Harbor district mobile security detection method based on image recognition, terminal and storage medium

Abstract

The invention relates to the technical field of image recognition, and particularly provides a harbor district mobile security detection method based on image recognition, a terminal and a storage medium, and the method comprises the steps: dividing a harbor district into a plurality of regions, and distributing a mobile patrol device for the plurality of regions; receiving inspection data uploaded by the mobile inspection device, and identifying a target from the inspection data by using a target identification algorithm; acquiring an authorized area of the target; and judging whether the authorized area is matched with the current area where the mobile patrol device is located or not, and if not, generating alarm prompt information. According to the invention, the security capability of the port is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of image recognition, and in particular relates to a port mobile security detection method, a terminal and a storage medium based on image recognition. Background Art

[0002] The port area usually constitutes a relatively closed working environment, which is characterized by a vast area and sparse personnel distribution. The main personnel are professionals who operate various types of working vehicles, and the working area is clearly divided. In the port area, most of the goods are directly piled on the ground, forming a dense storage scene.

[0003] In the port area's security protection strategy, the regional joint defense mechanism is considered an effective means of protection. However, given the vast geographical scope of the port area, if it relies on manual monitoring, it will not only require a large number of manpower for uninterrupted shift duty, resulting in high labor costs, but also appear to be unable to cope with the actual operation. On the other hand, if security cameras are widely installed to strengthen monitoring, there will be many challenges: the main operating vehicles in the port area are heavy engineering vehicles, excavators and loaders, etc. The special nature of these vehicles makes the planning of laying security lines extremely complicated; at the same time, the setting of poles may interfere with the normal stacking and operation process of goods, and even have the risk of causing safety accidents.

[0004] Therefore, there is an urgent need for an intelligent security system and device, which should have the ability to quickly identify and respond immediately when an alarm event occurs, so as to achieve efficient and accurate management of port security. Summary of the invention

[0005] In view of the above-mentioned deficiencies in the prior art, the present invention provides a port mobile security detection method, terminal and storage medium based on image recognition to solve the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides a port mobile security detection method based on image recognition, comprising: Divide the port area into multiple zones and assign mobile patrol units to multiple zones; Receiving inspection data uploaded by the mobile inspection device, and identifying the target from the inspection data using a target recognition algorithm; Obtaining the authorized area of ​​the target; It is determined whether the authorized area matches the current area where the mobile patrol device is located, and if the two do not match, an alarm prompt message is generated.

[0007] In an optional embodiment, the port area is divided into multiple areas, and mobile inspection devices are allocated to the multiple areas, including: Divide the port area into multiple areas and obtain the boundary coordinates of each area; Identify multiple mobile patrol devices, each of which includes a drive mechanism, a walking mechanism, a controller, a positioner, a communication module, a laser radar, a camera, a radio frequency identification device, and an audible and visual alarm device; An area is allocated to the mobile patrol device, and the boundary coordinates of the allocated area are sent to the corresponding mobile patrol device.

[0008] In an optional embodiment, the method further comprises: Multiple mobile inspection devices are connected to the network using a self-organizing network solution; designating a master station, a redundant master station, and a slave station among a plurality of said mobile inspection devices; Configure the data type and upload cycle uploaded from the slave to the master; Configure the data synchronization mode and in-place detection mode between the master station and the redundant master station; Configure the data type and upload cycle that the master station uploads to the upper terminal.

[0009] In an optional embodiment, a master station, a redundant master station and a slave station are designated in a plurality of the mobile inspection devices, including: Obtaining signal strength and hardware configuration information of a plurality of said mobile inspection devices; According to the pre-set quantization rules, the signal strength and hardware configuration information are converted into quantized values; sorting the plurality of mobile inspection devices from large to small according to the quantitative values; The mobile patrol units that are ranked the highest are designated as the master station and the redundant master station, and the remaining mobile patrol units are designated as slave stations.

[0010] In an optional embodiment, receiving inspection data uploaded by a mobile inspection device and identifying a target from the inspection data using a target recognition algorithm includes: Receiving inspection data uploaded by a mobile inspection device, wherein the inspection data includes laser point cloud data; Projecting the laser point cloud data onto a two-dimensional plane to obtain a two-dimensional feature map; Extracting contour features from the two-dimensional feature map to obtain target contour features; The target contour features are input into a pre-trained convolutional neural network to obtain target types, which include personnel and equipment.

