A construction worker safety interlocking method and system based on single-entry transaction and multi-source space-time binding

Abstract

The application discloses a kind of based on single entry transaction and multi-source space-time binding construction personnel safety interlocking method and system, it is related to construction site personnel management technical field, technical scheme is, including obtaining the multi-source perception data in access identification area, the space-time tracking trajectory of personnel to be entered is constructed and single entry transaction and transaction identification are generated;Identity verification, safety equipment detection and personnel state check result are bound with transaction identification, determine transaction state according to the effective state of each result and the continuous state of track, generate one-time release instruction when meeting the condition;Control channel traffic mechanism action, determine actual traffic result according to channel detection data and action feedback, close transaction when normal traffic, revoke instruction and update transaction state when abnormal traffic or track interruption.The application can reduce the detection result mismatch and reuse when multiple people continuously enter, realize the closed-loop interlocking of personnel verification, safety equipment detection, state check and actual traffic.

Description

Technical Field

[0001] This invention relates to the field of construction site personnel management technology, and in particular to a method and system for safety interlocking of construction personnel based on single entry transactions and multi-source spatiotemporal binding. Background Technology

[0002] Construction sites are typically characterized by a large number of personnel, frequent movement, and complex job types and work areas. To implement real-name management and safety production requirements, construction workers usually need to complete identity verification, safety training status check, job qualification verification, and safety equipment such as safety helmets and reflective vests before entering the site. Currently, construction sites generally manage personnel entry through real-name gates, facial recognition terminals, document reading devices, and a back-end management platform. After personnel swipe their cards, ID cards, or undergo facial recognition, the control equipment sends a release command to the gate.

[0003] In existing access management systems, identity verification, security equipment detection, and personnel status verification are typically performed separately by different devices or processing units. The results of each detection are primarily stored temporarily based on their generation time, lacking a data carrier capable of continuously linking multiple detection results to the same person and the same entry behavior. When multiple people enter consecutively, when human targets obscure each other, or when personnel move back and forth within the identification area, the system struggles to accurately determine whether detection results from different sources belong to the same person waiting to enter. This can easily lead to situations where the identity verification result, security equipment detection result, or personnel status verification result of a previous person is misused by a subsequent person.

[0004] Especially in the process of safety equipment inspection, existing technologies typically determine whether a person meets the wearing requirements based on the presence of targets such as safety helmets and reflective vests in a single frame image. When multiple people are standing adjacent to each other or when image areas overlap, the safety equipment detected in the image may not be worn by the person currently undergoing identity verification. Furthermore, if a person holds a safety helmet, briefly raises it near their head, or uses the safety equipment of an adjacent person to enter the inspection area, incorrect results may be obtained indicating that the safety equipment is qualified. Existing technologies lack a processing mechanism to determine the attribution of safety equipment based on the continuous correspondence between personnel and safety equipment in terms of spatial location, time of appearance, and movement.

[0005] Furthermore, identity verification results, safety equipment inspection results, and personnel status verification results have different time limits. Personnel may remove their safety helmets after completing safety equipment inspections, and the personnel status or the risk status of the construction area may change while waiting to pass. Existing systems typically retain the corresponding passage status after the inspection results are generated, without updating the release conditions based on whether the personnel tracking trajectory is continuous, whether the inspection results have exceeded their validity period, and whether the safety equipment status has changed. This makes it easy to use inspection results that are no longer valid or correspond to the current personnel to control the opening of the passage.

[0006] Existing turnstiles typically receive simple open or close signals, lacking a clear link between the pass signal and personnel identity, entry transaction, target channel, and permitted number of passes. During turnstile operation, situations may arise such as unverified personnel tailgating or verified personnel leaving and being replaced by others. Furthermore, some systems record passage completion immediately after issuing the opening command, without considering entrance and exit detection data, the actual direction of personnel movement, and turnstile feedback to confirm whether a person with access authorization has actually completed a normal passage. Therefore, it is difficult to identify tailgating, reverse passage, personnel delays, personnel substitutions, and abnormal turnstile operation.

[0007] Therefore, existing construction worker entry management technologies still suffer from problems such as unstable correlation between multi-source detection results and personnel and their single entry behavior, easy mismatch between detection results of different personnel, continued use of expired detection results, and lack of closed-loop correspondence between release instructions and actual personnel passage behavior. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a method and system for safety interlocking of construction workers based on single entry transactions and multi-source spatiotemporal binding.

[0009] Includes the following steps: S1. Acquire multi-source sensing data within the channel identification area, construct the spatiotemporal tracking trajectory of the person to be admitted, and generate a single admission transaction and transaction identifier accordingly. S2. Collect and verify the identity information of the personnel waiting to enter, obtain the identity verification result and personnel identifier, and bind it with the transaction identifier; S3. Identify security equipment based on multi-source sensing data and obtain security equipment detection results, determine the spatiotemporal relationship between security equipment and personnel waiting to enter the site, and bind the security equipment detection results with transaction identifiers; S4. Obtain and verify personnel status information based on personnel identifier, obtain personnel status verification results, and bind them with transaction identifier; S5. Determine the transaction status based on the identity verification result, security equipment test result, personnel status verification result, and the continuous status of the spatiotemporal tracking trajectory bound to the transaction identifier, and generate a one-time release instruction when permission is granted. S6. Control the passage mechanism according to the one-time release command, and determine the actual passage result based on the passage detection data and action feedback; S7. When passage is normal, invalidate the one-time release command and close the single entry transaction; when passage is abnormal or the spatiotemporal tracking trajectory is interrupted, cancel the one-time release command and update the single entry transaction to an abnormal state.

[0010] Furthermore, the multi-source sensing data includes at least personnel image data and channel space detection data; the personnel image data and channel space detection data are synchronized in time, the human targets of the personnel to be entered are identified from the personnel image data, and the human targets are tracked in association based on their spatial position and motion state at continuous time intervals to obtain a spatiotemporal tracking trajectory with trajectory identification.

[0011] Furthermore, when the spatiotemporal tracking trajectory first enters the preset transaction triggering area, a single entry transaction is created and a transaction identifier is generated, and the transaction identifier is uniquely associated with the trajectory identifier; before the single entry transaction is closed, no new single entry transaction is created based on the same spatiotemporal tracking trajectory.

[0012] Furthermore, during the identity verification process, the environmental quality parameters of the channel recognition area and the facial occlusion status of the person waiting to enter are obtained. Based on the environmental quality parameters and facial occlusion status, the target verification method is determined from a variety of preset identity verification methods, and the target verification method is used for identity verification. When the verification credibility is lower than a preset threshold, an identity verification method different from the target verification method is called for review. The identity verification result and personnel identification are determined based on the verification results of each identity verification method.

[0013] Furthermore, the identified security equipment is continuously tracked based on multi-source sensing data to obtain the security equipment trajectory; the spatiotemporal association result is determined based on the position matching relationship, time overlap relationship and motion consistency relationship between the security equipment trajectory and the spatiotemporal tracking trajectory, and when the spatiotemporal association result meets the preset association conditions, the corresponding security equipment detection result is bound to the transaction identifier.

[0014] Furthermore, the personnel status information includes at least one of the following: safety training status, job qualification status, work group authorization status, and prohibited entry status; obtain the target construction area and its operational risk status corresponding to the personnel to be admitted, determine the preset entry rules based on the personnel category and operational risk status of the personnel to be admitted, and verify the personnel status information according to the preset entry rules.

[0015] Furthermore, the generation time and validity period of the identity verification results, security equipment test results, and personnel status verification results are recorded respectively. The validity status of each result is determined based on whether each result is within its corresponding validity period and whether the spatiotemporal tracking trajectory is continuous. When any result exceeds its corresponding validity period, the spatiotemporal tracking trajectory is interrupted, or the security equipment status changes, the corresponding result becomes invalid, and the transaction status is updated based on the validity status of each result.

[0016] Furthermore, the one-time release instruction includes a transaction identifier, a target channel access authority identifier, the number of allowed passages, and an instruction validity period. The one-time release instruction is only used to control the corresponding target channel access authority to complete one passage, and becomes invalid when passage is completed, the instruction validity period expires, the spatiotemporal tracking trajectory is interrupted, or the transaction status changes.

[0017] Furthermore, a first detection area and a second detection area are set up along the direction of personnel passage. The number of people entering, leaving and the direction of passage are determined based on the channel detection data generated by the first detection area and the second detection area. The actual passage result is determined in combination with the action feedback of the channel passage mechanism. When the number of people entering and leaving is consistent with the number of allowed passages and the direction of passage is normal, it is determined as normal passage; otherwise, it is determined as abnormal passage.

