Mobile target data analysis method, system and storage medium based on field environment

Abstract

The application relates to the technical field of intelligent transportation, and discloses a mobile target data analysis method and system based on a field environment and a storage medium. The method comprises the following steps: setting a safety code corresponding to a plurality of color feature representations for a target feature; acquiring a site to which the target feature belongs, and calculating first distribution data of the safety code; calculating a dynamic level of the site according to the first distribution data; acquiring a running platform of the site, and calculating second distribution data of the dynamic level; calculating a risk level of the running platform according to the second distribution data; acquiring label information in real time, reducing the safety level if the content of the label information is illegal behavior, updating the safety code according to the safety level, and generating an execution log; updating the dynamic level and the risk level by a management background; triggering a site early warning if the dynamic level is lower than a dynamic reference level; triggering a platform early warning if the risk level is lower than a risk reference level; and the main body of the non-motor vehicle can be systematically managed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent transportation, in particular to a mobile target data analysis method and system based on a field environment and a storage medium. BACKGROUND

[0002] The rapid development of express delivery and take-out industries has made non-motor vehicles the main means of transportation for practitioners, but related illegal behaviors are common. Red light running, reverse driving, and occupying motor vehicle lanes are frequent, not only disrupting the order of urban road traffic, but also easily causing traffic accidents and threatening the safety of pedestrians and other vehicles, becoming a difficult problem for urban traffic management.

[0003] To strengthen the safety management of non-motor vehicles, various measures have been taken: traffic control departments increase road patrol frequency and intensify the handling of illegal behaviors; electronic monitoring and snapshot devices are installed on key road sections to record and punish illegal behaviors; traffic safety propaganda and education is carried out to improve the law-abiding consciousness of practitioners; and industry self-discipline mechanisms are implemented in some areas, requiring enterprises to strengthen the management of employees. These measures have to some extent curbed illegal activities, but there are still problems such as incomplete management coverage and lack of real-time performance.

[0004] Existing management measures mainly rely on on-site discovery and on-site punishment, and lack systematic management intervention means for the relevant subjects of non-motor vehicles. When non-motor vehicles commit illegal acts, practitioners cannot be reminded in time to correct them, resulting in the continuous occurrence of illegal acts and the difficulty of eliminating safety hazards. SUMMARY

[0005] In order to systematically manage and intervene the relevant subjects of non-motor vehicles, the present application provides a mobile target data analysis method and system based on a field environment and a storage medium.

[0006] In a first aspect, the present application provides a mobile target data analysis method based on a field environment, which adopts the following technical solution:

[0007] A mobile target data analysis method based on a field environment, comprising the following steps:

[0008] A safety code corresponding to a plurality of color feature representations is set for a target feature, and each safety code corresponds to a different safety level;

[0009] The site to which the target feature belongs is obtained, and the first distribution data of the safety code corresponding to the target feature owned by the site is calculated;

[0010] The dynamic level corresponding to the site is calculated according to the first distribution data, and the dynamic level has multiple levels;

[0011] acquire a running platform corresponding to the affiliated station, calculate second distribution data of the dynamic level corresponding to the affiliated station owned by the running platform;

[0012] calculate a risk level corresponding to the running platform according to the second distribution data, the risk level having multiple levels;

[0013] acquire a target label of the target feature in real time through label reading auxiliary camera video algorithm recognition, if the content of the label information of the target label corresponds to illegal behavior, reduce the security level of the target feature, update the security code of the target feature according to the security level, and generate an execution log;

[0014] upload the label information and the execution log to a management background;

[0015] the management background updates the dynamic level and the risk level according to the label information and the execution log;

[0016] if the dynamic level is lower than a preset dynamic reference level, trigger a station early warning;

[0017] if the risk level is lower than a preset risk reference level, trigger a platform early warning.

[0018] By adopting the above technical solution, the target feature, the affiliated station and the running platform are automatically rated according to the data indicators of the target feature. Among them, the target feature is marked with a security level by a security code in multiple colors, and the target feature can be queried in the management background, realizing one-to-one code and dynamic traceability. The affiliated station is rated by a multi-level dynamic level. The running platform is rated by multiple risk levels. The related subjects of the non-motor vehicle are systematically managed and intervened.

[0019] Optionally, the step of acquiring the target label of the target feature in real time further includes the following sub-steps:

[0020] acquire a live image through a camera video algorithm, identify a target area from the live image, and identify the target feature based on the target area;

[0021] if the target feature is identified, calculate a target position corresponding to the target feature based on the live image, and perform target early warning;

[0022] calculate a target route and a target direction according to multiple target positions;

[0023] acquire a camera or a label reader located on the target route and in front of the target direction;

[0024] select the camera or the label reader closest to the target position;

[0025] reading a target label on the target feature through a camera or a label reader;

[0026] if the target label is read, obtaining target data corresponding to the target label and uploading the target data;

[0027] if the target label is not read, prompting that the target feature does not have the target data, and uploading the on-site image and the target feature.