[0011] In an optional implementation, obtaining the authorized area of ​​the target includes: Confirm that the target is identified and extract the target's location data; Controlling a camera to capture an image containing the target based on the position data; If the target is a person, a facial image is captured from the image, and the facial image is matched with a standard facial image in a permission database, and the authorization area bound to the matched standard facial image is output as the authorization area of ​​the target; If the target is a device, the device type is identified from the image, and a matching authorization area is retrieved from a permission database based on the device type.

[0012] In an optional implementation, obtaining the authorized area of ​​the target includes: Radio frequency technology is used to identify the target's permission tag, and the target's authorization area is obtained based on the correspondence between the permission tag and the authorization area.

[0013] In an optional embodiment, the method further comprises: Confirming the generation of alarm prompt information, extracting the location data of the target, wherein the location data includes the distance and direction from the mobile inspection device; generating walking control parameters for the mobile patrol device based on the position data of the target, so that the mobile patrol device tracks the target; The tracking path of the mobile patrol device is obtained. If the end point of the tracking path is the boundary of the authorized area of ​​the mobile patrol device, the coordinates of the end point are sent to the mobile patrol devices in the adjacent area to continuously track the target.

[0014] In a second aspect, a terminal is provided, including: processor, memory, wherein: The memory is used to store computer programs. The processor is used to call and run the computer program from the memory, so that the terminal executes the above-mentioned terminal method.

[0015] According to a third aspect, a computer storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the computer executes the methods described in the above aspects.

[0016] The beneficial effect of the present invention lies in that the mobile port security detection method, terminal and storage medium based on image recognition provided by the present invention collect laser point cloud images of multiple areas in parallel through mobile patrol devices, realize the monitoring of personnel or equipment, and through authorized area matching, further realize the supervision of illegal cross-regional operations, thereby improving the security capability of the port.

[0017] In addition, the invention has a reliable design principle, a simple structure and a very broad application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention.

[0020] Figure 2 A schematic diagram of the structure of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0023] The method for detecting mobile security in a port area based on image recognition provided in the embodiment of the present invention is executed by a computer device, and accordingly, the system for detecting mobile security in a port area based on image recognition runs in the computer device.

[0024] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention. Figure 1 The execution subject may be a mobile port security detection system based on image recognition. According to different requirements, the order of the steps in the flowchart may be changed, and some may be omitted.

[0025] like Figure 1 As shown, the method includes: Step 110, dividing the port area into a plurality of areas, and assigning mobile inspection devices to the plurality of areas; Step 120, receiving inspection data uploaded by the mobile inspection device, and identifying the target from the inspection data using a target recognition algorithm; Step 130, obtaining the authorized area of ​​the target; Step 140, determining whether the authorized area matches the current area where the mobile patrol device is located, and generating an alarm prompt message if the two do not match.

[0026] To facilitate the understanding of the present invention, the following is a further description of the mobile security detection method for port areas based on image recognition provided by the present invention, based on the principle of the mobile security detection method for port areas based on image recognition and the process of performing mobile security detection on the port area in the embodiment.

[0027] Specifically, the port area mobile security detection method based on image recognition includes: S1. Divide the port area into multiple areas and assign mobile patrol devices to multiple areas.

[0028] The port area is divided into multiple areas, and the boundary coordinates of each area are obtained; multiple mobile patrol devices are identified, and the mobile patrol devices include a driving mechanism, a walking mechanism, a controller, a locator, a communication module, a laser radar, a camera, a radio frequency identification device, and an audio-visual alarm device; areas are allocated to the mobile patrol devices, and the boundary coordinates of the allocated areas are sent to the corresponding mobile patrol devices.

[0029] In order to more effectively manage and monitor the vast port area, it is first necessary to reasonably divide it into multiple clear areas and obtain the boundary coordinates of each area through accurate surveying and mapping data. These coordinate information will provide a basic framework for subsequent security deployment, ensuring that monitoring coverage is complete and resource allocation is more reasonable.

[0030] In this embodiment, the mobile inspection device is an automatic inspection vehicle, and an electric lifting rod is provided on the upper part of the vehicle body, and a radar and a radar controller are fixed on the top of the electric lifting rod. The electric lifting rod includes a rod body connected up and down and an electric cylinder arranged in the rod body. The wireless communication device integrates 4G communication, WIFI communication, 5G communication, and radio station communication modules. A lithium battery pack is provided inside the vehicle body, and a driving wheel and a steering wheel are provided on the vehicle body. The driving wheel is connected to a reducer and a servo motor; the steering wheel is connected to a steering motor; and a charging port is provided at the rear of the vehicle body. An alarm and an infrared camera are also provided on the vehicle body. A PLC controller is used, and the PLC controller is connected to an industrial control touch screen arranged on the surface of the vehicle body; the PLC controller is also connected to the steering motor, servo motor, radar controller, electric cylinder, alarm, infrared camera, and wireless communication device respectively.