[0018] This invention also provides a construction worker safety interlocking system based on single entry transactions and multi-source spatiotemporal binding, comprising: The multi-source sensing module is used to acquire multi-source sensing data within the channel identification area; The trajectory and transaction generation module is used to construct the spatiotemporal tracking trajectory of the personnel to be admitted based on multi-source sensing data, and generate a single admission transaction and its corresponding transaction identifier based on the spatiotemporal tracking trajectory. The identity verification module is used to collect and verify the identity information of people waiting to enter, obtain the identity verification result and personnel identifier, and bind the identity verification result and personnel identifier with the transaction identifier. The security equipment association module is used to identify security equipment based on multi-source sensing data, determine the spatiotemporal relationship between security equipment and personnel waiting to enter the site, and bind the corresponding security equipment detection results with transaction identifiers. The personnel status verification module is used to obtain and verify the personnel status information of the personnel to be admitted based on the personnel identifier, obtain the personnel status verification result, and bind the personnel status verification result with the transaction identifier. The transaction status control module is used to determine the transaction status based on the identity verification result bound to the transaction identifier, the security equipment detection result, the personnel status verification result, and the continuous status of the spatiotemporal tracking trajectory, and to generate a one-time release instruction when the transaction status is the allowable release status. The passage control module is used to control the movement of the passage mechanism according to a one-time release command; The passage result processing module is used to determine the actual passage result based on the channel detection data and the action feedback of the channel passage mechanism, and to close the single entry transaction or update the single entry transaction to an abnormal status based on the actual passage result.

[0019] The beneficial effects of the technical solution provided by the embodiments of the present invention are as follows: 1. By using a single entry transaction as a carrier, the spatiotemporal tracking trajectory of the person waiting to enter, the identity verification result, the security equipment detection result, the personnel status verification result, and the actual passage result are bound to the same transaction identifier, which can reduce the problem of mismatch and reuse of detection results between different people when multiple people enter continuously or when people block each other.

[0020] 2. The system determines the ownership of safety equipment based on the positional matching, temporal overlap, and consistent movement relationship between the trajectory of safety equipment and the spatiotemporal tracking trajectory of personnel. This system can distinguish between correctly worn and handheld safety equipment and interference from safety equipment of adjacent personnel, thereby improving the reliability of the correspondence between the safety equipment detection results and the personnel waiting to enter the site.

[0021] 3. Update the transaction status by combining the validity period of each verification result, the continuous status of personnel trajectory and the status changes of safety equipment, and adopt a one-time release instruction bound to the transaction identifier, target channel and allowed number of passes to avoid using expired or invalid verification results to control the passage mechanism.

[0022] 4. Based on the detection data at the entrance and exit of the passage and the action feedback of the passage passage mechanism, the actual passage result is determined. It can identify tailgating, reverse passage, personnel replacement, passage stagnation and abnormal mechanism action, so as to form a closed loop of identity verification, interlock release and actual passage. Attached Figure Description

[0023] Figure 1 This is a flowchart of the construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to the present invention; Figure 2 This is a structural block diagram of the construction worker safety interlocking system based on single entry transactions and multi-source spatiotemporal binding, as described in this invention. Figure 3 This is a schematic diagram illustrating the binding relationship between a single entry transaction and multi-source detection results in this invention; Figure 4 This is a schematic diagram of the access area for construction personnel and the passage detection process in this invention; Figure 5 This is a schematic diagram of the state transition of a single entry transaction in this invention. Detailed Implementation

[0024] Example 1 This embodiment provides a safety interlocking method for construction workers based on single entry transactions and multi-source spatiotemporal binding, which is applied to personnel entry channels at construction sites.

[0025] The personnel entry channel is equipped with a passage mechanism, which can be a swing gate, wing gate, tripod turnout, sliding door, or access control electric lock. In this embodiment, a swing gate is used. A passage identification area is formed on the entrance side of the passage mechanism. Image acquisition equipment, identity information acquisition equipment, and spatial detection equipment are installed in the passage identification area. The image acquisition equipment is used to collect image data of personnel waiting to enter and their security equipment. The identity information acquisition equipment is used to collect facial information, document information, or fingerprint information. The spatial detection equipment is used to detect the position, number, and direction of passage of personnel in the passage identification area.

[0026] Image acquisition equipment, identity information acquisition equipment, channel detection equipment, and channel access mechanism are all connected to the on-site control host. The on-site control host is connected to the back-end management platform, which stores the registration information, training information, job qualification information, work team authorization information, and access restriction information of construction personnel.

[0027] The method in this embodiment includes the following steps.

[0028] S1. Acquire multi-source sensing data within the channel identification area, construct the spatiotemporal tracking trajectory of the personnel to be admitted, and generate a single admission transaction and transaction identifier accordingly.

[0029] Specifically, when a person waiting to enters the channel recognition area, the image acquisition device collects the person's image data, the channel detection device collects the person's spatial detection data within the channel recognition area, the on-site control host synchronizes the person's image data and spatial detection data in time, and identifies the human target from the person's image data.

[0030] The on-site control host continuously tracks the human target based on its position, direction of movement, and speed at different times, forming a spatiotemporal tracking trajectory of the person to be entered, and generating a trajectory identifier for this spatiotemporal tracking trajectory.

[0031] When the time-space tracking trajectory first enters the preset transaction triggering area, the field control host creates a single entry transaction corresponding to the entry behavior, generates a transaction identifier, and associates the transaction identifier with the trajectory identifier, so that the identity verification results, security equipment detection results, personnel status verification results and passage results generated by the personnel to be entered during this entry process all belong to the same single entry transaction.

[0032] Before a single entry transaction is closed, the field control host will no longer create a new single entry transaction based on the same spatiotemporal tracking trajectory, thereby avoiding the formation of multiple independent verification records for the same person during a single entry process.

[0033] S2. Collect and verify the identity information of the personnel to be admitted, obtain the identity verification result and personnel identifier, and bind it with the transaction identifier.

[0034] Specifically, the identity information collection device collects the facial information and identification information of the people waiting to enter the venue. The on-site control host verifies the consistency between the collected facial information and the identification information, and matches the verification result with the personnel registration information pre-stored in the background management platform.

[0035] When the collected identity information matches the personnel registration information, the on-site control host generates an identity verification pass result and obtains the corresponding personnel identifier; when the collected identity information does not match the personnel registration information, an identity verification fail result is generated.

[0036] The on-site control host writes the identity verification result and personnel identification into the corresponding single entry transaction, and establishes the correlation between the identity verification result, personnel identification and spatiotemporal tracking trajectory through the transaction identification.

[0037] When a facial image is obscured by strong light, dust, a mask, or the brim of a safety helmet, resulting in insufficient credibility of facial verification, the on-site control host will prompt the person waiting to enter to supplement the verification with document verification or fingerprint verification, and determine the final identity verification result based on the supplementary verification results.

[0038] S3. Identify security equipment based on multi-source sensing data and obtain security equipment detection results, determine the spatiotemporal relationship between security equipment and personnel waiting to enter the site, and bind the security equipment detection results with transaction identifiers.

[0039] Specifically, the safety equipment to be detected in this embodiment includes safety helmets and reflective vests. The image acquisition device acquires images of the head area and upper body area of ​​the person to be entered. The on-site control host identifies the safety helmet target and the reflective vest target from the acquired images and determines whether the safety helmet and reflective vest are in the preset wearing position.

[0040] The on-site control host continuously tracks the identified safety helmet and reflective vest targets, forming the safety helmet trajectory and reflective vest trajectory respectively. Based on the position matching relationship, time overlap relationship and motion consistency relationship between the safety equipment trajectory and the spatiotemporal tracking trajectory of the personnel waiting to enter the site, it determines whether the identified safety equipment is being worn by the personnel currently waiting to enter the site.

[0041] When the safety equipment and the personnel waiting to enter meet the preset spatiotemporal association conditions, the on-site control host will bind the corresponding safety equipment detection result with the transaction identifier of the current single entry transaction; when the safety equipment is present in the image, but the safety equipment is not in the preset wearing position of the personnel waiting to enter, or the movement state of the safety equipment is inconsistent with that of the personnel waiting to enter, the safety equipment detection result will not be regarded as the valid detection result of the current personnel waiting to enter.

[0042] The above processing can prevent the misuse of safety equipment test results between adjacent personnel and can also identify situations such as personnel holding safety helmets or temporarily borrowing the test results of safety helmets from adjacent personnel.

[0043] S4. Obtain and verify personnel status information based on personnel identifier, obtain personnel status verification results, and bind them with transaction identifier.

[0044] Specifically, the on-site control host uses personnel identification as an index to obtain the personnel status information of personnel to be admitted from the back-end management platform or local cache. The personnel status information includes safety training status, job qualification status, team authorization status, and restricted entry status.

[0045] The on-site control host verifies the personnel status information based on the personnel category and the preset entry rules corresponding to the target construction area.

[0046] For example, ordinary construction workers need to meet the following conditions: their safety training is valid, their work team authorization is valid, and there are no prohibited entry signs. In addition to meeting the above conditions, special operation personnel also need to have their corresponding job qualification certificates be valid.

[0047] When the personnel status information meets the preset entry rules, a personnel status verification result that has passed the verification is generated; when any necessary status does not meet the preset entry rules, a personnel status verification result that has failed the verification is generated, and the specific reason for the failure is recorded.

[0048] The on-site control host binds the personnel status verification result with the transaction identifier of the current single entry transaction.

[0049] S5. Based on the identity verification results, security equipment detection results, personnel status verification results, and the continuous status of the spatiotemporal tracking trajectory bound to the transaction identifier, determine the transaction status and generate a one-time release instruction when permission is granted.

[0050] Specifically, a single entry transaction sequentially has the following states: pending verification, verification in progress, permission granted, passage in progress, normal closure, and abnormal status.

[0051] The on-site control host summarizes the identity verification results, safety equipment test results, and personnel status verification results, and determines whether each result is still valid and whether the spatiotemporal tracking trajectory of the personnel waiting to enter the site is continuous.

