Road constructor management method and system based on track positioning

Abstract

The invention discloses a road constructor management method and system based on track positioning, and relates to the technical field of data processing. According to the method, distance measurement data of a constructor is collected through a positioning device, image data of the constructor is collected through a camera device, a local monitoring device constructs an overdetermined equation set based on the distance measurement data of a first period and solves the overdetermined equation set to generate a first coordinate, and the local monitoring device generates a second coordinate based on the image data of a second period. And the trajectory prediction model generates a trajectory path of the constructor according to the first coordinate and the second coordinate, and outputs an early warning signal based on the trajectory path. And furthermore, the first period is reversely adjusted according to the image data, the second period is reversely adjusted according to the distance measurement data, the third period is adjusted according to the error between the first coordinate and the second coordinate, and the signal processing model and the image processing model are updated based on the third period, so that the supervision precision and the response efficiency in a complex construction environment can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a highway construction personnel management method and system based on trajectory positioning. Background Art

[0002] Safety protection work at construction sites is receiving more and more attention. A Chinese patent application with publication number CN118506265A discloses a machine vision-based method for identifying and alarming the trajectory of construction workers' edge operations. The method obtains site image information, uses AI virtual electronic fence technology to divide the construction area, detects and locates construction workers in the image information, and uses the ReInspect algorithm to build a multi-target tracking model for construction workers' edge operations in the image information, and tracks and collects the movement trajectory of construction workers. A construction worker edge operation trajectory prediction module based on Social-STGCNN is established to predict the movement trajectory of construction workers, analyze the trajectory information of construction workers that may be in danger, and thus realize the identification and alarm of edge operations. This method requires the terminal device to continuously upload site image information, requires the deployment of limited broadband to increase the network rate, and requires the deployment of hardware facilities with higher processing capabilities to train deep learning models. This method is not suitable for highway construction environments where the construction area is constantly changing. Therefore, it is necessary to propose a construction worker supervision method with low requirements for network rate and hardware facilities to meet the needs of highway construction. Summary of the invention

[0003] In view of the above problems, the present invention provides a highway construction personnel management method and system based on trajectory positioning, which realizes the trajectory positioning of construction personnel by combining ranging data and image data. By optimizing the signal period of ranging data and image data and the model update period, the data transmission volume and processing difficulty are reduced while ensuring the accuracy of trajectory warning, and the network rate and hardware requirements for terminal equipment are reduced.

[0004] The invention objectives of this application can be achieved through the following technical means: A highway construction personnel management method based on trajectory positioning comprises the following steps: Step 1: Generate an electronic fence for the construction road, set a buffer area and a pass area on the construction road based on the electronic fence, set a monitoring device in the pass area, and set a throttling device in the buffer area; Step 2: The monitoring device predicts the traffic flow in the buffer area, the throttling device adjusts the traffic flow in the buffer area, the positioning device periodically collects the distance measurement data of the construction personnel, and the camera device periodically collects the image data of the construction personnel; Step 3: The positioning device reports the distance measurement data at a first period interval, the camera device reports the image data at a second period interval, and the local monitoring device predicts the first coordinates of the construction personnel based on a signal processing model and predicts the second coordinates of the construction personnel based on an image processing model; Step 4: The early warning device adjusts the trajectory prediction model based on the first cycle and the second cycle, inputs multiple sets of first coordinates and second coordinates into the trajectory prediction model to obtain the trajectory path of the construction personnel, and outputs an early warning signal if at least one trajectory coordinate point of the trajectory path is outside the electronic fence; Step 5: extracting the extreme distance between the trajectory path and the electronic fence, and updating the signal period of the throttling device according to the traffic flow and multiple sets of extreme distances; Step 6: adjusting the first period according to the image data of different camera devices at the same time, adjusting the second period according to the ranging data of different positioning devices at the same time, and adjusting the third period according to the first coordinates and the second coordinates at the same time; Step 7: The local monitoring device reports the distance measurement data and image data every third period, and the remote monitoring device updates the signal processing model and the image processing model, and returns to step 2.