[0028] By adopting the technical solution, the illegal behavior of the target feature (non-motor vehicle) can be identified in real time and a warning can be given on site (motor vehicle lane on the road) to stop the illegal behavior in time, reduce safety hazards, and significantly improve the timeliness of management. Through hierarchical image recognition, position calculation, and route direction analysis, combined with double verification of the label reader (such as an RFID label), the illegal subject and behavior can be accurately locked, the identification accuracy can be greatly improved, and the misjudgment rate can be reduced. Through the method, illegal behavior can be intervened in real time, and the label data can be used for post-factum accountability. The label-free target can also be recorded, solving the problem of incomplete coverage of traditional management. Moreover, relying on existing monitoring equipment and label readers, the automatic process reduces hardware investment and labor costs, and improves management efficiency.

[0029] Optionally, the step of identifying a target feature based on the target area further includes the following sub-steps:

[0030] extracting a license plate feature information from a moving target in the target area;

[0031] analyzing the license plate feature information, and distinguishing the moving target as an electric bicycle or a motorcycle according to the coding rule or the identification feature of the license plate.

[0032] By adopting the technical solution, the electric bicycle or the motorcycle can be distinguished by the license plate, which can improve the degree of target identification refinement and enhance the accuracy of law enforcement management.

[0033] Optionally, the method further includes the following steps:

[0034] if the moving target is an electric bicycle or a motorcycle, obtaining the license plate feature information of the electric bicycle or the motorcycle;

[0035] obtaining corresponding ownership information according to the license plate feature information;

[0036] querying a warning record amount based on the ownership information;

[0037] if the warning record amount of the ownership information is greater than a preset warning reference amount, sending a warning information according to the ownership information.

[0038] By adopting the technical scheme, the enterprise with illegal problems highlighted in the express delivery and take-out industry can be quickly found by associating the early warning record with the information (such as the information of the operating company of the non-motor vehicle operator in the express delivery and take-out industry) of the operator, so that the management measures of the traffic control department are more targeted; and the sending of the warning information can be beneficial to the internal management of the enterprise and the education and constraint of illegal behaviors.

[0039] Optionally, the method further comprises the following sub-steps:

[0040] Triggering a camera or a tag reader in the next order located on the target route and in front of the target direction to read the target tag on the target feature.

[0041] By adopting the technical scheme, the target tag reading success rate can be improved, the information loss caused by single reading failure can be reduced, and the effective capture and recording of non-motor vehicle illegal information can be ensured.

[0042] Optionally, the method further comprises the following steps:

[0043] According to the target position and the recording time corresponding to the target position, a target speed is calculated;

[0044] Based on the target speed, the number of cameras or tag readers to be started is determined, and the number and the target speed are in a positive correlation; the greater the number, the greater the target speed; the smaller the number, the smaller the target speed;

[0045] Simultaneously start a corresponding number of tag readers located on the target route and in front of the target direction to read the target tag on the target feature.

[0046] By adopting the technical scheme, the number of readers started according to the speed is increased, the effective reading range and time window can be expanded through the relay or simultaneous capture of multiple devices, the risk of information missing reading caused by high speed can be greatly reduced, and the label information of the high-speed moving target can be timely and accurately acquired.

[0047] Optionally, the method further comprises the following steps after the step of sending the warning information according to the information:

[0048] Calculate the early warning ratio of the early warning record quantity and the early warning reference quantity;

[0049] Determine the number of cameras or tag readers to be started based on the early warning ratio, the number is positively correlated with the early warning ratio; the larger the number, the larger the early warning ratio; the smaller the number, the smaller the early warning ratio;

[0050] Start the corresponding number of cameras or tag readers located on the target route and in front of the target direction at the same time, and read the target tag on the target feature.

[0051] By adopting the above technical scheme, for the vehicle of the enterprise with high illegal rate, the number of tag readers started at the same time is increased, the reading failure probability caused by single device signal interference, angle deviation and other factors can be reduced through multi-device collaborative reading, and information loss is minimized. The reading success rate and information integrity of the target tag are greatly improved.

[0052] Optionally, the step of reading the target tag further includes the following sub-steps:

[0053] The started tag sensors all generate reading data, and the content of the reading data includes complete reading, partial reading and non-reading;

[0054] Calculate the reading completeness = the number of tag sensors with complete reading / (the number of tag sensors with partial reading + the number of tag sensors with non-reading);

[0055] If the reading completeness > the preset reference completeness, the target tag is read.

[0056] By adopting the above technical scheme, multi-device redundant verification avoids single sensor misjudgment and reduces the missed detection rate; the reading quality is quantitatively evaluated by replacing the binary judgment with the completeness index, and the system fault tolerance is improved; the resource allocation is dynamically adjusted in combination with the early warning ratio or the target speed to balance the supervision efficiency and cost.

[0057] In a second aspect, the application provides a mobile target data analysis system based on a field environment, which adopts the following technical scheme:

[0058] A mobile target data analysis system based on a field environment, comprising a processor, wherein the processor executes the steps of the mobile target data analysis method based on a field environment according to any one of the above.