[0031] In other embodiments of the present invention, an automatic inspection robot or an automatic inspection robot dog may also be used.

[0032] Each mobile inspection unit is equipped with: Driving mechanism: responsible for providing forward power for the patrol device, ensuring that it can move freely within the port area and cover all areas that need to be monitored.

[0033] Walking mechanism: Designed with the ability to adapt to different terrains and environments, it can maintain stable and efficient movement regardless of whether it is a flat warehouse area or a complex loading and unloading area.

[0034] Controller: As the brain of the entire device, it is responsible for receiving instructions, processing data, coordinating the work of various parts, and ensuring the orderly execution of inspection tasks.

[0035] Locator: Combined with GPS or other high-precision positioning technologies, the location of the patrol device can be determined in real time, providing a basis for accurate monitoring and scheduling.

[0036] Communication module: ensures the transmission of inspection data between the inspection device and the central control center, including video streams, alarm information, etc., to achieve seamless information connection.

[0037] LiDAR: Used to build three-dimensional environment models, detect obstacles, avoid collisions, and assist in precise navigation.

[0038] Camera: Provides high-definition video surveillance to capture abnormal behaviors or events within the port area and provide visual evidence for security analysis.

[0039] Radio Frequency Identification Device: Used to identify specific vehicles, goods or personnel within the port area, and enhance the tracking and management of important assets.

[0040] Sound and light alarm device: When an abnormal situation is detected, it will be activated immediately and sound an alarm in the form of sound and light to remind on-site personnel to pay attention and take corresponding measures.

[0041] After clarifying the regional division and boundary coordinates of the port area, the corresponding monitoring areas are allocated to the mobile patrol devices according to the characteristics, risk level and patrol requirements of each area. This allocation process needs to comprehensively consider factors such as the coverage range, endurance, and response speed of the patrol device to ensure the optimal allocation of resources. The specific allocation principles are as follows: Conduct a detailed analysis of the geographical, climatic, transportation and other characteristics of each area in the port area to understand the natural environmental conditions and infrastructure conditions of each area.

[0042] Different types of monitoring areas are divided according to regional characteristics, such as high-risk areas (such as cargo loading and unloading areas, dangerous goods storage areas), medium-risk areas (such as warehouse areas, yard areas) and low-risk areas (such as office areas, living areas).

[0043] Combining historical data and expert evaluation, a risk level assessment is conducted on each monitored area to determine the security risk level of each area.

[0044] Based on the risk level, differentiated patrol strategies and monitoring requirements are formulated to ensure that high-risk areas receive more intensive patrols and monitoring.

[0045] Conduct performance evaluation on existing patrol devices, including key indicators such as coverage, battery life, response speed, and camera clarity.

[0046] According to the performance evaluation results, the patrol devices are divided into different levels to meet the needs of different monitoring areas.

[0047] According to the risk level and patrol requirements of each monitoring area, patrol devices with matching performance are allocated to the corresponding area.

[0048] Ensure that high-risk areas are equipped with patrol devices with superior performance and wide coverage, medium-risk areas are equipped with patrol devices with moderate performance and long endurance, and low-risk areas are equipped with patrol devices with basic performance.

[0049] After the allocation is completed, the boundary coordinate information of the allocated area is sent to the corresponding mobile patrol device through the communication module, so that it can accurately understand the scope of its responsibilities and start performing security tasks according to the preset patrol routes and strategies.

[0050] In one embodiment of the present invention, a plurality of mobile inspection devices form a self-organizing network with one device as a master station and the rest as slave stations. All slave station data are aggregated to the master station, and the master station exchanges information with the server.

[0051] Multiple mobile inspection devices are connected to the network using a self-organizing network solution; through wireless communication between the inspection devices, a stable network topology is automatically formed and maintained. This flexible network architecture can adapt to the complex and changing environment of the port area and ensure the reliability and real-time performance of data communication.

[0052] A master station, a redundant master station and a slave station are designated among the plurality of mobile patrol devices; when designating the master station and the redundant master station, the signal strength and hardware configuration information of the plurality of mobile patrol devices are obtained; according to a pre-set quantization rule, the signal strength and the hardware configuration information are converted into a quantized value; the plurality of mobile patrol devices are sorted from large to small according to the quantized value; the master station and the redundant master station are designated from the mobile patrol device with the highest sorting, and the remaining mobile patrol devices are designated as slave stations.