[0052] When the identity verification result is passed, the safety equipment test result meets the wearing requirements, the personnel status verification result is passed, and the spatiotemporal tracking trajectory remains continuous, the on-site control host will update the single entry transaction to the allowed release status and generate a one-time release instruction corresponding to the transaction identifier.

[0053] A one-time clearance instruction includes at least a transaction identifier, a target channel access authority identifier, a permitted number of passages, and an instruction validity period. In this embodiment, the permitted number of passages is one.

[0054] When any verification result fails, the verification result expires, or the spatiotemporal tracking trajectory is interrupted, the field control host will not generate a one-time release command, but will output corresponding prompt information through the display device or voice device.

[0055] S6. Control the passage mechanism according to the one-time release command, and determine the actual passage result based on the passage detection data and action feedback.

[0056] Specifically, the field control host sends a one-time release command to the passage mechanism. After receiving a valid one-time release command, the passage mechanism performs the opening action and sends the gate opening status, opening time, and reset status as action feedback to the field control host.

[0057] As people enter through the passage, the space detection equipment sequentially acquires passage detection data of people entering the passage, being inside the passage, and leaving the passage. The on-site control host determines the number of people entering, leaving, and the direction of passage based on the passage detection data.

[0058] The on-site control host compares the number of people entering, leaving, and the direction of passage with the number of passages allowed in the one-time release instruction, and combines the action feedback of the passage mechanism to determine the actual passage result corresponding to the current single entry transaction.

[0059] When the passage mechanism completes opening and resetting according to the one-time release instruction, and the number of people entering and leaving is consistent with the allowed number of passages, and the people pass through the passage in the predetermined direction, the actual passage result is normal passage.

[0060] When the number of people entering or leaving is inconsistent with the allowed number of passages, people pass in the wrong direction, people stay in the passage, or the passage mechanism fails to complete the action as instructed, the actual passage result is abnormal passage.

[0061] S7. Close or update a single entry transaction based on the actual passage results.

[0062] Specifically, when the actual passage result is normal passage, the on-site control host invalidates the one-time release command, updates the single entry transaction to the normal closed state, and records the personnel identification, transaction identification, entry time, safety equipment test results, personnel status verification results, and the action status of the passage mechanism.

[0063] When the actual passage result is abnormal, the on-site control host cancels the one-time release command, updates the single entry transaction to an abnormal state, controls the passage mechanism to remain closed or performs a reset, and outputs abnormal prompts or alarm information.

[0064] When a person waiting to enter leaves the channel identification area before completing the passage, or when the spatiotemporal tracking trajectory is interrupted due to obstruction and fails to recover within a preset time, the on-site control host terminates the current release process, cancels the one-time release instruction, and updates the corresponding single entry transaction to an abnormal status.

[0065] After a single entry transaction is closed, the on-site control host uploads the identity verification results, safety equipment test results, personnel status verification results, actual passage results, and abnormal information to the back-end management platform to form a complete entry record corresponding to this entry behavior.

[0066] Example 2 The access control area is divided into a pre-detection area, a transaction triggering area, and a passage waiting area along the personnel entry direction. The pre-detection area is located outside the transaction triggering area and is used to identify human targets and establish an initial spatiotemporal tracking trajectory before personnel enter the transaction triggering area. The transaction triggering area is used to trigger a single entry transaction. The passage waiting area is located between the transaction triggering area and the access control mechanism and is used for personnel waiting to enter to complete identity verification, security equipment inspection, and personnel status verification.

[0067] Multi-source sensing data includes at least personnel image data and channel spatial detection data. Personnel image data is acquired by a first image acquisition device positioned directly in front of the channel recognition area and a second image acquisition device positioned above the channel recognition area. The first image acquisition device is used to acquire frontal and upper-body images of the personnel about to enter, while the second image acquisition device is used to acquire the position of the personnel about to enter within the channel recognition area and the relative positions between different personnel.

[0068] The spatial detection data for the channel is acquired by spatial detection devices positioned on either side or above the channel, and is used to determine the position, direction of movement, and occupancy status of human targets. These spatial detection devices can be one or more of a depth camera, laser rangefinder, millimeter-wave detection device, or infrared array detection device, used to detect the spatial position, direction of movement, and occupancy status of human targets within the channel recognition area.

[0069] The first image acquisition device, the second image acquisition device, and the spatial detection device each add a timestamp to the acquired data. The field control host corrects the timestamps of each device according to a unified time base. For any image frame in the personnel image data, the field control host selects the data from the channel spatial detection data whose acquisition time is closest to that image frame and whose time difference is within a preset synchronization range, forming the fused perception data for that time.

[0070] When different sensing devices use different internal clocks, the field control host corrects the device timestamps based on the time deviation between each sensing device and a unified time reference. The corrected acquisition time is represented as:

[0071] in, Indicates the first The first sensing device collects the data. The device timestamp generated when grouping data. Indicates the first Time correction of each sensing device relative to a unified time base This indicates the corrected data collection time.

[0072] The on-site control host identifies human targets from personnel image data and determines the spatial location of the human target within the channel recognition area by combining the channel spatial detection data.

[0073] For the distance, outline, or area occupancy position of the human body obtained by the spatial detection equipment, the on-site control host matches it with the spatial position obtained from the personnel image data. When the position difference between the two is within a preset range, the corresponding data is determined to belong to the same human target, and the spatial position of the human target is corrected using the channel spatial detection data.

[0074] The on-site control host obtains the spatial position of the human target at each continuous acquisition moment, determines the movement speed and direction of the human target based on the changes in the human target's position at adjacent acquisition moments, and arranges the spatial position, movement speed, movement direction and human characteristics of the human target at each acquisition moment in chronological order to form a candidate spatiotemporal tracking trajectory.

[0075] For a human target detected at the current acquisition time, the on-site control host determines the degree of proximity of the human target to the existing candidate spatiotemporal tracking trajectory, the degree of consistency of the movement speed, the degree of consistency of the movement direction, and the degree of similarity of the human appearance features, and determines the trajectory association result accordingly.

[0076] No. Personal goals and the first The trajectory correlation between candidate spatiotemporal tracking trajectories is expressed as:

[0077] in, Indicates the degree of proximity in location. Indicates the degree of consistency in movement speed. Indicates the degree of consistency in the direction of movement. Indicates the degree of similarity in human physical appearance. , , and Each of these represents a weight, and the sum of all weights is 1.

[0078] When the trajectory correlation reaches the preset trajectory correlation threshold, the perception data of the current human target is updated to the corresponding candidate spatiotemporal tracking trajectory; when the trajectory correlation between the current human target and each candidate spatiotemporal tracking trajectory does not reach the preset trajectory correlation threshold, a new candidate spatiotemporal tracking trajectory is established for the current human target.

[0079] Once a candidate spatiotemporal tracking trajectory has been continuously maintained for a preset duration, or after a preset number of valid human targets have been continuously acquired, the on-site control host confirms the candidate spatiotemporal tracking trajectory as a valid spatiotemporal tracking trajectory and generates a trajectory identifier for it. The spatiotemporal tracking trajectory includes at least a trajectory identifier and the spatial location, movement speed, and movement direction of the personnel arranged in chronological order.

[0080] When multiple people enter the pre-detection area simultaneously, the on-site control host establishes spatiotemporal tracking trajectories for different human targets. In cases where human targets occlude each other, the on-site control host predicts the human position during the occlusion period based on the spatial position and motion state of each spatiotemporal tracking trajectory before the occlusion occurs, and combines top image data and channel spatial detection data to maintain the distinction between different trajectories.

[0081] After the occlusion ends, the on-site control host will associate the newly detected human target with the spatiotemporal tracking trajectory before the occlusion. If the position, direction of movement, and physical characteristics of the newly detected human target match the spatiotemporal tracking trajectory before the occlusion, the original trajectory identifier will continue to be used, and a new trajectory identifier will not be generated.

[0082] Example 3 The access control area is equipped with facial image capture devices, upper body image capture devices, document reading devices, and auxiliary identity collection devices. Facial image capture devices capture facial images of individuals waiting to enter; upper body image capture devices capture images of the body and security equipment of individuals waiting to enter; document reading devices read identity information from ID cards, real-name work cards, or other registered documents; and auxiliary identity collection devices can be fingerprint scanners, palm vein scanners, or other devices capable of outputting identity verification results.

[0083] The on-site control host pre-stores multiple identity verification methods, including at least facial verification, document verification, and biometric-assisted verification. Based on the environmental quality parameters of the channel recognition area and the facial occlusion status of the person waiting to enter, the on-site control host determines the target verification method to be used from among the multiple identity verification methods.

[0084] The on-site control host extracts parameters from the face image, including illumination uniformity, image clarity, local glare, image noise, and dust or lens contamination levels. Based on these parameters, it determines the environmental quality parameters, which are expressed as follows:

[0085] in, Indicates environmental quality parameters, Indicates the uniformity of illumination. Indicates the image sharpness. Indicates the degree of localized reflection. Indicates the level of image noise. Indicates the degree of dust or lens contamination. , , , and These represent the weights of the corresponding parameters.

[0086] Before calculation, the uniformity of illumination, image clarity, local reflectivity, image noise, and dust or lens contamination are normalized to eliminate differences in the dimensions of different parameters. The weights of each parameter are pre-calibrated based on sample images collected at the construction site.

