1 As shown, the present invention provides a method for seamless passage at a cross-border port based on front and rear cameras, comprising the following steps:

[0030] Step A1: Install a front camera at the cross-border entry area and a rear camera at the cross-border exit area;

[0031] In step A1, both the front camera and the rear camera are high-definition cameras with a resolution of 1920x1080, and have the conditions for night shooting, so as to ensure that the target can be clearly photographed at night and under complex light conditions; support remote control of the angle and focal length parameters of the camera, so as to flexibly adjust the monitoring range according to actual needs; at the same time, support viewing the real-time monitoring screen through terminals such as computers or mobile phones, which is convenient for remote monitoring and command;

[0032] Among them, high-definition camera equipment must be resistant to high temperatures, waterproof and dustproof, and be able to cope with the complex and changeable environmental conditions in border areas; at the same time, it is necessary to ensure the stability and security of data transmission to prevent information leakage or tampering.

[0033] The specific method of installing a high-definition camera is as follows:

[0034] Install the high-definition camera above the entrance area, ensure that the camera is flush with the equipment panel and the cable connection is stable and reliable; at the same time, pay attention to protecting the camera lens to avoid scratches or damage during the installation process, and ensure that the camera can fully cover the target area and is not easily blocked or damaged; after the installation is completed, debug and test the equipment; ensure that the camera can work normally, the image is clear, and the transmission is stable; at the same time, adjust the various parameters of the camera according to actual needs to obtain the best monitoring effect.

[0035] Step A2: The facial image information of the passenger at the entrance is collected by the front camera, and the real-time collected facial image data is transmitted to the central processing system through the network. The median filter denoising processing method is used to denoise the image, and the facial feature information is extracted from the denoised image;

[0036] In step A2, the method of performing median filtering to denoise the image is specifically as follows:

[0037] Step B1: Select a 5x5 filter window size and apply the median filter algorithm to each pixel in the image;

[0038] Step B2: For the pixel currently being processed, the grayscale values ​​of all pixels in the surrounding filtering window are collected; this includes the current pixel itself and the pixels at other positions in the window;

[0039] Step B3: Sort the collected pixel values; calculate the median according to the sorted pixel value sequence; if the number of pixels in the window is an odd number, the median is the pixel value in the middle position after sorting; if the number of pixels in the window is an even number, the median is the average of the two middle pixel values;

[0040] Step B4: Using the calculated median as the new grayscale value of the current pixel to replace the original pixel value, and completing the median filtering process on the current pixel;

[0041] Step B5: Repeat the above steps to perform median filtering on each pixel in the image until all pixels are processed to obtain an image denoised by median filtering.

[0042] The specific method for extracting image features based on the image denoised by median filtering is as follows:

[0043] The center of the face in the image is used as the origin to establish a coordinate system, and points (a1, b1) and (a2, b2) are randomly selected. The specific method for calculating the Euclidean distance is:

[0044] Among them, C represents the Euclidean distance, a1, a2, b1, b2 represent the coordinate values;

[0045] The specific calculation method of the normalized value of the face-eye distance in the image is:

[0046] Among them, D represents the standardized value of the face-eye distance in the image, C represents the Euclidean distance, E represents the key point prediction value, E0 represents the true value, and F represents the face-eye distance value in the image.

[0047] Step A3: The central processing system quickly compares the extracted facial features with the pre-entered passenger identity information database and uses the face recognition algorithm to verify the identity;

[0048] In step A3, the normalized value of the face-eye distance calculated in step A2 is matched with the passenger identity information database stored in the central processing system for first face recognition. During the matching process, the system applies the similarity calculation technology in the face recognition algorithm for matching. When matching, the system sets a similarity threshold to determine whether the two belong to the same identity. If the similarity between the two exceeds the preset threshold, the system will determine that the identity verification is passed. If the similarity between the two does not exceed the preset threshold, the first face recognition fails and further processing is performed.

[0049] The database contains pre-entered data on passengers’ facial features;

[0050] If the first face recognition fails, the further processing method is as follows:

[0051] The principal component analysis is used to extract the main features of the face image to form eigenfaces; these eigenfaces are the directions with the largest variance in the data. The calculation method of the second face eigenvalue is as follows:

[0052] Among them, H n Represented as the second face feature value, ε n Expressed as the corresponding eigenvalue, x i It is represented as the i-th sample in the data, ρ is represented as the average vector of all samples; arcmax is represented as finding the value of the parameter or independent variable that makes the function take the maximum value in a set of data;

[0053] The passengers who fail the first facial recognition are subjected to a second facial recognition, and the calculated second facial feature value is compared with the preset second facial feature threshold. If the second facial feature value is equal to the preset second facial feature threshold, a pass instruction is output; if the second facial feature value is not equal to the preset second facial feature threshold, a fail instruction is output, and the process returns to step A2 to re-collect the passenger's facial image.