[0053] When building a self-organizing network solution for mobile inspection devices, it is crucial to reasonably specify the master station and redundant master station to ensure network stability and data communication efficiency. In order to scientifically and fairly select the master station and redundant master station, a quantitative evaluation method based on signal strength and hardware configuration information is designed.

[0054] Signal strength measurement: First, use professional signal measurement equipment to measure the signal strength of multiple mobile inspection devices in the assigned area. Signal strength is an important indicator to measure the communication capability of wireless devices. It reflects the signal quality of the device when sending and receiving data.

[0055] Hardware configuration information collection: Collect the hardware configuration information of each mobile inspection device. This includes the device's processor model, memory size, storage capacity, wireless communication module type, etc. This information is critical for evaluating the device's processing power and communication performance.

[0056] Quantification rule setting: In order to convert signal strength and hardware configuration information into comparable quantitative values, we set a set of quantitative rules. These rules are based on multiple dimensions such as the device's communication performance, processing power, and stability, and assign a weight and scoring standard to each indicator.

[0057] Quantitative value calculation: According to the quantitative rules, the signal strength and hardware configuration information of each mobile inspection device are quantitatively calculated to obtain the corresponding quantitative value. The higher the quantitative value, the stronger the communication performance and processing capacity of the device, and the more suitable it is to be used as a master station or redundant master station.

[0058] Sorting: Sort multiple mobile inspection devices by their quantitative values ​​from large to small. The sorting results reflect the priority and applicability of the devices in the self-organizing network solution.

[0059] Master station designation: From the top ranked mobile inspection devices, the device with the highest signal strength and the best hardware configuration is selected as the master station. The master station will be responsible for the management and maintenance of the network, as well as the aggregation, processing and analysis of data.

[0060] Redundant master station designation: Among the remaining devices, the device with better signal strength and hardware configuration is selected again as the redundant master station. The redundant master station will serve as a backup for the main station. When the main station fails or cannot work properly, it can quickly take over the work of the main station to ensure the stability of the network and the continuity of data.

[0061] Slave station designation: The remaining mobile inspection devices that have not been designated as master stations or redundant master stations will be designated as slave stations. The slave stations will be responsible for collecting laser point cloud data, video images, etc. in the port area, and upload the data to the master station according to the predetermined data type and upload cycle.

[0062] Configure the data type and upload cycle of the slave station to upload to the master station; according to the function and monitoring requirements of each slave station, configure the data type and upload cycle of each slave station to upload to the master station. The data type is specified as laser point cloud data and video data.

[0063] Configure the data synchronization mode and in-situ detection mode between the master station and the redundant master station; specifically, when the master station receives the data uploaded by the slave station, it will immediately synchronize the data to the redundant master station. After the data is updated, the redundant master station will feed back the MD5 value of the stored data to the master station. The master station calculates the MD5 value of its own stored data, and then verifies the data consistency by comparing the consistency of the MD5 value. If there is data inconsistency, it is necessary to compare the missing data and perform data synchronization. In addition, configure the in-situ detection mode to detect whether the master station and the redundant master station are both online and working normally. If the in-situ detection finds that the master station has a fault, the redundant master station will immediately take over the work of the master station to ensure the continuity of data communication. If the master station finds that the redundant master station has a fault, it generates an alarm message and sends a remote restart command to the redundant master station.

[0064] Configure the data type and upload cycle that the master station uploads to the superior terminal. The master station uploads the processed data to the superior terminal according to the predetermined data type and upload cycle. The superior terminal can be the port's management center, security monitoring center, etc. The uploaded data type can include but is not limited to laser point cloud data, video images, alarm information, etc. The upload cycle is determined according to the port's operational needs and the processing capacity of the superior terminal.

[0065] S2. Receive inspection data uploaded by the mobile inspection device, and use a target recognition algorithm to identify targets from the inspection data.

[0066] The laser point cloud data, video data, etc. collected by each mobile inspection device can be processed locally for target recognition and the processing results can be uploaded to the superior server, or the inspection data can be directly uploaded to the superior server through the main station for unified processing, depending on the computing power of the mobile inspection device.

[0067] Receive inspection data uploaded by a mobile inspection device, wherein the inspection data includes laser point cloud data; project the laser point cloud data onto a two-dimensional plane to obtain a two-dimensional feature map; extract contour features from the two-dimensional feature map to obtain target contour features; input the target contour features into a pre-trained convolutional neural network to obtain target types, wherein the target types include personnel and equipment.