[0087] The on-site control unit also identifies the eyes, nose, mouth, facial contours, and other key facial areas from the facial image. Based on the visibility of each key facial area, it determines the degree of facial occlusion of the person entering the venue. The degree of facial occlusion is expressed as follows:

[0088] in, Indicates the degree of facial occlusion. Indicates the first Area weighting of key facial regions Indicates the first The visibility state parameters of key facial regions are as follows: when the key facial regions can be fully identified, the corresponding visibility state parameters take a larger value; when the key facial regions are obscured by the brim of a safety helmet, mask, goggles, or other objects, the corresponding visibility state parameters take a smaller value.

[0089] When the environmental quality parameters reach the preset environmental threshold and the degree of facial occlusion is lower than the preset occlusion threshold, the on-site control host will determine facial verification as the target verification method and match the collected facial features with the facial features in the personnel registration information.

[0090] When the environmental quality parameters are lower than the preset environmental threshold, or when the degree of facial occlusion reaches the preset occlusion threshold, the on-site control host will determine the document verification as the target verification method, read the document information of the personnel to be admitted, and compare the document information with the personnel registration information stored in the back-end management platform.

[0091] When the verification credibility output by the target verification method is lower than the corresponding credibility threshold, the on-site control host calls a different identity verification method for review. For example, if the facial verification credibility is insufficient, the document reading device is called to read the ID card or real-name work card; if the document information is abnormal, the auxiliary identity collection device is called to collect fingerprint information or palm vein information.

[0092] For each identity verification method involved, the on-site control host obtains the corresponding verification result and verification credibility, and determines the overall identity credibility based on the reliability of each identity verification method in the current environment. The overall identity credibility is expressed as:

[0093] in, Indicates the overall credibility of the identity. Indicates the first The credibility of various identity verification methods Indicates the first The reliability of various identity verification methods in the current environment. Indicates the first Preset weights for various identity verification methods.

[0094] When the overall identity credibility reaches the preset identity threshold and the personnel identity obtained by different identity verification methods is consistent, the on-site control host generates an identity verification pass result and obtains the corresponding personnel identifier from the personnel registration information.

[0095] When the overall identity credibility does not reach the preset identity threshold, or when the identity of the person obtained by different identity verification methods is inconsistent, the on-site control host generates an identity verification failure result and prompts the person waiting to enter the site to re-collect identity information or transfer to manual review through the display screen and voice prompt device.

[0096] After the identity verification result and personnel identification are generated, the on-site control host searches for the transaction identification associated with the trajectory identification corresponding to the spatiotemporal tracking trajectory, and writes the identity verification result and personnel identification into the corresponding single entry transaction.

[0097] The on-site control unit identifies the safety equipment of personnel about to enter the site based on multi-source sensing data generated by upper body and top image acquisition devices. Safety equipment includes safety helmets and reflective vests, and may also include goggles, safety belts, face shields, or other personal protective equipment depending on the operational requirements of the target construction area.

[0098] The on-site control host identifies human targets and safety equipment targets from upper body image data, and determines the expected wearing area of ​​various safety equipment based on the head area, torso area and limb area of ​​the human target.

[0099] The expected wearing area for a safety helmet is the head area of ​​the human target; the expected wearing area for a reflective vest is the torso area of ​​the human target; the expected wearing area for goggles is the eye area of ​​the human target; and the expected wearing area for a seat belt is the waist and torso of the human target.

[0100] The on-site control host tracks and associates the safety equipment targets identified in continuous images, generating a safety equipment trajectory for each target. The safety equipment trajectory includes at least the safety equipment type, trajectory identifier, spatial location at each acquisition time, movement speed, movement direction, and wearing status.

[0101] For the The spacetime tracking trajectory and the first The safety equipment trajectory is determined by the on-site control host, which determines the position matching degree, time overlap degree, and motion consistency degree.

[0102] Position matching degree is determined based on the degree of overlap between the safety equipment target and the corresponding expected wearing area of ​​the human body, as well as the distance between their center positions. Taking a safety helmet as an example, the position matching degree between the safety helmet trajectory and the spatiotemporal tracking trajectory is expressed as:

[0103] in, Indicates the first The human head region corresponding to the spatiotemporal tracking trajectory. Indicates the first The target area of ​​the safety helmet corresponding to the trajectory of the safety equipment. The time overlap is determined based on the relationship between the duration of simultaneous existence of the spatiotemporal tracking trajectory and the safety equipment trajectory and the duration of both. The time overlap is expressed as:

[0104] in, Indicates the first The time interval corresponding to the spatiotemporal tracking trajectory Indicates the first The time interval corresponding to each safety equipment trajectory.

[0105] Motion consistency is determined based on the difference in movement speed and direction between the spatiotemporal tracking trajectory and the safety equipment trajectory at consecutive moments, as well as the positional change of the safety equipment relative to the intended wearing area. Motion consistency is expressed as:

[0106] in, and These represent the movement speeds of personnel waiting to enter and safety equipment, respectively. and These indicate the direction of movement for personnel and safety equipment waiting to enter, respectively. This indicates the amount of positional change of the safety equipment relative to the intended wearing area. , and These represent the corresponding scale parameters.

[0107] The on-site control host determines the spatiotemporal correlation between the spatiotemporal tracking trajectory and the safety equipment trajectory based on position matching degree, time overlap degree, and motion consistency degree. The spatiotemporal correlation degree is expressed as:

[0108] in, , γ and γ represent the association weights corresponding to position matching degree, time overlap degree and motion consistency degree, respectively, and the sum of the three is 1.

[0109] When the spatiotemporal correlation reaches the preset correlation threshold and remains at the preset correlation threshold for multiple consecutive data collection times, the field control host will determine the trajectory of the corresponding safety equipment as associated with the person to be admitted.

[0110] When multiple people are simultaneously located within the channel identification area, the on-site control host calculates the spatiotemporal correlation between each safety equipment trajectory and each spatiotemporal tracking trajectory. For the same safety equipment trajectory, the spatiotemporal tracking trajectory with the highest spatiotemporal correlation is selected as the candidate associated trajectory.

[0111] When the difference between the highest and second-highest spatiotemporal correlation reaches a preset distinction threshold, the safety equipment trajectory is associated with the candidate associated trajectory; when the difference between the highest and second-highest spatiotemporal correlation is lower than the preset distinction threshold, the safety equipment trajectory is not bound to any spatiotemporal tracking trajectory, and subsequent multi-source sensing data continues to be acquired.

[0112] The on-site control unit determines whether safety equipment is worn correctly based on the trajectory of the safety equipment associated with the personnel to be on site. For safety helmets, it determines whether the safety helmet target is continuously located in the head area of ​​the person and whether the relative position between the safety helmet target and the head area is stable; for reflective vests, it determines whether the reflective vest target covers the torso area of ​​the person and whether the reflective vest target moves synchronously with the torso.

[0113] When the safety helmet is located in the area of ​​a person's hands, or only passes through the area of ​​a person's head for a short period of time, the position matching relationship and motion consistency relationship between the safety helmet trajectory and the spatiotemporal tracking trajectory cannot continuously meet the preset association conditions, and the on-site control host will determine that the corresponding safety helmet is not worn correctly.

[0114] When a person waiting to enter passes by an adjacent person, the safety helmet worn by the adjacent person may be in the currently acquired image at the same time. However, the trajectory of the safety helmet is inconsistent with the spatiotemporal tracking trajectory of the person waiting to enter in terms of spatial location, direction of movement, or temporal continuity. The on-site control host does not bind the safety helmet detection result to the transaction identifier corresponding to the person waiting to enter.

[0115] The results of safety equipment testing should include at least the type of safety equipment, whether the safety equipment exists, whether the safety equipment is worn correctly, the trajectory markers of the safety equipment, the spatiotemporal correlation results, and the time of testing.

[0116] When the trajectory of the safety equipment and the spatiotemporal tracking trajectory of the personnel to be admitted meet the preset association conditions, the on-site control host will write the corresponding safety equipment detection result into the single entry transaction associated with the spatiotemporal tracking trajectory.

[0117] After the safety equipment detection results are bound to the transaction identifier, the field control host continues to detect the spatiotemporal correlation between the safety equipment trajectory and the spatiotemporal tracking trajectory. When the safety equipment target leaves the expected wearing area, or when the spatiotemporal correlation between the two is lower than the preset correlation threshold and continues for more than a preset time, the field control host updates the corresponding safety equipment detection results to a failure status or an incorrect wearing status.

[0118] The first and second personnel entering the access area one after the other, with the first person correctly wearing a safety helmet and the second person not wearing one, are identified. The on-site control host establishes a first spatiotemporal tracking trajectory and a second spatiotemporal tracking trajectory, and also establishes a safety helmet trajectory for the target. If the safety helmet trajectory and the first spatiotemporal tracking trajectory meet the conditions of position matching, time overlap, and consistent movement, the qualified safety helmet detection result is bound to the transaction identifier corresponding to the first person entering the area. If the safety helmet trajectory and the second spatiotemporal tracking trajectory do not meet the preset association conditions, the safety equipment detection result for the second person entering the area remains "no valid safety helmet detected."