[0005] In the present invention, in step 3, I groups of positioning devices that have collected the distance measurement data of the construction personnel are extracted, the distance measurement data are substituted into the signal processing model to calculate the signal propagation time of each positioning device, a reference positioning device is extracted from the I groups of positioning devices, and the reference coordinates (x0, y0, z0) and signal propagation time t0 of the reference positioning device are created. i ,y i ,z i ) and the signal propagation time t i The overdetermined set of equations is used to calculate the first coordinate of the construction worker, i=1,2,...,I-1.

[0006] In the present invention, in step 3, the image processing model includes a monocular sub-model and a multi-eye sub-model, J groups of camera devices that collect image data of the construction worker are extracted, the image contour of the construction worker in the image data is extracted, the monocular sub-model predicts the observation coordinates of the image contour according to the intrinsic parameter matrix of the camera device, and the multi-eye sub-model predicts the viewing angle weight of the image contour according to the extrinsic parameter matrix of the camera device, and then the second coordinate of the construction worker is calculated according to multiple observation coordinates and corresponding viewing angle weights.

[0007] In the present invention, in step 4, the noise covariance matrix of the ranging data is adjusted according to the first cycle, and the noise covariance matrix of the image data is adjusted according to the second cycle, and the weight w1 of the ranging data and the weight w2 of the image data in the trajectory prediction model are respectively calculated based on the adjusted noise covariance matrix.

[0008] In the present invention, in step 4, multiple sets of first coordinates and second coordinates are input into the trajectory prediction model. If the current moment is an integer multiple of the second period, the trajectory coordinate point is generated according to the first coordinate and the second coordinate. Otherwise, the first coordinate is directly used as the trajectory coordinate point.

[0009] In the present invention, in step 6, J groups of image data of different camera devices at the same time are obtained, and then J groups of observation coordinates are calculated according to the image data, the difference value Δ1 between the J groups of observation coordinates is calculated, and the first period is adjusted according to the difference value Δ1.

[0010] In the present invention, in step 6, the ranging data of I groups of positioning devices at the same time are obtained, three groups of positioning devices in the ranging data are selected to generate a positioning device combination, candidate coordinates are calculated according to the three-sided positioning method and the positioning device combination, multiple groups of candidate coordinates are generated, the difference value Δ2 between the multiple groups of candidate coordinates is calculated, and the second period is adjusted according to the difference value Δ2.

[0011] In the present invention, in step 6, the first coordinate and the second coordinate at the same time are obtained, the difference value Δ3 between the first coordinate and the second coordinate is calculated, and the third period is adjusted according to the difference value Δ3.

[0012] In the present invention, in step 7, the remote monitoring device calculates the theoretical signal propagation time between any non-reference positioning device and the reference positioning device based on the ranging data, calculates the time difference between the actual signal propagation time and the theoretical signal propagation time, takes the mean of the time difference as the time difference error coefficient δ, reprojects the image data, calculates the viewing angle deviation between the reprojected pixel coordinates and the pixel coordinates of the image contour of each camera device, takes the mean of the viewing angle deviation as the viewing angle error coefficient γ, and the local detection device updates the signal processing model and the image processing model according to δ and γ, respectively.

[0013] A system for implementing the highway construction personnel management method based on trajectory positioning, comprising: a monitoring device, the monitoring device being configured to predict traffic flow in the buffer area; a throttling device, the throttling device being configured to adjust the flow of traffic in the buffer area; A positioning device, the positioning device is configured to collect distance measurement data of construction personnel; A camera device, the camera device is configured to collect image data of construction personnel; A local monitoring device, the local monitoring device is configured to predict a first coordinate and a second coordinate of a construction worker; An early warning device, the early warning device is configured to predict the trajectory path of the construction personnel and output an early warning signal; A remote monitoring device, the remote monitoring device is configured to update the signal processing model and the image processing model, wherein: The positioning device reports the distance measurement data at intervals of a first cycle, the camera device reports the image data at intervals of a second cycle, and the local monitoring device reports the distance measurement data and the image data at intervals of a third cycle. The local monitoring device adjusts the first period according to the image data of different camera devices at the same time, adjusts the second period according to the ranging data of different positioning devices at the same time, and adjusts the third period according to the first coordinates and the second coordinates at the same time.