[0068] Specifically, the laser point cloud data contains the three-dimensional coordinates (x, y, z) of each point in space.

[0069] Projection process: In order to project the 3D laser point cloud data onto a 2D plane, you need to select a projection plane. In this example, the horizontal plane (i.e. the plane with z=0) is selected as the projection plane.

[0070] For each three-dimensional point (x, y, z), set its z coordinate to 0 to obtain its projection point (x, y, 0) on the horizontal plane. That is, in a two-dimensional image representation, usually only the x and y coordinates are retained, so the projected point can be represented as (x, y).

[0071] In this way, all points in the 3D laser point cloud data can be projected onto a 2D plane to form a 2D feature map. This 2D feature map is actually a projection of a point cloud on a horizontal plane, which retains some of the spatial structure information of the original point cloud.

[0072] Image preprocessing: Before extracting contour features, it is usually necessary to preprocess the 2D feature map. This includes steps such as denoising and smoothing to reduce noise and interference in the image. Use relevant functions in image processing libraries (such as OpenCV) for preprocessing operations.

[0073] Edge detection: Edge detection is one of the key steps in extracting contour features. It aims to identify the edges of objects in the image and thus outline the contours of the objects. In this embodiment, classic algorithms such as the Canny edge detection algorithm are used for edge detection. These algorithms can identify areas with large gradient changes in the image and use them as edges.

[0074] Contour tracing: Based on edge detection, a contour tracing algorithm can be used to extract complete contour features. Contour tracing algorithms usually start from the edge points obtained by edge detection and gradually trace along the edge of the object until it returns to the starting point or encounters a breakpoint. Through contour tracing, a series of point sets representing the contour of the object can be obtained, and these point sets are the target contour features.

[0075] Post-processing: After extracting the target contour features, post-processing operations are performed to optimize the results. This includes steps such as removing redundant points and smoothing contours. Specifically, relevant functions in the image processing library are used for post-processing operations to obtain more accurate and clear contour features.

[0076] The following is a sample code for extracting contour features from a 2D feature map using Python and the OpenCV library: import cv2 import numpy as np # Assume that we already have a 2D feature map (use an example image instead here) image = cv2.imread('example_feature_map.png', cv2.IMREAD_GRAYSCALE) # Perform edge detection edges = cv2.Canny(image, 100, 200) # Find contours contours, hierarchy = cv2.findContours(edges, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) # Draw contours output_image = cv2.cvtColor(edges, cv2.COLOR_GRAY2BGR) cv2.drawContours(output_image, contours, -1, (0, 255, 0), 2) # Display the result cv2.imshow('Contours', output_image) cv2.waitKey(0) cv2.destroyAllWindows().

[0077] The optimized target contour features are input into a pre-trained convolutional neural network (CNN). CNN is a deep learning model that can automatically learn and extract high-dimensional features in images for target classification and recognition. CNN processes the input target contour features and outputs the target type. Target types include personnel and equipment, which are further refined into different types of equipment (such as engineering vehicles, transport vehicles, and other different equipment types).

[0078] Before implementing this solution, a large amount of training data containing target profile features and corresponding target types needs to be prepared. Use these training data to train the convolutional neural network so that it can learn the mapping relationship between target profile features and target types. During the training process, it is necessary to adjust the network's hyperparameters (such as learning rate, batch size, etc.) and use appropriate optimization algorithms (such as Adam, SGD, etc.) to optimize the network's performance.

[0079] S3. Obtain the authorized area of ​​the target.

[0080] In one embodiment, image recognition technology is used to obtain the authorized area of ​​the target, confirm that the target is identified, and extract the target's location data based on the target's laser point cloud data; based on the location data, the camera is controlled to capture an image containing the target; if the target is a person, a facial image is captured from the image, and the facial image is matched with a standard facial image in a permission database, and the authorized area bound to the matched standard facial image is output as the authorized area of ​​the target; if the target is a device, the device type is identified from the image, and the matching authorized area is retrieved from the permission database based on the device type.

[0081] Specifically, the center point of the target is obtained from the laser point cloud data of the target, and the relative position of the target and the mobile inspection device is obtained based on the coordinates (x, y, z) of the center point to obtain the position data of the target.