[0119] Example 4 The construction site is pre-divided into multiple construction zones, each with a corresponding zone identifier. These zones include general construction zones, restricted construction zones, and high-risk construction zones. The on-site control host or back-end management platform stores the correspondence between each construction zone and access control mechanisms, and determines the target construction zone for personnel entering the site based on their work tasks, work teams, job types, or personnel selection results.

[0120] After identity verification is completed and a personnel identification tag is obtained, the on-site control host uses the personnel identification tag as an index to retrieve the personnel status information of the personnel to be admitted from the back-end management platform or the local personnel information database. The personnel status information includes safety training status, job qualification status, team authorization status, and prohibited entry status. It may also include the affiliated unit, personnel category, job type, validity period of qualification certificate, and historical entry anomaly information.

[0121] Safety training status indicates whether personnel entering the site have completed the required safety training for the current construction project or target construction area, and includes at least three statuses: completed, incomplete, and expired. Job qualification status indicates whether personnel entering the site possess the necessary qualifications to perform the target operation. For special operations personnel, the job qualification status also includes the type of qualification certificate, the validity period of the qualification certificate, and the scope of applicable operations. Team authorization status indicates whether the team to which the personnel entering the site belong has the authority to enter the target construction area. Prohibited entry status indicates whether personnel entering the site are marked with a project-level prohibited entry sign, a region-level prohibited entry sign, or a temporary restricted entry sign.

[0122] The on-site control host simultaneously acquires the operational risk status of the target construction area. The operational risk status is determined based on one or more of the following: construction plan information, environmental monitoring data, construction equipment status, and personnel occupancy status in the area.

[0123] The on-site control host converts various risk factors affecting the safety status of the target construction area into corresponding risk quantities, determines the operational risk value of the target construction area based on each risk quantity, and makes corrections based on the correlation between different risk factors. The operational risk value is expressed as:

[0124] in, Indicates the target construction area At any moment The operational risk value, Indicates the first The risk amount corresponding to each risk factor Indicates the first The individual risk coefficient of each risk factor. Indicates the first Risk factors and the first The coupling coefficient when multiple risk factors exist simultaneously.

[0125] All risk quantities are normalized to the same numerical range before calculation. The individual risk coefficient is preset according to the degree of impact of each risk factor on the safety of personnel entering the site, and the coupling coefficient is determined according to the superimposed effect when different risk factors occur at the same time.

[0126] For example, when hot work is carried out simultaneously in the target construction area and the concentration of combustible gas increases, the coupling coefficient between the risk of hot work and the risk of combustible gas is taken as a positive value, so that the risk value of the work is higher than the sum of the risk values ​​when the two risk quantities are applied separately.

[0127] The on-site control unit compares the operational risk value with a preset risk threshold to determine the operational risk level of the target construction area. Operational risk levels include low risk, medium risk, high risk, and no entry.

[0128] When the operational risk value is below the first risk threshold, the target construction area is in a low-risk state; when the operational risk value reaches the first risk threshold but is below the second risk threshold, the target construction area is in a medium-risk state; when the operational risk value reaches the second risk threshold but is below the third risk threshold, the target construction area is in a high-risk state; when the operational risk value reaches the third risk threshold, the target construction area is in a prohibited entry state.

[0129] The on-site control host determines the applicable entry rules for the current entry transaction from a pre-stored set of rules, based on the personnel category, target construction area, and operational risk status of the personnel to be admitted. Entry rules include at least identity verification requirements, personnel status requirements, and safety equipment requirements.

[0130] When ordinary construction workers enter low-risk construction areas, they must meet the following requirements: pass identity verification, be in valid safety training status, be in valid work team authorization status, not be in a prohibited area, and correctly wear safety helmets and reflective vests.

[0131] When special operations personnel enter low-risk construction areas, in addition to meeting the entry requirements for ordinary construction personnel, their job qualification category must match the target job type, and their job qualification certificate must be within the validity period.

[0132] When the target construction area is in a hot work situation, the entry rules will add requirements for hot work qualifications and the wearing of safety goggles. When the target construction area is in a height work situation, the entry rules will add requirements for height work qualifications and the wearing of safety belts. When the target construction area is in a confined space work situation, the entry rules will add requirements for confined space work training and corresponding protective equipment.

[0133] When the target construction area is in a high-risk state, the on-site control host will adjust the identity verification requirement from a single identity verification to a combination of two different identity verification methods, and can add conditions for secondary authorization by management personnel. When the target construction area is in a prohibited state, regardless of whether the status of other personnel waiting to enter is qualified, the personnel status verification result will be determined as failed.

[0134] The on-site control host compares the acquired personnel status information with the current entry rules item by item. When all necessary personnel statuses meet the entry rules, a personnel status verification pass result is generated; when any necessary personnel status does not meet the entry rules, a personnel status verification fail result is generated, and the failed status item and corresponding reason are recorded.

[0135] After the personnel status verification result is generated, the field control host searches for the transaction identifier associated with the current spatiotemporal tracking trajectory identifier, and writes the personnel status verification result into the corresponding single entry transaction.

[0136] The on-site control host records the generation time of identity verification results, safety equipment test results, and personnel status verification results, and sets corresponding validity periods for different types of results.

[0137] The identity verification result is used to confirm the identity of the personnel corresponding to the current spatiotemporal tracking trajectory. Its validity period can be set from the time the identity verification result is generated until the current single entry transaction is closed. The validity period of the personnel status verification result is determined based on the update frequency of personnel status information and the operational risk status of the target construction area. The safety equipment detection result reflects the current actual wearing status of the personnel to be entered, and its validity period is shorter than that of the identity verification result and the personnel status verification result.

[0138] No. The validity status of an item at the current moment can be determined based on the validity period or validity calculation result:

[0139] in, Indicates the first The result at time effectiveness, This indicates the initial confidence level when the result was generated. This indicates the time when the result was generated. This represents the validity decay coefficient of the result. This indicates the correlation between this result and the current spatiotemporal tracking trajectory.

[0140] When the spatiotemporal tracking trajectory is continuous and the result is still associated with the current person waiting to enter, the association status takes a valid value; when the spatiotemporal tracking trajectory is interrupted, the person target is replaced, or the result loses its association with the current spatiotemporal tracking trajectory, the association status takes an invalid value, making the corresponding result invalid.

[0141] Different types of results use different validity decay coefficients. The validity period for identity verification results is longer than that for safety equipment testing results. The decay coefficient for personnel status verification results is determined based on the operational risk level of the target construction area. The higher the operational risk level of the target construction area, the larger the decay coefficient for personnel status verification results, and the shorter the corresponding validity period.

[0142] When the validity of any result falls below the corresponding validity threshold, the field control host updates the result to an invalid state and re-executes the verification process corresponding to that result. When the identity verification result is invalid, the identity information of the personnel to be admitted is re-collected and verified; when the safety equipment detection result is invalid, the safety equipment is re-identified and the spatiotemporal relationship between the safety equipment and the personnel to be admitted is determined; when the personnel status verification result is invalid, the personnel status information and the operational risk status of the target construction area are re-acquired.

[0143] The field control host can also directly determine the validity of a result based on its generation time and corresponding validity period. If the time interval between the current time and the result generation time exceeds the corresponding validity period, the result becomes invalid.

[0144] After the safety equipment detection results are bound to the transaction identifier, the field control host continues to acquire the safety equipment trajectory. When the safety equipment leaves the corresponding preset wearing area, or when the spatiotemporal correlation between the safety equipment trajectory and the spatiotemporal tracking trajectory no longer meets the preset correlation conditions, the safety equipment detection results are updated to a failure state or an incorrect wearing state.

[0145] Each entry transaction can be set to the following statuses: pending verification, verification in progress, permission granted, pending passage, passage in progress, normal closure, and abnormal status.

[0146] After a single entry transaction is created, the transaction status is initialized to "Pending Verification". When identity verification, security equipment detection, or personnel status verification begins, the transaction status is updated to "Verification in Progress".

[0147] When the identity verification result is passed, the security equipment test result meets the current entry rules, the personnel status verification result is passed, all results are in a valid state, and the spatiotemporal tracking trajectory remains continuous, the transaction status is updated to allow passage.

[0148] The normal sequence of transaction state transitions is represented as follows: Status pending verification → Verification in progress → Allowed to proceed → Pending passage → Proceeding through → Normally closed.

[0149] If any verification result fails, any result becomes invalid, the spatiotemporal tracking trajectory is interrupted, or the actual passage process is abnormal, the transaction status will be updated from the current status to the abnormal status, or it will be rolled back to the verification in progress status according to the re-verification requirement.

[0150] After the transaction status is updated to allow passage, the field control host generates a one-time release instruction corresponding to the transaction identifier. The one-time release instruction includes the transaction identifier, the identifier of the target passage access organization, the number of allowed passages, the instruction generation time, and the instruction validity period. It may also include personnel identifiers and target construction area identifiers.

[0151] The transaction identifier is used to limit the one-time release instruction to only belong to the current single entry transaction; the target channel access authority identifier is used to limit the one-time release instruction to only control the corresponding channel access authority; the allowed number of passes is used to limit the number of times the one-time release instruction can authorize personnel to pass; the instruction validity period is used to limit the time range within which the one-time release instruction can be executed.

[0152] In this embodiment, the one-time release instruction allows passage only once. After the one-time release instruction is generated, the field control host updates the transaction status to the pending passage status and sends the one-time release instruction to the target channel access mechanism.