[0059] In a third aspect, the application provides a storage medium, which adopts the following technical scheme:

[0060] A storage medium, wherein the storage medium stores a program, and the program is executed by a processor to implement the steps of the mobile target data analysis method based on a field environment according to any one of the above.

[0061] In summary, the present application includes at least one of the following beneficial technical effects: through on-site image recognition and target position calculation, real-time early warning and suppression of non-motor vehicle illegal behavior is realized, the management timeliness is significantly improved, and the security risks are reduced. Combined with number plate feature analysis and RFID tag double verification, both electric bicycles or motorcycles can be distinguished, and the illegal subject can be accurately locked, greatly reducing the misjudgment rate and enhancing the precision of law enforcement. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 is a step diagram of a mobile target data analysis method based on an on-site environment.

[0063] Figure 2 is a step diagram of a mobile target being an electric bicycle or a motorcycle.

[0064] Figure 3 is a step diagram of reading the target tag. DETAILED DESCRIPTION

[0065] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0066] In the description of the present specification, the description of the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the described embodiments or examples are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0067] The present application embodiment discloses a mobile target data analysis method based on an on-site environment, referring to Figure 1 , comprising the following steps:

[0068] setting a security code corresponding to a plurality of color feature representations for the target feature, each security code corresponding to a different security level;

[0069] acquiring the site to which the target feature belongs, and calculating first distribution data of the security code corresponding to the target feature owned by the site;

[0070] calculating the dynamic level corresponding to the site according to the first distribution data;

[0071] acquiring the running platform corresponding to the site, and calculating second distribution data of the dynamic level corresponding to the site owned by the running platform;

[0072] According to the second distribution data, the risk level corresponding to the operation platform is calculated;

[0073] Real-time acquisition of the target label of the target feature, if the content of the label information of the target label corresponds to illegal behavior, the security level of the target feature is reduced, the security code of the target feature is updated according to the security level, and an execution log is generated;

[0074] The label information and the execution log are uploaded to the management background;

[0075] The management background updates the dynamic level and the risk level according to the label information and the execution log;

[0076] If the dynamic level is lower than the preset dynamic reference level, a site early warning is triggered;

[0077] If the risk level is lower than the preset risk reference level, a platform early warning is triggered.

[0078] The target feature is a working rider, according to the data indexes such as the use situation of the special license plate of the working rider, the traffic violation and accident handling situation, and the implementation situation of traffic safety training and education, the implementation loopholes of the main responsibility of the enterprise and the site can be automatically identified, and the safety level of the rider, the site and the enterprise is automatically evaluated. The site belongs to the distribution site of the working rider, the operation platform belongs to the platform enterprise of the distribution site, and the management background belongs to the administrative platform, such as "Shanghai Application APP". Among them, the safety level of the working rider is marked with "green, yellow and red" safety code, and the working rider can also query in "Shanghai Application APP", realizing "one code for one person, dynamic traceability"; the distribution site is implemented with dynamic star rating from 1 to 5 stars, and 5 stars are the best; the platform enterprise is implemented with "high risk" and "low risk" two risk level evaluation.

[0079] Through the deployment of high-definition monitoring cameras along the road (such as fixed monitoring on both sides of the motor lane, pole type mobile snapshot equipment, etc.), real-time collection of on-site environment images is realized. The collection frequency can be dynamically adjusted according to the road traffic flow, set to 25 frames per second during peak period and 15 frames per second during flat peak period, to ensure that the dynamic characteristics of moving targets can be clearly captured. For example, at the crossroads of urban trunk roads, the camera can cover the motor lane, non-motor lane and pedestrian area within a radius of 50 meters, providing complete image data source for subsequent target identification.

[0080] An image segmentation algorithm based on deep learning (such as the U-Net model) is used to divide the on-site image into regions, extracting the pre-set target area from the complex background. The target area is the range of motor vehicle lane that needs to be supervised. Specifically, through the iterative optimization of the model by training samples (containing motor vehicle lane images under different light, weather, and time periods), the model can accurately identify lane lines, barriers, and other features, thereby defining the pixel range of the motor vehicle lane. For example, in the image, the motor vehicle lane area is marked as a red rectangular frame, and the non-motor vehicle lane and pedestrian lane are marked as other colors, achieving clear distinction between the target area and the non-target area.

[0081] Within the determined motor vehicle lane area, a target detection algorithm (such as YOLOv5) is used to scan the image and identify moving targets such as vehicles, pedestrians, etc., and further extract target features. For non-motor vehicles, the focus is on identifying their shape features, riding status (such as whether they carry passengers or not, and whether they are reversing), and attached signs (such as number plates, vehicle color, etc.). For example, when an electric bicycle or motorcycle is detected entering the motor vehicle lane, the system will automatically extract the feature of "no engine exhaust pipe" and preliminarily determine it as a non-motor vehicle, marking it as a target feature to be verified. Motorcycles include electric motorcycles and gasoline motorcycles.