[0082] The pre-processed laser point cloud data is processed using an advanced center point detection algorithm (such as the CenterPoint algorithm). The algorithm can accurately identify the center point of the target and output its three-dimensional coordinates (x, y, z). The extracted center point is verified to ensure its accuracy and stability. This can be achieved by comparing with the center point of a known target, or by scanning multiple times and calculating the average value of the center point. Use GPS, inertial navigation systems or other positioning technologies to determine the position of the mobile inspection device in the current environment. This usually includes the three-dimensional coordinates of the device (x_device, y_device, z_device) and possible posture information. Based on the center point coordinates of the target (x, y, z) and the position of the mobile inspection device (x_device, y_device, z_device), calculate the relative position between the target and the device. This can be achieved by calculating the Euclidean distance or vector difference between the two. The calculated relative position data is output for subsequent decision-making and control. These data may include information such as the straight-line distance, direction angle, etc. between the target and the device.

[0083] According to the extracted position data, the camera's angle, focal length, exposure and other parameters are automatically adjusted to ensure that the target can be captured clearly. The camera is triggered to collect images to ensure that the image contains complete target information. The collected images are pre-processed, such as denoising and contrast enhancement, to improve the accuracy of subsequent processing.

[0084] Face image capture and matching (for personnel): Accurately detect and capture the face area in the image. Preprocess the captured face image, such as grayscale, histogram equalization, etc., to improve the matching accuracy. Compare the preprocessed face image with the standard face image in the permission database, and use advanced matching algorithms (such as deep learning feature matching) to ensure the accuracy of the match. If the match is successful, extract the authorization area information bound to the standard face image as the target's authorization area. At the same time, record the matching results and related information for subsequent audits.

[0085] Device type identification and permission retrieval (for devices): Use image recognition algorithms to identify the type of device (such as transport vehicles, engineering vehicles, etc.). According to the device type, retrieve the authorized area information that matches the device from the permission database. The authorized area information includes specific areas that the device is allowed to access, time restrictions, etc.

[0086] In another embodiment of the present invention, the permission tag of the target may be identified by using radio frequency technology, and the authorization area of ​​the target may be obtained based on the corresponding relationship between the permission tag and the authorization area.

[0087] Specifically, system configuration: First, ensure that the radio frequency identification system (RFID system) has been correctly installed and configured. This includes the initialization of the RFID reader, antenna, and permission tag.

[0088] Database synchronization: Import the corresponding relationship data between permission tags and authorized areas into the system database. This usually includes the unique identifier of the tag (such as UID) and the corresponding authorized area information (such as area name, scope, access rights, etc.).

[0089] RF signal transmission: The RF reader transmits RF signals through the antenna to activate the authorization tag on the target. Tag response: When the target enters the coverage area of ​​the RF signal, the authorization tag on it will be activated and send a response signal containing its unique identifier. Signal reception and processing: The RF reader receives the response signal from the tag and extracts the unique identifier of the tag through decoding processing.

[0090] S4. Determine whether the authorized area matches the current area where the mobile patrol device is located, and generate an alarm prompt message if the two do not match.

[0091] The current area where the mobile patrol device is located is obtained, and it is determined whether the current area belongs to the authorized area of ​​the target. If it does not belong to the authorized area, an alarm prompt message is generated; if it belongs to the authorized area, no alarm prompt message is generated.

[0092] S5. Behavior recognition.

[0093] Scan the location of the wall and trigger the wall-climbing system alarm for the positioning tag holder or illegal person who suddenly appears or disappears in the wall area.

[0094] Specifically, the coordinates of the wall monitoring area are stored in advance. If the actual coordinates of the mobile patrol device enter the wall monitoring area, when the laser point cloud data recognizes a human target, it triggers the acquisition of the target's video data, and identifies the wall-climbing behavior from the video data.

[0095] Convert surveillance videos to formats suitable for video analysis, such as MP4 or AVI. Adjust the frame rate of the video according to the quality and requirements of the surveillance video to ensure the accuracy and real-time nature of the video analysis. De-noise the video to improve the image quality; at the same time, enhance the image to highlight the target features.

[0096] Use deep learning algorithms (such as YOLO, SSD, etc.) to detect human bodies in each frame of the video and extract the position and size information of human targets. Use target tracking algorithms (such as SORT, DeepSORT, etc.) to continuously track human targets to obtain their motion trajectories. Use posture estimation algorithms (such as OpenPose, HRNet, etc.) to estimate the posture of human targets and extract the position information of joint points. According to the motion trajectory and posture information of human targets, calculate relevant behavioral characteristics, such as speed, acceleration, posture change, etc.