[0153] Upon receiving a one-time clearance instruction, the target channel access authority verifies the transaction identifier, the target channel access authority identifier, and the instruction's validity period. If the verification passes, the target channel access authority performs an opening action; if the target channel access authority identifier is inconsistent or the instruction has exceeded its validity period, the opening action is not performed.

[0154] A one-time clearance instruction becomes invalid under any of the following circumstances: the target access mechanism completes a valid passage according to the instruction; the current time exceeds the validity period of the instruction; the spatiotemporal tracking trajectory of the personnel waiting to enter is interrupted; the identity verification result, security equipment test result, or personnel status verification result changes from a valid state to an invalid state; the single entry transaction changes from a permitted clearance state or a pending passage state to another state; or the access mechanism requesting the execution of the instruction is inconsistent with the target access mechanism recorded in the instruction.

[0155] After a one-time release command fails, the field control host sends a cancellation signal to the target passage mechanism. If the target passage mechanism has not yet started to open, the opening action will no longer be performed; if the target passage mechanism has already started to open, it will maintain a safe open position or complete a safe reset based on the personnel's location and the current gate status.

[0156] Personnel A, who is awaiting entry, is a regular construction worker, and the target construction area is in a low-risk state. Personnel A's identity verification result is successful, their safety training status and team authorization status are both valid, there is no prohibited entry status, and their safety helmet and reflective vest are correctly worn. When all results are valid and Personnel A's spatiotemporal tracking trajectory is continuous, the on-site control host updates the corresponding single entry transaction to the permitted release status and generates a one-time release instruction allowing passage once.

[0157] Personnel B, who is to be admitted, is a welding worker, and hot work is currently being carried out in the target construction area. Personnel B has completed safety training and is correctly wearing a safety helmet and reflective vest, but their hot work qualification certificate has expired. The on-site control host determines that Personnel B's personnel status verification result is unsuccessful and keeps the corresponding single entry transaction in the "under verification" state or updates it to an abnormal state, without generating a one-time release instruction.

[0158] Personnel C, upon completion of identity verification, personnel status verification, and safety equipment inspection, removes their safety helmet. Once the safety helmet's trajectory leaves the area where Personnel C's head is worn, the on-site control host updates the safety equipment inspection result to an invalid state, cancels the generated but not yet executed one-time release command, and reverts the corresponding single entry transaction from the pending passage state to the verification in progress state.

[0159] Example 5 The personnel entry channel is set up with a first detection area, a channel passage mechanism and a second detection area in sequence along the predetermined passage direction. The first detection area is located on the entrance side of the channel passage mechanism and is used to detect the time, number of people and direction of movement of people entering the channel. The second detection area is located on the exit side of the channel passage mechanism and is used to detect the time, number of people and direction of movement of people leaving the channel.

[0160] The first detection area and the second detection area form an internal channel area, which includes the operating range of the channel passage mechanism and the space where personnel are located when passing through the channel passage mechanism.

[0161] The first and second detection areas are respectively equipped with a depth image acquisition component and a multi-column infrared detection component. The depth image acquisition component is located above the personnel entry channel and is used to acquire depth images of the first detection area, the area inside the channel, and the second detection area, and to identify different personnel based on the human head and shoulder contour, human height, and spatial intervals between human targets.

[0162] The multi-row infrared detection assembly includes multiple sets of through-beam infrared detection units spaced apart along the personnel passage direction, with each set of through-beam infrared detection units laterally passing through the personnel entry channel. When personnel pass through the corresponding detection area, they sequentially block different positions of the through-beam infrared detection units along the passage direction. The on-site control host determines the personnel's movement direction based on the order in which the through-beam infrared detection units are blocked and then restored.

[0163] An action detection component is installed on the passageway mechanism. The action detection component is used to detect the current position and action status of the passageway mechanism. The action status includes open state, closed state, obstructed state and reset state.

[0164] When a swing gate is used as the passageway mechanism, the motion detection component includes an angle detector installed at the swing gate's pivot. The field control host determines whether the gate has started to open, reached the preset opening position, started to close, and returned to the preset closing position based on the gate angle output by the angle detector.

[0165] After the on-site control host generates a one-time release command, it updates the corresponding single entry transaction to the pending status and sends the one-time release command to the target channel access authority. The target channel access authority verifies the transaction identifier, target channel access authority identifier, and command validity period in the one-time release command. If the verification is successful, it executes the opening action.

[0166] When the target passage mechanism begins its opening action, it sends an opening start message to the field control host; when the target passage mechanism reaches the preset opening position, it sends an opening completion message to the field control host. Upon receiving the opening completion message, the field control host updates the corresponding single entry transaction to the "passing" status.

[0167] Before the passageway mechanism is activated, the field control host confirms, based on the spatiotemporal tracking trajectory, that the person waiting to enter, bound to the current transaction identifier, is located at the entrance side of the first detection area. When the person enters the first detection area, the field control host generates an entry event based on the passageway detection data generated by the first detection area.

[0168] An entry event includes a transaction identifier, entry time, entry direction, human target identifier, and number of people entering. The field control host associates the human target in the entry event with the spatiotemporal tracking trajectory corresponding to the transaction identifier. When the human target in the entry event continuously corresponds to the spatiotemporal tracking trajectory, the human target is identified as the authorized person for the current single entry transaction.

[0169] When personnel move from the first detection area to the inner area of ​​the passage, the through-beam infrared detection unit near the entrance is blocked first, and the through-beam infrared detection unit near the passage passage mechanism is blocked later. Based on this, the on-site control host determines that the personnel are moving in the predetermined entry direction.

[0170] When the blocking sequence of each through-beam infrared detection unit is opposite to the blocking sequence corresponding to the predetermined entry direction, the on-site control host determines that the personnel are moving in the opposite direction.

[0171] The depth image acquisition component counts human targets entering the first detection area. For two people standing adjacent to each other, the on-site control host identifies them as two passing human targets based on the boundary between the head and shoulder contours of the two people, the spatial distance between the human targets, and their respective movement trajectories.

[0172] When one human target is identified in the first detection area, the number of people entering is determined to be one; when two or more independent human targets are identified, the number of people entering is determined according to the number of human targets identified.

[0173] When personnel pass through the access control mechanism and enter the second detection area, the on-site control host generates a departure event based on the access detection data generated by the second detection area. The departure event includes a transaction identifier, departure time, departure direction, human target identifier, and number of people leaving.

[0174] The on-site control host associates human targets entering the first detection area with human targets leaving the second detection area based on human characteristics, changes in spatial position, direction of movement, and temporal continuity during entry and exit events, in order to determine whether the personnel have completed passage.

[0175] The on-site control host also determines the current number of people in the channel's internal area based on the changes in the number of people between the first detection area, the internal area of ​​the channel, and the second detection area. The current number of people in the internal area of ​​the channel is expressed as follows:

[0176] in, This indicates the number of people inside the passage at the current moment. This indicates the number of people in the area inside the channel at the previous detection time. This indicates the number of people who entered the internal area of ​​the passage from the first testing area during the current testing period. This indicates the number of people who entered the second testing area from the area inside the channel during the current testing period.

[0177] When a single passage instruction allows only one passage, the number of people entering and leaving under normal passage conditions is one, the direction of passage is the predetermined entry direction, and the current number of people in the internal area of ​​the passage returns to zero after the people pass through.

[0178] The on-site control host synchronously verifies the operation process of the passageway mechanism. The normal operation process of the passageway mechanism includes receiving a one-time release command, starting to open, opening to the final position, maintaining the open position, starting to close, and closing to the final position.

[0179] When the passageway mechanism reaches the preset open position within the preset opening time limit and returns to the preset closed position within the preset reset time limit after personnel leave the second detection area, the on-site control host determines that the action feedback of the passageway mechanism is normal.

[0180] If the passageway mechanism fails to reach the preset opening position within the preset opening time limit, closes prematurely before personnel have passed, fails to close within the preset reset time limit after personnel have passed, or the gate operation is obstructed, the on-site control host will determine that the passageway mechanism is malfunctioning.

[0181] The on-site control host determines the actual passage result based on the number of people entering, leaving, passage direction, the continuous correspondence between human targets and spatiotemporal tracking trajectories, and the action feedback of the passage mechanism.

[0182] When the number of people entering and leaving both match the number of passes allowed in the one-time release instruction, the spatiotemporal tracking trajectories of the human targets and transaction identifiers in the entry and exit events continuously correspond, and the personnel pass through the first detection area, the internal area of ​​the passage, and the second detection area in sequence along the predetermined entry direction, the passage passage mechanism completes the opening and closing actions, and the current number of people in the internal area of ​​the passage is zero after the personnel pass through, the on-site control host determines the actual passage result as normal passage.

[0183] When the actual passage result is normal passage, the on-site control host invalidates the one-time release command, updates the single entry transaction to the normal closed state, and records the entry time, exit time, passage direction, number of people entering, number of people leaving, and the action status of the passage mechanism.

[0184] When the number of people entering or leaving is inconsistent with the allowed number of passages, the direction of personnel passage is abnormal, the human target is not continuous with the spatiotemporal tracking trajectory, personnel linger in the internal area of ​​the passage, or the passage passage mechanism operates abnormally, the on-site control host will determine the actual passage result as abnormal passage.