[0014] The beneficial effects of implementing the highway construction personnel management method and system based on trajectory positioning of the present invention are as follows: the present invention collects the distance measurement data of the construction personnel through the positioning device, collects the image data through the camera device, and realizes the trajectory positioning of the construction personnel by combining the distance measurement data and the image data, thereby ensuring the accuracy of the trajectory positioning. The first period is adjusted when the image data captured by different cameras are quite different, and the second period is adjusted when the distance measurement data collected by different positioning devices are quite different, thereby reducing the data transmission volume and processing difficulty on the basis of ensuring the accuracy of trajectory positioning. Furthermore, the present invention adjusts the third period according to the error between the first coordinate and the second coordinate, and when the working environment changes and affects the data collection, the signal processing model and the image processing model are updated in time to improve the supervision efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the highway construction personnel management method based on trajectory positioning of the present invention; Figure 2 A schematic diagram of a highway construction method according to the present invention; Figure 3 A schematic diagram of the present invention showing the collection of distance measurement data by multiple positioning devices; Figure 4 A schematic diagram of an image contour in image data of the present invention; Figure 5 A schematic diagram of observation coordinates for predicting image contours according to the present invention; Figure 6 It is a schematic diagram of the first cycle, the second cycle and the trajectory generation cycle of the present invention; Figure 7 A schematic diagram of the construction worker trajectory path of the present invention; Figure 8 A block diagram of a system for implementing the highway construction personnel management method based on trajectory positioning according to the present invention; Fig. 9 A block diagram of a wireless transceiver device for construction personnel of the present invention; Fig.10 The figure is a structural diagram of the intelligent helmet with a wireless transceiver of the present invention.

[0016] Explanation of some reference numerals: cap body 10 , cap brim 20 , waterproof box body 30 , warning module 31 , charging interface 32 . DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0018] The highway construction environment has significant complexity and dynamic characteristics, which are specifically manifested in high-risk working conditions such as high-frequency movement of construction machinery, real-time changes in the boundaries of the working area, and intensive human-machine interaction operations. Traditional personnel management methods mainly rely on the following two technical paths: one is a static monitoring system based on fixed cameras, and the other is a positioning device based on radio frequency signals. However, the above single technical solutions all have inherent limitations: although the static monitoring system can achieve visual coverage, the image data processing is complex, and there are problems such as low monitoring efficiency and delayed response. The signal positioning device is easily blocked by the metal structure of large construction equipment, resulting in deterioration of positioning accuracy. In response to this technical bottleneck, the present invention innovatively proposes a solution for multi-source data fusion, combining the advantages of ranging data and image data to complementarily locate construction personnel. This dual-modal data collaboration mechanism not only overcomes the technical shortcomings of a single signal source, but also improves the robustness of the system through data complementarity. Embodiment 1

[0019] Reference Figure 1 The highway construction personnel management method based on trajectory positioning of the present invention described in detail in this embodiment includes the following steps: Step 1: Generate an electronic fence for the construction road, set up a buffer area and a pass area on the construction road based on the electronic fence, set up a monitoring device in the pass area, and set up a throttling device in the buffer area. Figure 2 The electronic fence is a virtual boundary generated according to the geographic information of the construction road. It is used to define the construction area of ​​the construction personnel. The construction area is the current working range. The buffer area is a triangular area, one side of which is connected to the electronic fence and the vertex points to the direction of traffic. The area of ​​the buffer area is determined by the construction specifications. The monitoring device is usually an infrared counter or a camera, and the throttling device is, for example, a traffic light.

[0020] Step 2: The monitoring device predicts the traffic flow in the buffer area, the throttling device adjusts the traffic flow in the buffer area, the positioning device periodically collects the distance measurement data of the construction workers, and the camera device periodically collects the image data of the construction workers. The monitoring device predicts the traffic flow in the buffer area in the next minute or one traffic light cycle based on the vehicle information in the traffic area. Figure 3The positioning device is, for example, a UWB base station arranged around the electronic fence. The construction workers carry smart helmets, and the wireless transceiver in the smart helmet can receive the ranging signal sent by the UWB base station. Multiple UWB base stations use different codes to send ranging signals to the construction workers at the same time. The wireless transceiver responds to the signal quickly, and the UWB base station records the signal propagation time to generate ranging data. Since construction equipment will block the propagation of part of the ranging signal, it is necessary to combine image data for auxiliary positioning. The camera device is, for example, multiple industrial cameras installed on a 3-6 meter high bracket. Multiple industrial cameras synchronously capture the scene on the construction road to generate image data.