[0097] Using a large amount of video data of circumvention behavior, a deep learning model (such as a convolutional neural network (CNN) or a recurrent neural network (RNN)) is trained to identify circumvention behavior. The extracted behavior features are input into the trained deep learning model to classify the behavior. If the model determines that it is circumvention behavior, the recognition result is output.

[0098] S6. Illegal target tracking.

[0099] After the alarm is triggered, the regional security device moves to the target point. The server pre-mobilizes the nearby regional security devices to move in the target direction according to its moving direction. After the regional security device moves to the boundary of its own moving area, it stops and returns to its initial position. Its moving path is uploaded to the server. The server summarizes the moving paths of the relevant regional security devices to form a target moving path.

[0100] Confirm the generation of alarm prompt information, extract the target's location data, the location data including the distance and direction to the mobile patrol device; generate walking control parameters for the mobile patrol device based on the target's location data, so that the mobile patrol device tracks the target; obtain the tracking path of the mobile patrol device, and if the end point of the tracking path is the boundary of the authorized area of ​​the mobile patrol device, send the end point coordinates to the mobile patrol device in the adjacent area to continuously track the target.

[0101] Specifically, based on the extracted target location data, the intelligent monitoring system will generate a set of walking control parameters for the mobile patrol device closest to the target. These parameters include walking speed, steering angle, driving path, etc., to ensure that the patrol device can reach the target location at the fastest speed and shortest path. After receiving the walking control parameters, the mobile patrol device will immediately start and move to the target location along the specified path. During the movement, the patrol device will continuously use its own laser radar and camera to perceive the surrounding environment and update its own location information in real time. The movement state of the patrol device is continuously tracked, and the walking control parameters are adjusted as needed to ensure that the patrol device can accurately reach the target location. When the mobile patrol device successfully reaches the target location and starts tracking the target, the intelligent monitoring system will record the tracking path of the patrol device. If the end point of the tracking path happens to be the boundary of the authorized area of ​​the patrol device, the coordinate information of the end point will be automatically sent to the mobile patrol device in the adjacent area. After receiving the coordinate information, the patrol device in the adjacent area will immediately start and continue to track the target along the specified path to ensure that the target can be continuously tracked and monitored throughout the port area. The present invention takes each area security device as the center point, and forms regional protection of the entire port area by overlapping and covering each small area. At the same time, through the preset judgment logic, timely warning is given for wall-climbing actions. Whether it is illegal personnel or illegal operations by legal personnel, timely alarm processing can be achieved. Figure 2 A schematic diagram of the structure of a terminal 200 provided in an embodiment of the present invention. The terminal 200 can be used to execute the port mobile security detection method based on image recognition provided in an embodiment of the present invention.

[0102] The terminal 200 may include: a processor 210, a memory 220 and a communication unit 230. These components communicate via one or more buses. Those skilled in the art will appreciate that the server structure shown in the figure does not limit the present invention, and it may be a bus structure or a star structure, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0103] The memory 220 can be used to store the execution instructions of the processor 210, and the memory 220 can be implemented by any type of volatile or non-volatile storage terminal or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. When the execution instructions in the memory 220 are executed by the processor 210, the terminal 200 can perform some or all of the steps in the following method embodiments.

[0104] The processor 210 is the control center of the storage terminal, and uses various interfaces and lines to connect various parts of the entire electronic terminal. It runs or executes software programs and / or modules stored in the memory 220, and calls data stored in the memory to perform various functions of the electronic terminal and / or process data. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of multiple packaged ICs with the same or different functions. For example, the processor 210 can only include a central processing unit (CPU). In the embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.

[0105] The communication unit 230 is used to establish a communication channel so that the storage terminal can communicate with other terminals, receive user data sent by other terminals or send user data to other terminals.

[0106] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, the program may include some or all of the steps in each embodiment provided by the present invention. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0107] Therefore, the present invention collects laser point cloud images of multiple areas in parallel through mobile patrol devices to achieve monitoring of personnel or equipment, and further realizes supervision of illegal cross-regional operations through authorized area matching, thereby improving the security capabilities of the port. The technical effects that can be achieved by this embodiment can be found in the description above and will not be repeated here.

[0108] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention, in essence or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes, including several instructions for enabling a computer terminal (which can be a personal computer, a server, or a second terminal, a network terminal, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention.

[0109] In this specification, the same or similar parts between the various embodiments can be referred to each other. In particular, for the terminal embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiment.

[0110] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or modules, which can be electrical, mechanical or other forms.