[0082] If a non-motor vehicle is identified entering the motor vehicle lane, i.e., the target feature exists, the actual geographic location of the target feature is calculated through an image coordinate conversion algorithm. Specifically, based on the installation parameters of the monitoring camera (such as altitude, depression angle, focal length) and the pre-set road coordinate system (with the intersection center point as the origin, establishing an x-y plane coordinate), the pixel coordinates of the target feature in the image are converted into actual latitude and longitude coordinates.

[0083] At the same time, the system triggers a real-time warning mechanism: through the deployment of sound and light alarms at the intersection, such as LED warning screens on the edge of the motor vehicle lane and high-pitched loudspeakers; for example, the screen displays "Non-motor vehicles are not allowed to occupy the motor vehicle lane", and the loudspeaker simultaneously broadcasts a voice warning; if the target feature is an electric bicycle or electric motorcycle, it can also send a text warning through its vehicle terminal, such as "You have entered the motor vehicle lane, please be careful and leave immediately", achieving immediate intervention on illegal behavior.

[0084] By continuously collecting multiple location coordinates of the target feature, such as recording the latitude and longitude every 1 second, a trajectory fitting algorithm such as the least squares method is used to generate the target route. For example, the location coordinates of a non-motor vehicle in 3 seconds are (x1, y1), (x2, y2), and (x3, y3), and the system determines its driving trajectory as east to west along the motor vehicle lane by fitting the straight line formed by these three points.

[0085] At the same time, according to the time sequence of the position coordinates, such as the offset direction of the coordinates at the next moment relative to the previous moment, the target direction is determined to be "west", providing a direction basis for the selection of subsequent camera or tag reader.

[0086] Pre-installed cameras or RFID tag readers along the road, such as one every 20 meters, at a height of 1.5 meters, with a coverage radius of 8 meters, and their position information (latitude and longitude) and coverage range are recorded in the system database. After determining the target route and direction, the system filters out the readers located on the route and in front of the target direction from the database. For example, if the target route is "from east to west", all RFID readers distributed along the route on the west side of the target current position are selected.

[0087] Calculate the straight-line distance between each selected reader and the target current position (latitude and longitude), and select the one closest to the target as the priority reading device. For example, if the target current position is (x0, y0), the three readers on the west side are 5 meters, 10 meters, and 15 meters away, respectively, then the reader 5 meters away is activated first to ensure that the tag reading is done in the first time when the target reaches the coverage range of the reader.

[0088] The target feature (such as a non-motor vehicle) is pre-installed with an RFID electronic tag, which stores unique identification information (such as vehicle number, owner information, etc.). When the target enters the coverage range of the selected reader, the reader transmits a radio frequency signal to activate the tag and receives the identification information returned by the tag to complete the reading operation. For example, after the RFID tag of the delivery vehicle in the express delivery and takeout industry is activated, the reader can obtain its corresponding "rider ID + enterprise code + vehicle registration number" and other data.

[0089] If the reader successfully obtains the tag information (i.e. returns complete identification data), the system associates the data with the identified target feature (such as vehicle shape, driving state) to generate a complete record containing "illegal time, illegal location, target feature, tag data", and uploads it to the back-end database of the traffic management department. For example, the record content is "2023-10-0109:30, XX intersection motor vehicle lane, electric bicycle or motorcycle, rider ID: 12345, enterprise code: ABC takeout".

[0090] If the reader does not obtain the tag information, such as tag damage, falling off or the target not installing the tag, the system determines that the target feature does not have traceable tag data, at which time the obtained on-site image (containing the target close-up), the extracted target feature (such as a number plate, vehicle color) and the calculated position information are packaged and uploaded to the background. For example, for an electric bicycle or motorcycle without a tag, the uploaded information includes "vehicle close-up image (containing a blurred number plate), time and place of entering the motor vehicle lane, driving trajectory", ensuring that even without a tag, the illegal behavior can be recorded through image recording.

[0091] The method can identify illegal behavior of non-motor vehicles on motor vehicle lanes in real time and provide on-site early warning, timely stop illegal behavior to reduce safety hazards, and significantly enhance the timeliness of management; with the help of hierarchical image recognition, position calculation and route direction analysis, combined with RFID and other tag readers for double verification, the illegal subject and behavior can be accurately locked, the identification accuracy is greatly improved, and the misjudgment probability is reduced; through this method, illegal behavior can be intervened in real time, and tag data can be used to implement post accountability, and targets without tags can also be recorded, solving the problem of incomplete coverage in traditional management; at the same time, relying on existing monitoring equipment and tag readers, the automatic process reduces hardware investment and labor costs, and effectively improves management efficiency.