[0185] The current one-time clearance instruction allows one passage. However, if the first detection area detects two or more human targets entering the channel, or the second detection area detects more people leaving than the allowed number of passages, the on-site control host will determine the actual passage result as tailgating.

[0186] For example, Person A has completed identity verification, security equipment inspection, and personnel status verification, and has obtained a one-time clearance instruction. Person B, without obtaining clearance authorization, follows Person A into the passage. The first detection area identifies two human targets entering the passage area, but the current one-time clearance instruction only allows one person to pass. The on-site control host marks this passage as an abnormal tailgating.

[0187] When tailgating is detected but the person following has not yet entered the gate's operating range, the on-site control host will stop widening the gate's opening and output a tailgating alarm. Once the person following has entered the gate's operating range, the on-site control host will maintain the passage mechanism in a safe position to prevent injury and send a response prompt to the on-site management terminal.

[0188] If the second detection area detects a human target entering first, and then the first detection area detects the human target leaving, or if the blocking sequence of the through-beam infrared detection unit is opposite to the predetermined entry direction, the on-site control host will determine the actual passage result as reverse passage.

[0189] When reverse passage is detected, the field control host does not use the one-time release command that has not yet been executed, keeps the passage mechanism of the control channel closed, and outputs a reverse passage prompt.

[0190] If the first detection area has detected personnel entering, and the current number of people in the passage area has increased, but the second detection area has not detected the corresponding personnel leaving within the preset passage time limit, and the current number of people in the passage area has not returned to zero, the on-site control host determines that the personnel are in a state of being stuck in the passage.

[0191] When personnel remain within the operating range of the passage mechanism, the on-site control host suspends the gate closing action and provides voice prompts to personnel to continue passage or exit the passage. If passage is not completed after the preset dwell time limit, the on-site control host determines the actual passage result as passage congestion and updates the single entry transaction to an abnormal status.

[0192] If the spatiotemporal tracking trajectory of the human target entering the first detection area is discontinuous with the transaction identifier, or if the human target leaving the second detection area is inconsistent with the human target entering the first detection area, the on-site control host will determine the actual passage result as a personnel mismatch.

[0193] For example, if Person A, who has obtained a one-time clearance instruction, leaves the channel identification area after the channel access mechanism is opened, and Person B, who has not obtained clearance authorization, enters the channel, even if the first detection area detects one person entering and the second detection area detects one person leaving, the on-site control host will still determine this passage as a personnel mismatch because the spatiotemporal tracking trajectory of the human target entering the channel is discontinuous with that of Person A.

[0194] If the passage mechanism fails to reach the preset opening position after receiving a one-time release command, the gate position feedback is inconsistent with the control command, the gate is continuously obstructed during the closing process, or the gate fails to return to the preset closing position, the on-site control host will determine the actual passage result as an abnormal mechanism action.

[0195] When an abnormal mechanism action occurs, the field control host writes the abnormal status into the corresponding single entry transaction, controls the passage mechanism to stop its current action or return to a safe position, and sends the equipment abnormality information to the back-end management platform.

[0196] When one person enters, zero people leave, and the person exits the first detection area in the opposite direction, the on-site control host determines that the passage is not completed based on the reverse spatiotemporal tracking trajectory of the person, thus invalidating the one-time release command and updating the single entry transaction to an active exit status or an abnormal status.

[0197] When zero people enter but the second detection area detects people leaving, the on-site control host combines the current number of people in the channel area with historical detection events to determine whether there are people who have been delayed or whether the data in the first detection area is missing, and marks this passage as a channel detection anomaly.

[0198] When the number of people obtained by the depth image acquisition component and the multi-column infrared detection component is inconsistent, the on-site control host verifies the data by combining data from multiple consecutive detection times and changes in the number of people in the passage area. If subsequent data shows that the number of people entering, leaving, and the current number of people in the passage area are consistent, the verified number of people is adopted; if the consistency cannot be met, the actual passage result is determined to be an anomaly.

[0199] When a person carrying a large toolbox passes through the first detection area, the toolbox may cause multiple occlusion areas to be generated by multiple rows of infrared detection components. Based on the human body outline in the depth image and the state of the toolbox moving synchronously with the human target, the on-site control host determines that the situation is a person carrying an item and does not increase the number of people entering.

[0200] After a single entry transaction is updated to an abnormal state, the on-site control host cancels the one-time release instruction that has not yet expired, and writes the abnormal type, the time of the abnormality, the detection data of the first detection area, the detection data of the second detection area, the number of personnel in the internal area of ​​the passage, and the action feedback of the passage passage mechanism into the corresponding transaction record.

[0201] Example 6 A safety interlocking system for construction workers based on single entry transactions and multi-source spatiotemporal binding includes a multi-source sensing module, a trajectory and transaction generation module, an identity verification module, a safety equipment association module, a personnel status verification module, a transaction status control module, an access control module, and an access result processing module.

[0202] Each module is located in the field control host, which is installed in a cabinet on one side of the construction site access passage and is connected to the image acquisition equipment, spatial detection equipment, identity information acquisition equipment, passage access mechanism and back-end management platform.

[0203] The side of the cabinet facing the personnel entering the premises is equipped with facial image acquisition equipment, upper body image acquisition equipment, display equipment, document reading equipment, and identity auxiliary acquisition equipment. A top image acquisition device is installed above the channel recognition area, and spatial detection equipment is installed on both sides or above the channel. The entrance and exit sides of the channel passage mechanism form a first detection area and a second detection area, respectively. Depth image acquisition components and multi-column infrared detection components are installed in each of the two detection areas. Motion detection components are installed on the channel passage mechanism to detect its open, closed, obstructed, and reset states.

[0204] The multi-source sensing module is connected to the face image acquisition device, upper body image acquisition device, top image acquisition device and spatial detection device respectively, and is used to acquire personnel image data, security equipment image data and channel spatial detection data in the channel recognition area, and to process the various types of data synchronously according to a unified time reference.

[0205] The trajectory and event generation module identifies human targets based on synchronized multi-source sensing data, and performs correlation tracking based on the spatial position, movement speed, movement direction and human characteristics of the human targets at continuous time intervals to form the spatiotemporal tracking trajectory of the personnel to be entered, and generates trajectory identifiers for each valid spatiotemporal tracking trajectory.

[0206] When the time-space tracking trajectory first enters the transaction triggering area and moves towards the passageway mechanism, the trajectory and transaction generation module creates a single entry transaction, generates a transaction identifier, and uniquely associates the transaction identifier with the trajectory identifier. If the current trajectory identifier is already associated with a single entry transaction that has not yet been closed, the original transaction identifier will continue to be used, and no new single entry transaction will be created.

[0207] When a human target is briefly occluded, the trajectory and event generation module predicts its movement position based on the human's position and movement state before occlusion, and combines top image data and channel space detection data to maintain trajectory continuity; if the human target does not reappear after the preset recovery time limit, the corresponding spatiotemporal tracking trajectory is determined to be interrupted.

[0208] The identity verification module connects to the facial image acquisition device, the document reading device, and the identity auxiliary acquisition device to collect the identity information of people waiting to enter. The identity verification module determines the target verification method based on the facial image quality and the degree of facial occlusion. When the verification credibility of the target verification method is lower than a preset threshold, another identity verification method is called for verification.

[0209] When identity verification is successful, the identity verification module obtains the corresponding personnel identifier and writes the identity verification result and personnel identifier into the single entry transaction associated with the current trajectory identifier; when identity verification fails, the reason for the verification failure is written into the corresponding single entry transaction and a prompt is output through the display device or voice device.

[0210] The safety equipment association module receives safety equipment image data and spatiotemporal tracking trajectories, which are used to identify and continuously track safety helmets, reflective vests, goggles, safety belts, or protective masks to form corresponding safety equipment trajectories.

[0211] The safety equipment association module determines the spatiotemporal association result based on the position matching relationship, time overlap relationship, and motion consistency relationship between the safety equipment trajectory and the spatiotemporal tracking trajectory, and generates the safety equipment detection result based on whether the safety equipment is located in the preset wearing area.

[0212] When the trajectory of the safety equipment meets the preset association conditions, the safety equipment association module writes the safety equipment detection result into the corresponding single entry transaction; when the safety equipment leaves the preset wearing area, or no longer meets the preset association conditions with the spatiotemporal tracking trajectory, the safety equipment detection result is updated to a failure state or an incorrect wearing state.

[0213] Based on the personnel identifier output by the identity verification module, the personnel status verification module retrieves the safety training status, job qualification status, team authorization status, and restricted entry status from the local personnel information database or the back-end management platform, and determines the target construction area corresponding to the personnel to be admitted.

[0214] The personnel status verification module determines the entry rules based on the personnel category and the operational risk status of the target construction area, compares the personnel status information with the entry rules, generates the personnel status verification result, and writes the personnel status verification result into the corresponding single entry transaction.

[0215] The transaction status control module is connected to the trajectory and transaction generation module, the identity verification module, the security equipment association module, and the personnel status verification module, respectively, and is used to determine the transaction status of a single entry transaction.

[0216] A single entry transaction includes the following states: pending verification, verification in progress, permission granted, waiting to pass, passing in progress, normal closure, and abnormal status. The transaction status control module records the generation time and validity period of the identity verification result, security equipment detection result, and personnel status verification result, and determines the validity of each result by combining the continuous status of the spatiotemporal tracking trajectory.