[0111] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0113] Although the present invention has been described in detail with reference to the accompanying drawings and in combination with the preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, a person of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and these modifications or substitutions shall be within the scope of the present invention. Any person of ordinary skill in the art may easily think of changes or substitutions within the technical scope disclosed by the present invention, and these shall be within the scope of protection of the present invention.

Claims

1. A method for detecting mobile security in port areas based on image recognition, characterized in that: include: Divide the port area into multiple zones and assign mobile patrol units to multiple zones; Receiving inspection data uploaded by the mobile inspection device, and identifying the target from the inspection data using a target recognition algorithm; Obtaining the authorized area of ​​the target; It is determined whether the authorized area matches the current area where the mobile patrol device is located, and if the two do not match, an alarm prompt message is generated.

2. The method according to claim 1, characterized in that The port area is divided into multiple areas and mobile patrol units are assigned to multiple areas, including: Divide the port area into multiple areas and obtain the boundary coordinates of each area; Identify multiple mobile patrol devices, each of which includes a drive mechanism, a walking mechanism, a controller, a positioner, a communication module, a laser radar, a camera, a radio frequency identification device, and an audible and visual alarm device; An area is allocated to the mobile patrol device, and the boundary coordinates of the allocated area are sent to the corresponding mobile patrol device.

3. The method according to claim 1, characterized in that The method further comprises: Multiple mobile inspection devices are connected to the network using a self-organizing network solution; designating a master station, a redundant master station, and a slave station among a plurality of said mobile inspection devices; Configure the data type and upload cycle uploaded from the slave to the master; Configure the data synchronization mode and in-place detection mode between the master station and the redundant master station; Configure the data type and upload cycle that the master station uploads to the upper terminal.

4. The method according to claim 3, characterized in that Designating a master station, a redundant master station and a slave station in a plurality of the mobile inspection devices includes: Obtaining signal strength and hardware configuration information of a plurality of said mobile inspection devices; According to the pre-set quantization rules, the signal strength and hardware configuration information are converted into quantized values; sorting the plurality of mobile inspection devices from large to small according to the quantitative values; The mobile patrol units that are ranked the highest are designated as the master station and the redundant master station, and the remaining mobile patrol units are designated as slave stations.

5. The method according to claim 1, characterized in that Receiving inspection data uploaded by the mobile inspection device and identifying the target from the inspection data using a target recognition algorithm, including: Receiving inspection data uploaded by a mobile inspection device, wherein the inspection data includes laser point cloud data; Projecting the laser point cloud data onto a two-dimensional plane to obtain a two-dimensional feature map; Extracting contour features from the two-dimensional feature map to obtain target contour features; The target contour features are input into a pre-trained convolutional neural network to obtain target types, which include personnel and equipment.

6. The method according to claim 5, characterized in that Get the authorized area of ​​the target, including: Confirm that the target is identified and extract the target's location data; Controlling a camera to capture an image containing the target based on the position data; If the target is a person, a facial image is captured from the image, and the facial image is matched with a standard facial image in a permission database, and the authorization area bound to the matched standard facial image is output as the authorization area of ​​the target; If the target is a device, the device type is identified from the image, and a matching authorization area is retrieved from a permission database based on the device type.

7. The method according to claim 1, characterized in that Get the authorized area of ​​the target, including: Radio frequency technology is used to identify the target's permission tag, and the target's authorization area is obtained based on the correspondence between the permission tag and the authorization area.

8. The method according to claim 1, characterized in that The method further comprises: Confirming the generation of alarm prompt information, extracting the location data of the target, wherein the location data includes the distance and direction from the mobile inspection device; generating walking control parameters for the mobile patrol device based on the position data of the target, so that the mobile patrol device tracks the target; The tracking path of the mobile patrol device is obtained. If the end point of the tracking path is the boundary of the authorized area of ​​the mobile patrol device, the coordinates of the end point are sent to the mobile patrol devices in the adjacent area to continuously track the target.

9. A terminal, characterized in that: include: A memory, used for storing a port area mobile security detection program based on image recognition; A processor is used to implement the steps of the port area mobile security detection method based on image recognition as described in any one of claims 1 to 8 when executing the port area mobile security detection program based on image recognition.

10. A computer-readable storage medium storing a computer program, characterized in that: The readable storage medium stores a port mobile security detection program based on image recognition, and when the port mobile security detection program based on image recognition is executed by the processor, the steps of the port mobile security detection method based on image recognition as described in any one of claims 1-8 are implemented.

Images