[0092] In the step of identifying the target feature based on the target area, the following sub-steps are further included:

[0093] The number plate feature information is extracted from the moving target in the target area; through edge detection and contour recognition technology, the number plate area is framed in the vehicle image, such as the rectangular identification area of the tail or the head of the electric bicycle or motorcycle, excluding interference elements such as vehicle body advertisements and decorative stickers. For example, for electric bicycles or motorcycles, the number plate is usually a rectangular plate with blue background and white characters, fixed on the rear license plate holder; for gasoline motorcycles, it is a yellow background with black characters or a white background with black characters, and the image segment of the area can be accurately intercepted through positioning algorithm; the general resolution is adjusted to 300x100 pixels to ensure clear characters. The positioned number plate image is preprocessed, including grayscale, deblurring and tilt correction, and then the characters (including numbers, letters and special symbols) on the number plate are extracted one by one through character segmentation algorithm to form structured feature information.

[0094] Based on the preset coding rules and identification feature library, the extracted number plate information is analyzed to accurately distinguish electric bicycles or motorcycles:

[0095] The system has built-in non-motor vehicle and motor vehicle number plate coding rules input in advance. If the coding rules are ambiguous, such as non-uniform number plate formats in the transition period of some areas, the number plate physical characteristics are further distinguished.

[0096] ReferenceFigure 2 , the method further comprises the following steps:

[0097] After distinguishing that the moving target is an electric bicycle or a motorcycle, the system calls the extracted license plate feature information and performs structured processing thereon to extract key identification fields, including a region code, a serial number, and a special identification, such as a special number segment assigned to an electric vehicle of the express delivery and take-out industry in some regions. Specifically, the core fields associated with the affiliation information are filtered from the license plate information by a character matching algorithm to provide a data basis for subsequent association of the affiliation information.

[0098] The system presets a "license plate-affiliation information" mapping database, which stores the association between the license plate of an electric bicycle or a motorcycle and the subject (such as an operating company of the express delivery and take-out industry, a logistics enterprise, or a personal user).

[0099] For an electric vehicle of a personal user, the license plate serial number is associated with the identity card information of the vehicle owner.

[0100] When the input license plate feature information is input, the system obtains the corresponding affiliation information, such as "affiliation enterprise: XX delivery limited company, enterprise code: WMD001", by a combination of fuzzy matching and accurate query, such as matching the region code to narrow the range and then accurately retrieving by the serial number.

[0101] The system background establishes an association library of affiliation information and early warning records, and records the illegal early warning history of the electric bicycle or motorcycle of each subject in real time, including the early warning time, the illegal location, and the illegal behavior type, such as "driving on a motor vehicle lane" and "running a red light". When the affiliation information is obtained, the system automatically queries the cumulative early warning record quantity of the subject within a preset period (such as the last 30 days). For example, it is found that "XX delivery limited company" has been warned 15 times in the last 30 days due to "non-motor vehicle driving into motor vehicle lane", that is, the early warning record quantity of the enterprise is 15.

[0102] The system presets an early warning reference quantity for different subjects according to the average level of the industry and management needs, such as setting the early warning reference quantity of the enterprise of the express delivery and take-out industry to 10 times per 30 days. When the queried early warning record quantity (such as 15 times) is greater than the early warning reference quantity (10 times), the warning information sending mechanism is triggered:

[0103] Directly pushed to the management background (such as the enterprise safety management system) of the affiliated enterprise and the mobile terminal bound to the person in charge of the enterprise, such as a short message, a special APP notification. The sending information content includes "early warning exceeding reminder" (such as "your company has 15 early warning records in the last 30 days, which exceeds the reference value of 10 times"), "illegal details", and "rectification requirements" (such as "please complete the rider safety education within 3 days, and will be jointly checked by law enforcement officers if overdue"); the triple push of "system pop-up window + short message + email" can be used to ensure that the enterprise receives the information in a timely manner.

[0104] By associating the pre-warning record with the information about the target (e.g., after the staff of express delivery and take-out industry is identified to drive non-motor vehicles on the motor vehicle lane, the system can automatically obtain the information about the operating company to which the staff belongs), the enterprise with frequent illegal problems in the express delivery and take-out industry can be quickly located, and the management measures of the traffic control department are more targeted. At the same time, sending warning information to the enterprise helps to promote the enterprise to strengthen internal management, effectively educate and restrict the illegal behavior of employees, and reduce the occurrence of illegal phenomena from the source.

[0105] If the target tag method is not read, the following sub-steps are also included in the step:

[0106] When the camera or tag reader (such as an RFID reader) closest to the target position fails to successfully read the target tag on the target feature, for example, due to the tag signal being blocked, temporary equipment failure, or the target moving too fast causing the reading to time out, the system automatically triggers the next-in-line activation mechanism:

[0107] Based on the calculated target route (such as "drive along XX trunk road from east to west") and target direction (such as "west"), the camera or tag reader located on the target route and in front of the current reader (i.e., the next node in the target direction of travel) is selected from the preset camera or tag reader deployment database as the "next-in-line reader". For example, the current reader is device R1 deployed 50 meters west of intersection A, and the target direction is west, so the next-in-line reader is device R2 30 meters west of intersection B (the distance between R1 along the target route is 80 meters).