[0054] Step A4: The rear camera collects images of the luggage items carried by the passenger and identifies the type and quantity characteristics of the luggage items through image processing technology;

[0055] In step A4, the collected baggage image is preprocessed, and the preprocessing includes contrast enhancement, brightness adjustment and color balance steps to improve image quality and make subsequent feature extraction more accurate; at the same time, image cropping or scaling may also be required to adapt to baggage images of different sizes and angles;

[0056] Using edge detection, texture analysis, and shape recognition image processing techniques, significant features of luggage items are extracted from pre-processed images; features include the outline, color distribution, texture pattern, and size ratio of the items; the extracted features are matched with a pre-trained classification model, which is trained based on a large amount of labeled data and can recognize a variety of common types of luggage items, such as electronic products, liquid containers, clothing, books, and food; the classification algorithm uses support vector machine technology to achieve high-precision recognition.

[0057] After identifying the type of items, the system further uses image processing technology to calculate the number of each type of items. The identified items are grouped, counted, and deduplicated to ensure the accuracy of the statistical results. For stacked or overlapping items, image segmentation and depth analysis technology are used to accurately separate and count them.

[0058] Step A5: Compare the baggage identification results with the passenger's declared information to confirm the legitimacy of the baggage; and achieve seamless passage through passenger identity verification and baggage verification.

[0059] In step A5, the system will compare the type and quantity of luggage items obtained by image recognition technology with the list declared by the passenger one by one. In this process, it will not only check whether the basic attributes of the items are consistent, but also pay attention to whether the difference in quantity is within a reasonable range and whether there are any undeclared items;

[0060] After the comparison is completed, the system will evaluate the legality of the luggage items according to the preset rules and algorithms. These rules include the limited quantity of different types of items, the carrying regulations of special items, and the specific safety requirements of the country and region;

[0061] If the passenger identity verification and baggage inspection are both passed, the passenger will be automatically released. For passengers who have passed all inspections, the system will realize touchless passage through linkage with other airport facilities (such as gates, electronic display screens, etc.). Passengers do not need to stop and wait for manual inspection. They can quickly pass through the security inspection area by just moving forward. If any link is not passed, the system will immediately sound an alarm and notify the passenger and security personnel through voice broadcast. Depending on the severity of the violation, the system may also trigger further inspection procedures, such as requiring passengers to open their bags for inspection, conduct stricter identity verification, or contact relevant departments for assistance. For passengers who fail the inspection, the system will record the processing process and results in detail, including the specific information of the illegal items, the passenger's response, and the final solution.

[0062] These records are of great significance for subsequent supervision, auditing and data analysis, and help airport management departments continuously optimize security inspection processes and improve safety and efficiency.

[0063] It should be noted that the preset values ​​in the present invention are determined according to specific circumstances, that is, this embodiment does not specifically limit the specific values.

[0064] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for seamless passage at a cross-border port based on front and rear cameras, characterized in that: include: Step A1: Install a front camera at the cross-border entry area and a rear camera at the cross-border exit area; Step A2: The facial image information of the passenger at the entrance is collected by the front camera, and the real-time collected facial image data is transmitted to the central processing system through the network. The median filter denoising processing method is used to denoise the image, and the facial feature information is extracted from the denoised image; Step A3: The central processing system quickly compares the extracted facial features with the pre-entered passenger identity information database and uses the face recognition algorithm to verify the identity; Step A4: The rear camera collects images of the luggage items carried by the passenger and identifies the type and quantity characteristics of the luggage items through image processing technology; Step A5: Compare the baggage identification results with the passenger's declared information to confirm the legitimacy of the baggage; and achieve seamless passage through passenger identity verification and baggage verification.

2. According to claim 1, a method for seamless passage through a cross-border port based on front and rear cameras is characterized in that: In step A1, both the front camera and the rear camera use high-definition cameras with a resolution of 1920x1080, and have the conditions for night shooting, ensuring that the target can be clearly photographed at night and under complex lighting conditions; support remote control of the camera's angle and focal length parameters, so as to flexibly adjust the monitoring range according to actual needs; at the same time, support viewing of real-time monitoring images through terminals such as computers or mobile phones, which is convenient for remote monitoring and command.

3. According to claim 1, a method for seamless passage through a cross-border port based on front and rear cameras is characterized in that: In step A2, the method of performing median filtering to denoise the image is specifically as follows: Step B1: Select a 5x5 filter window size and apply the median filter algorithm to each pixel in the image; Step B2: For the pixel currently being processed, the grayscale values ​​of all pixels in the surrounding filtering window are collected; this includes the current pixel itself and the pixels at other positions in the window; Step B3: sorting the collected pixel values; calculating the median value according to the sorted pixel value sequence; If the number of pixels in the window is an odd number, the median is the pixel value in the middle after sorting; If the number of pixels in the window is even, the median is the average of the two middle pixel values; Step B4: Using the calculated median as the new grayscale value of the current pixel to replace the original pixel value, and completing the median filtering process on the current pixel; Step B5: Repeat the above steps to perform median filtering on each pixel in the image until all pixels are processed to obtain an image denoised by median filtering.