[0217] When the identity verification result is passed, the security equipment test result and the personnel status verification result both meet the entry rules, all results are in a valid state and the spatiotemporal tracking trajectory is continuous, the transaction status control module will update the transaction status to the allowed release state and generate a one-time release instruction.

[0218] The one-time release instruction includes a transaction identifier, the target channel access authority identifier, the allowed number of passes, and the instruction's validity period. After the one-time release instruction is generated, the transaction status control module updates the transaction status to pending passage and sends the one-time release instruction to the access control module.

[0219] If any test result exceeds its validity period, the status of the security equipment changes, or the spatiotemporal tracking trajectory is interrupted, the transaction status control module invalidates the corresponding result, cancels the one-time release instruction that has not yet been executed, and rolls the transaction status back to the verification status or updates it to the abnormal status.

[0220] The access control module connects to the access control mechanism and its motion detection components to verify the transaction identifier, target access control mechanism identifier, and instruction validity period in a one-time release instruction. If the verification passes, the target access control mechanism is opened; if the verification fails, the opening action is not performed.

[0221] When the passage mechanism starts to open, opens to the correct position, begins to close, closes to the correct position, or is obstructed, the motion detection component sends the corresponding motion feedback to the passage control module and the passage result processing module.

[0222] The passage result processing module determines the number of people entering, leaving, and the direction of passage based on the channel detection data generated by the first and second detection areas, and judges whether the human targets entering the channel are continuous with the spatiotemporal tracking trajectory corresponding to the transaction identifier.

[0223] When the number of people entering and leaving both match the number of passages allowed in the one-time release instruction, the personnel pass through in the predetermined direction, the human target and the spatiotemporal tracking trajectory are continuous, and the passage mechanism completes the opening and resetting actions, the passage result processing module determines the actual passage result as normal passage.

[0224] When the actual passage result is normal passage, the transaction status control module invalidates the one-time release command and updates the single entry transaction to the normal closed state.

[0225] When the number of people entering or leaving is inconsistent with the allowed number of passages, people pass in the wrong direction, the spatiotemporal tracking trajectory of the human target and the authorized personnel is discontinuous, people are stuck in the passage, or the passage mechanism operates abnormally, the passage result processing module will determine the actual passage result as abnormal passage.

[0226] When the actual passage result is abnormal passage, the transaction status control module cancels the one-time release instruction that has not yet expired and updates the single entry transaction to an abnormal status; the passage control module controls the passage mechanism to stop its operation, maintain a safe position, or perform a reset based on the personnel's location and the current status of the passage mechanism.

[0227] The on-site control host is also equipped with a data storage module and a communication module. The data storage module is used to store personnel registration information, personnel status information, entry rules, single entry transaction records, and channel access mechanism operation records; the communication module is used to upload closed or abnormal single entry transactions to the back-end management platform, and to receive updated personnel information and entry rules from the back-end management platform.

[0228] When communication between the on-site control host and the back-end management platform is interrupted, the identity verification module and the personnel status verification module call the personnel information and entry rules cached in the data storage module to continue the verification; after communication is restored, the communication module uploads the transaction records formed during the communication interruption to the back-end management platform.

[0229] The above description of the top and bottom of the form is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for safety interlocking of construction workers based on single entry transactions and multi-source spatiotemporal binding, characterized in that, Includes the following steps: S1. Acquire multi-source sensing data within the channel identification area, construct the spatiotemporal tracking trajectory of the person to be admitted, and generate a single admission transaction and transaction identifier accordingly. S2. Collect and verify the identity information of the personnel to be admitted, obtain the identity verification result and personnel identifier, and bind them with the transaction identifier; S3. Identify security equipment based on the multi-source sensing data and obtain security equipment detection results, determine the spatiotemporal correlation between the security equipment and the personnel waiting to enter, and bind the security equipment detection results with the transaction identifier; S4. Obtain and verify personnel status information based on the personnel identifier, obtain personnel status verification result, and bind it with the transaction identifier; S5. Determine the transaction status based on the identity verification result, security equipment detection result, personnel status verification result and the continuous status of the spatiotemporal tracking trajectory bound to the transaction identifier, and generate a one-time release instruction when permission is granted. S6. Control the passage mechanism to operate according to the one-time release command, and determine the actual passage result based on the passage detection data and action feedback; S7. When passage is normal, invalidate the one-time release command and close the single entry transaction; when passage is abnormal or the spatiotemporal tracking trajectory is interrupted, cancel the one-time release command and update the single entry transaction to an abnormal state.

2. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 1, characterized in that, In step S1, the multi-source sensing data includes at least personnel image data and channel spatial detection data; The personnel image data and the channel space detection data are synchronized in time. The human targets of the personnel to enter are identified from the personnel image data, and the spatial position and motion state of the human targets at continuous time are correlated and tracked to obtain the spatiotemporal tracking trajectory with trajectory identification.

3. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 2, characterized in that, In step S1, when the spatiotemporal tracking trajectory first enters the preset transaction triggering area, the single entry transaction is created and the transaction identifier is generated. The transaction identifier is uniquely associated with the trajectory identifier, and no new single entry transaction is created based on the same spatiotemporal tracking trajectory before the single entry transaction is closed.

4. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 1, characterized in that, In step S2, the environmental quality parameters of the channel recognition area and the facial occlusion status of the person to be admitted are obtained. Based on the environmental quality parameters and the facial occlusion status, a target verification method is determined from a variety of preset identity verification methods. The target verification method is used to verify the identity. If the verification credibility is lower than a preset threshold, a different identity verification method than the target verification method is called for review. The identity verification result and the person identifier are determined based on the verification results of each identity verification method.

5. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 1, characterized in that, In step S3, the identified safety equipment is continuously tracked based on the multi-source sensing data to obtain the trajectory of the safety equipment. The spatiotemporal association result is determined based on the position matching relationship, time overlap relationship and motion consistency relationship between the safety equipment trajectory and the spatiotemporal tracking trajectory. When the spatiotemporal association result meets the preset association conditions, the corresponding safety equipment detection result is bound to the transaction identifier.

6. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 1, characterized in that, In step S4, the personnel status information includes at least one of safety training status, job qualification status, team authorization status, and restricted entry status; Obtain the target construction area and its operational risk status corresponding to the personnel to be admitted. Determine the preset entry rules based on the personnel category and operational risk status of the personnel to be admitted, and verify the personnel status information according to the preset entry rules.

7. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 1, characterized in that, In step S5, the generation time and validity period of the identity verification result, security equipment detection result and personnel status verification result are recorded respectively. The validity status of each result is determined according to whether each result is within the corresponding validity period and whether the spatiotemporal tracking trajectory is continuous. When any result exceeds its corresponding validity period, the spatiotemporal tracking trajectory is interrupted, or the security equipment status changes, the corresponding result becomes invalid, and the transaction status is updated according to the validity status of each result.

8. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 7, characterized in that, In step S5, the one-time release instruction includes the transaction identifier, the target channel access authority identifier, the number of permitted passes, and the instruction validity period; The one-time release instruction is only used to control the corresponding target channel passage mechanism to complete one passage, and it becomes invalid when the passage is completed, the validity period of the instruction is exceeded, the spatiotemporal tracking trajectory is interrupted, or the transaction status changes.

9. The construction worker safety interlocking method based on single entry transaction and multi-source spatiotemporal binding according to claim 5, characterized in that, In steps S6 and S7, the number of people entering, the number of people leaving, and the direction of passage are determined based on the channel detection data generated by the first and second detection areas set along the passage direction. The actual passage result is determined in combination with the action feedback of the passage mechanism. When the number of people entering and leaving is consistent with the allowed number of passages and the direction of passage is normal, it is determined as normal passage; otherwise, it is determined as abnormal passage.

10. A safety interlocking system for construction workers based on single entry transactions and multi-source spatiotemporal binding, characterized in that, include: The multi-source sensing module is used to acquire multi-source sensing data within the channel identification area; The trajectory and transaction generation module is used to construct the spatiotemporal tracking trajectory of the person to be admitted based on the multi-source sensing data, and generate a single admission transaction and corresponding transaction identifier based on the spatiotemporal tracking trajectory. The identity verification module is used to collect and verify the identity information of people waiting to enter, obtain the identity verification result and the personnel identifier, and bind the identity verification result and the personnel identifier with the transaction identifier. The security equipment association module is used to identify security equipment based on the multi-source sensing data, determine the spatiotemporal association between the security equipment and the personnel to be admitted, and bind the corresponding security equipment detection results with the transaction identifier. The personnel status verification module is used to obtain and verify the personnel status information of the personnel to be admitted based on the personnel identifier, obtain the personnel status verification result, and bind the personnel status verification result with the transaction identifier; The transaction status control module is used to determine the transaction status based on the identity verification result bound to the transaction identifier, the security equipment detection result, the personnel status verification result, and the continuous status of the spatiotemporal tracking trajectory, and to generate a one-time release instruction when the transaction status is an allowed release status. The passage control module is used to control the operation of the passage mechanism according to the one-time release command; The passage result processing module is used to determine the actual passage result based on the channel detection data and the action feedback of the channel passage mechanism, and to close the single entry transaction or update the single entry transaction to an abnormal state based on the actual passage result.

Images