[0108] The system sends an activation instruction to the next-in-line reader through a wired network (such as optical fiber) or a wireless network (such as 4G / 5G), which contains the real-time position, moving speed, and tag signal characteristics (such as tag frequency, encoding format) of the target feature, so that the reader enters an active state in advance (such as increasing the receiving sensitivity from the default -85dBm to -90dBm, and shortening the signal response time to within 100ms).

[0109] The target tag is read by the next-in-line camera or tag reader: after receiving the activation instruction, the next-in-line reader immediately adjusts the reading parameters for the target feature's travel trajectory:

[0110] According to the target speed (such as 20 km / h), the estimated time for it to reach the reader's coverage range (radius 10 meters) is calculated, such as 14.4 seconds for a distance of 80 meters, and the continuous scanning mode is started in advance, with the scanning frequency increased from 1 Hz to 5 Hz, to ensure that the target enters the range and is captured immediately.

[0111] The reader emits radio frequency signals through a directional antenna (in the direction of the target route), while recording data such as signal strength, reflection time, etc. If the signal that meets the target tag coding rules, such as "16-bit binary code with prefix 'WM'", is captured for 3 consecutive times, it is determined that the reading is successful. If it is still not read, the next order reader (such as R3) is started, and this process continues until the reading is successful or the maximum order number is reached, such as 3 times.

[0112] Through the order starting mechanism, information loss caused by single reader failure or transient interference is avoided. For high-speed moving non-motor vehicles, such as delivery vehicles in the express delivery and take-out industries, the "relay" reading ensures that the target is always within the supervision coverage range throughout the entire route, solving the problem of "coverage blind area" in traditional single device reading, and making the traceability chain of illegal behavior more complete.

[0113] The method further comprises the following steps:

[0114] The target speed is calculated according to the target position and the recording time corresponding to the target position. Two sets of continuous position data (such as (x1, y1, t1) and (x2, y2, t2)) with a time interval Δt ≥ 1 second are selected to avoid calculation errors caused by excessive sampling. For example, if the target position is (120.001°E, 30.001°N) at t1=09:00:00 and the position is (120.002°E, 30.001°N) at t2=09:00:02, then Δt=2 seconds. Based on the latitude and longitude coordinate conversion formula (such as the Haversine formula), the distance s (unit: meters) between the two points on the earth's surface is calculated. Taking the above coordinates as an example, s=111 meters (approximate value). The instantaneous speed is converted to km / h unit by the formula v=(s / Δt)×3.6. In the above example, v=(111 / 2)×3.6=199.8 km / h, which is an extreme example. The actual speed of a non-motor vehicle is usually ≤40 km / h.

[0115] The number of cameras or tag readers to be started is determined based on the target speed, and the number is positively related to the target speed. The larger the number, the greater the target speed; the smaller the number, the smaller the target speed. The specific dynamic relationship is as follows:

[0116] When v≤10 (km / h), start 1 camera or tag reader, and select the camera or tag reader closest to the target position;

[0117] When 10 (km / h) < v≤20 (km / h), start 2 cameras or tag readers, and select the 2 closest consecutive order cameras or tag readers;

[0118] When 20 (km / h) < v≤30 (km / h), start 3 cameras or tag readers, select the nearest 3 consecutive order cameras or tag readers;

[0119] When v>30 (km / h), start 4 cameras or tag readers, select the nearest 4 consecutive order cameras or tag readers.

[0120] At the same time, start the corresponding number of cameras or tag readers located on the target route and in front of the target direction, and read the target label on the target feature. Synchronous triggering mechanism: through the millisecond time synchronization protocol (such as NTP protocol) to ensure that the selected N readers enter the active state at the same time (error≤50ms), avoiding missing reading caused by start time difference. The reader adjusts the working parameters according to the target speed-for example, for a target with v>30km / h, the signal transmission power is increased from 1W to 2W, the reading sensitivity is optimized from-85dBm to-90dBm, and the effective identification distance is extended to 15 meters. Multiple readers are deployed along the target route (such as 20 meters each), forming a continuous coverage band. For example, 3 readers can cover a 60-meter road section, ensuring that a target with a speed of 20km / h (10.8 seconds are needed to pass 60 meters) has sufficient time to be read by at least one device.

[0121] Increasing the number of readers reduces the missing reading rate and improves the reading power of high-speed targets; for low-speed targets, reducing the number can reduce energy consumption and avoid resource waste.

[0122] In other embodiments, the step of sending warning information according to the ownership information is followed by the following steps:

[0123] Calculate the warning ratio of the warning record quantity and the warning reference quantity; retrieve the warning record quantity (denoted as A) of the ownership information in the preset statistical period (such as the last 30 days) from the system background database, for example, a certain enterprise has accumulated 25 warnings in the last 30 days; At the same time, extract the preset warning reference quantity (denoted as B), which is dynamically adjusted according to the industry average level, enterprise size and historical illegal data, for example, the warning reference quantity of the express delivery and take-out industry is set to 10 times / 30 days. Calculate the quantitative ratio by the formula "warning ratio R=A / B", in the above example R=25 / 10=2.5, which directly reflects that the illegal risk of the enterprise exceeds the reference level by a multiple.