4. According to claim 3, a method for seamless passage through a cross-border port based on front and rear cameras is characterized in that: The specific method for extracting image features based on the image denoised by median filtering is as follows: The center of the face in the image is used as the origin to establish a coordinate system, and points (a1, b1) and (a2, b2) are randomly selected. The specific method for calculating the Euclidean distance is: Among them, C represents the Euclidean distance, a1, a2, b1, b2 represent the coordinate values; The specific calculation method of the normalized value of the face-eye distance in the image is: Among them, D represents the standardized value of the face-eye distance in the image, C represents the Euclidean distance, E represents the key point prediction value, E0 represents the true value, and F represents the face-eye distance value in the image.

5. According to claim 1, a method for seamless passage through a cross-border port based on front and rear cameras is characterized in that: In step A3, the standardized value of the face-eye distance calculated in step A2 is matched with the passenger identity information database stored in the central processing system for the first face recognition. During the matching process, the system applies the similarity calculation technology in the face recognition algorithm for matching; when matching, the system sets a similarity threshold to determine whether the two belong to the same identity; if the similarity between the two exceeds the preset threshold, the system will determine that the identity authentication is passed; if the similarity between the two does not exceed the preset threshold, the first face recognition fails and further processing is performed.

6. The method for seamless passage through a cross-border port based on front and rear cameras according to claim 5 is characterized in that: If the first face recognition fails, the further processing method is as follows: The principal component analysis is used to extract the main features of the face image to form eigenfaces; these eigenfaces are the directions with the largest variance in the data. The calculation method of the second face eigenvalue is as follows: Among them, H n Represented as the second face feature value, ε n Expressed as the corresponding eigenvalue, x i It is represented as the i-th sample in the data, ρ is represented as the average vector of all samples; arcmax is represented as finding the value of the parameter or independent variable that makes the function take the maximum value in a set of data; The passengers who fail the first facial recognition are subjected to a second facial recognition, and the calculated second facial feature value is compared with the preset second facial feature threshold. If the second facial feature value is equal to the preset second facial feature threshold, a pass instruction is output; if the second facial feature value is not equal to the preset second facial feature threshold, a fail instruction is output, and the process returns to step A2 to re-collect the passenger's facial image.

7. According to claim 1, a method for seamless passage through a cross-border port based on front and rear cameras is characterized in that: In step A4, the collected baggage image is preprocessed, and the preprocessing includes contrast enhancement, brightness adjustment and color balance steps to improve image quality and make subsequent feature extraction more accurate; at the same time, image cropping or scaling may also be required to adapt to baggage images of different sizes and angles; Using edge detection, texture analysis, and shape recognition image processing techniques, significant features of luggage items are extracted from pre-processed images; features Including the outline, color distribution, texture pattern, and size ratio of the item; the extracted features are matched with a pre-trained classification model, which is trained based on a large amount of labeled data and can recognize a variety of common types of luggage items, such as electronic products, liquid containers, clothing, books, and food; the classification algorithm uses support vector machine technology to achieve high-precision recognition.

8. The method for seamless passage through a cross-border port based on front and rear cameras according to claim 1 is characterized in that: In step A5, the system will compare the type and quantity of luggage items obtained by image recognition technology with the list declared by the passenger one by one. In this process, it will not only check whether the basic attributes of the items are consistent, but also pay attention to whether the difference in quantity is within a reasonable range and whether there are any undeclared items; After the comparison is completed, the system will evaluate the legality of the luggage items based on preset rules and algorithms.

9. The method for seamless passage through a cross-border port based on front and rear cameras according to claim 8, characterized in that: If the passenger's identity verification and baggage inspection are both passed, the passenger will be automatically released. For passengers who have passed all inspections, the system will realize contactless passage through linkage with other airport facilities (such as gates, electronic display screens, etc.). Passengers do not need to stop and wait for manual inspection, but only need to keep moving forward to quickly pass through the security inspection area; if any link is not passed, the system will immediately sound an alarm and notify the passenger and security personnel through voice broadcast; depending on the severity of the violation, the system may also trigger further inspection procedures, such as requiring passengers to open their bags for inspection, conduct stricter identity verification, or contact relevant departments for assistance; for passengers who fail the inspection, the system will record the processing process and results in detail, including specific information about the illegal items, the passenger's response, and the final solution.

Images