[0124] Based on the warning ratio, determine the number of cameras or tag readers to be started, which is positively related to the warning ratio; the larger the number, the larger the warning ratio; the smaller the number, the smaller the warning ratio. Dynamically allocate reading resources according to the ratio, the specific rules are as follows (taking the city main road scene as an example):

[0125] When 1 < early warning ratio R ≤ 2, start 2 cameras or tag readers, select the 2 closest continuous order cameras or tag readers from the target position;

[0126] When 2 < early warning ratio R ≤ 3, start 3 cameras or tag readers, select the 3 closest continuous order cameras or tag readers from the target position;

[0127] When 3 < early warning ratio R, start 4 cameras or tag readers, select the 4 closest continuous order cameras or tag readers from the target position.

[0128] At the same time, start the corresponding number of cameras or tag readers located on the target route and in front of the target direction, and read the target label on the target feature. Send a synchronous start instruction to the selected N readers through an industrial Ethernet or 5G private network to ensure that all devices enter the working state at the same time within 100ms, avoiding reading gaps caused by start time differences. For high early warning ratio enterprise vehicles, the reader automatically adjusts the working parameters, such as expanding the signal receiving bandwidth from 2MHz to 4MHz to resist interference, and extending the label response time threshold from 50ms to 100ms to adapt to possible label aging problems. Multiple readers are deployed in a staggered and complementary manner, such as 3 readers installed at 1.2m, 1.5m, and 1.8m heights on the road stand, forming a three-dimensional coverage area, reducing reading dead angles caused by vehicle tilting and label blocking.

[0129] Through the above method, the dynamic matching of supervision resources and enterprise illegal risk is realized, which not only ensures the information capture quality of high-risk objects, but also improves the operation efficiency of the overall supervision system.

[0130] With reference to Figure 3 , the step of reading the target label further includes the following sub-steps:

[0131] During the process of reading the target label through the tag reader, all activated tag sensors (such as RFID radio frequency sensors) generate and upload reading data in real time. The data content is divided into three categories according to the capture integrity of the label information:

[0132] Complete reading: the sensor successfully obtains all the identification information of the target label (such as 16-bit character code containing unique device number, subject code, and record time), and the data is correct (passing CRC cyclic redundancy check);

[0133] Partial reading: the sensor only captures part of the label information field (such as only the first 8 digits), or the data has a verification error (such as missing characters, abnormal format);

[0134] Not read: The sensor does not receive any signal returned by the tag (such as signal shielding, tag failure), return empty data.

[0135] For example, when 3 tag sensors are started, the combination data of "2 complete readings, 1 partial reading" and "1 complete reading, 1 partial reading, 1 not read" may be generated.

[0136] Calculate the reading completeness = the number of complete reading tag sensors / (the number of partial reading tag sensors + the number of not read tag sensors).

[0137] If 4 sensors are started in a certain scene, 3 of which are "complete reading" and 1 is "partial reading", then the reading completeness = 3 / 1 = 3.0;

[0138] If 2 sensors are "complete reading", 1 is "partial reading", and 1 is "not read", then the reading completeness = 2 / (1+1) = 1.0;

[0139] If 0 is "complete reading", 2 is "partial reading", and 1 is "not read", then the reading completeness = 0 / 3 = 0.

[0140] This index quantitatively reflects the effectiveness of multi-sensor cooperative reading, and the higher the value, the more reliable the quality of tag information capture.

[0141] If the reading completeness is greater than the preset reference completeness, it is read to the target tag. The system presets the reference completeness threshold, which is dynamically adjusted according to the application scenario:

[0142] For high-risk scenarios, such as enterprise vehicles with early warning ratio R>3 and high-speed moving targets with target speed v>30km / h, the reference completeness is set to 2.0, that is, the number of complete readings must be at least 2 times the number of incomplete readings.

[0143] For regular scenarios, such as ordinary non-motor vehicles and low-speed moving targets, the reference completeness is set to 1.0, that is, the number of complete readings is not less than the number of incomplete readings.

[0144] When the calculated reading completeness is greater than the preset reference completeness, it is determined that the target tag is read, and the system accepts the complete reading tag information; otherwise, it is determined that it is not effectively read, and the next order reader starting mechanism is triggered.

[0145] Through multi-sensor redundancy verification, false positives caused by single device failure or transient interference are avoided. The completeness index replaces the traditional binary judgment of "success / failure", allowing a certain proportion of sensor reading abnormalities. Combined with the early warning ratio, target speed and other parameters to adjust the reference completeness threshold, high-risk high-standard, low-risk moderate relaxation is achieved.

[0146] The embodiment of the present application also discloses a mobile target data analysis system based on a field environment, comprising a processor, wherein the steps of the mobile target data analysis method based on the field environment are executed.

[0147] The embodiment of the present application also discloses a storage medium, wherein a program is stored in the storage medium, and the program is executed by a processor to realize the steps of the mobile target data analysis method based on the field environment.

[0148] Although the embodiments of the present application have been shown and described above, it should be understood by those skilled in the art that the above embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

Claims

1. A method for analyzing mobile target data based on a field environment, characterized by, The method comprises the following steps: setting a security code corresponding to a plurality of color feature representations for a target feature, each of the security codes corresponding to a different security level; acquiring a site to which the target feature belongs, and calculating first distribution data of the security code corresponding to the target feature owned by the site; calculating a dynamic level corresponding to the site according to the first distribution data; acquiring a running platform corresponding to the site, and calculating second distribution data of the dynamic level corresponding to the site owned by the running platform; calculating a risk level corresponding to the running platform according to the second distribution data; real-time acquisition of a target label of the target feature through label reading auxiliary camera video algorithm recognition, and if the content of the label information of the target label corresponds to illegal behavior, the security level of the target feature is reduced, the security code of the target feature is updated according to the security level, and an execution log is generated; uploading the label information and the execution log to a management background; the management background updates the dynamic level and the risk level according to the label information and the execution log; if the dynamic level is lower than a preset dynamic reference level, a site early warning is triggered; if the risk level is lower than a preset risk reference level, a platform early warning is triggered.

2. The on-premise environment-based mobile target data analysis method of claim 1, wherein, The step of real-time acquisition of the target label of the target feature further comprises the following sub-steps: acquiring a live image; identifying a target area from the live image through a camera video algorithm; identifying a target feature based on the target area; if a target feature is identified, a target position corresponding to the target feature is calculated based on the live image, and a target early warning is performed; a target route and a target direction are calculated according to a plurality of target positions; a camera or a label reader located on the target route and in front of the target direction is acquired; the camera or the label reader closest to the target position is selected; a target label on the target feature is read through the camera or the label reader; if the target label is read, target data corresponding to the target label is acquired, and the target data is uploaded; if the target label is not read, it is prompted that the target feature does not have the target data, and the live image and the target feature are uploaded.

3. The on-premise environment-based mobile target data analysis method of claim 2, wherein, The step of identifying a target feature based on the target area further comprises the following sub-steps: extracting a number plate feature information from a mobile target in the target area; analyzing the number plate feature information, and distinguishing the mobile target as an electric bicycle or a motorcycle according to a coding rule or an identification feature of a number plate.

4. The on-premise environment-based mobile target data analysis method of claim 3, wherein, The method further comprises the following steps: if the mobile target is an electric bicycle or a motorcycle, the number plate feature information of the electric bicycle or the motorcycle is acquired; corresponding site information is acquired according to the number plate feature information; an early warning record amount is queried based on the site information; if the early warning record amount of the site information is greater than a preset early warning reference amount, warning information is sent according to the site information.

5. The live environment-based mobile target data analysis method of claim 2, wherein, The step of the method if the target label is not read further comprises the following sub-steps: Triggering a camera or a tag reader next in order located on the target route and in front of the target direction, and reading a target tag on the target feature by the camera or the tag reader.

6. The live environment-based mobile target data analysis method of claim 2, wherein, The method further comprises the following steps: calculating a target speed according to the target position and a recording time corresponding to the target position; determining a number of cameras or tag readers to be triggered based on the target speed, the number being in positive correlation with the target speed; the greater the number, the greater the target speed; the smaller the number, the smaller the target speed; simultaneously triggering the corresponding number of cameras or tag readers located on the target route and in front of the target direction to read the target tag on the target feature.

7. The mobile target data analysis method based on a live environment according to claim 4, characterized by, After the step of sending the warning information according to the belonging information, the method further comprises the following steps: calculating a warning ratio of the warning record amount and the warning reference amount; determining a number of cameras or tag readers to be triggered based on the warning ratio, the number being in positive correlation with the warning ratio; the greater the number, the greater the warning ratio; the smaller the number, the smaller the warning ratio; simultaneously triggering the corresponding number of cameras or tag readers located on the target route and in front of the target direction to read the target tag on the target feature.

8. The mobile target data analysis method based on a live environment according to claim 2, characterized by, In the step of reading the target tag, the method further comprises the following sub-steps: the triggered tag sensors all generate reading data, the content of the reading data including complete reading, partial reading, and no reading; calculating a reading completeness = the number of tag sensors with complete reading / (the number of tag sensors with partial reading + the number of tag sensors with no reading); if the reading completeness > a preset reference completeness, the target tag is read.

9. A mobile target data analysis system based on a field environment, characterized by, The processor executes the steps of the method for analyzing mobile target data based on a live environment according to any one of claims 1-8.

10. A storage medium, characterized by The storage medium stores a program, and the program is executed by the processor to realize the steps of the method for analyzing mobile target data based on a live environment according to any one of claims 1-8.

Images