[0021] Step 3: The positioning device reports the ranging data at intervals of the first cycle, and the camera device reports the image data at intervals of the second cycle. The local monitoring device predicts the first coordinates of the construction worker based on a signal processing model and predicts the second coordinates of the construction worker based on an image processing model. Extract a group of positioning devices that have collected the ranging data of the construction worker, calculate the signal propagation time of each positioning device through the signal processing model, select a reference positioning device and establish an overdetermined equation group, and solve the overdetermined equation group to obtain the first coordinates of the construction worker. The image processing model includes a monocular sub-model and a multi-eye sub-model, refer to Figure 4 , combined with the contour size features of the construction worker, extract the image contour of the construction worker. Figure 5 The monocular sub-model predicts the observation coordinates of the image contour according to the intrinsic parameter matrix of the camera device. The multi-eye sub-model predicts the viewing angle weight of the image contour according to the extrinsic parameter matrix of the camera device, and then calculates the second coordinate of the construction worker according to the multiple observation coordinates and the corresponding viewing angle weights. The specific calculation method of the first coordinate and the second coordinate refers to Example 2.

[0022] Step 4: The early warning device adjusts the trajectory prediction model based on the first cycle and the second cycle, inputs multiple sets of first coordinates and second coordinates into the trajectory prediction model to obtain the trajectory path of the construction personnel, and outputs an early warning signal if at least one trajectory coordinate point of the trajectory path is outside the electronic fence. The noise covariance matrix of the ranging data is adjusted according to the first cycle, and the noise covariance matrix of the image data is adjusted according to the second cycle. The weight w1 of the ranging data and the weight w2 of the image data in the trajectory prediction model are respectively calculated based on the adjusted noise covariance matrix. Multiple sets of first coordinates and second coordinates are input into the trajectory prediction model. If the current moment is an integer multiple of the second cycle, the trajectory coordinate point is generated according to the first coordinate and the second coordinate. Otherwise, the first coordinate is directly used as the trajectory coordinate point. For the specific method of generating the trajectory route, refer to Example 2.

[0023] Step 5: Extract the extreme distance between the trajectory path and the electronic fence, and update the signal period of the throttling device according to the traffic flow and multiple sets of extreme distances. The smaller the extreme distance between the construction worker's trajectory path and the electronic fence, the closer the construction worker is to the traffic area, and the greater the safety risk. At the same time, the increase in traffic flow will increase the probability of vehicles in the traffic area accidentally invading the electronic fence due to congestion or lane changes, further exacerbating the safety hazards of construction workers. Therefore, it is necessary to adjust the signal period of the throttling device in real time based on the extreme distance and traffic flow to effectively reduce the operation risk of construction workers.

[0024] Calculate the instantaneous distance from each track coordinate point in the track path to the electronic fence, and store the minimum distance within the preset time as the extreme distance D, or generate an instantaneous distance function. When the derivative of the instantaneous distance function changes from a negative value to a positive value (the derivative is zero), it is the extreme distance D. The passage time T=T0[1+σ(Q / Q0)+ξ(D0 / D)], where Q is the traffic flow, Q0 is the maximum allowed flow (usually 10 vehicles / minute), D0 is the safety distance threshold (usually 0.5 meters), T0 is the basic passage time, σ is the traffic flow weight coefficient, ξ is the extreme distance weight coefficient, σ is usually 0.2-0.5, and ξ is usually 0.1-0.3. The signal cycle of the throttling device (traffic light) consists of the passage time and the prohibition time.

[0025] Step 6: Adjust the first period according to the image data of different camera devices at the same time, adjust the second period according to the distance measurement data of different positioning devices at the same time, and adjust the third period according to the first coordinate and the second coordinate at the same time. Obtain J groups of image data of different camera devices at the same time, and then calculate J groups of observation coordinates according to the image data, calculate the difference value Δ1 between the J groups of observation coordinates, and adjust the first period according to the difference value Δ1. The larger the difference value Δ1, the larger the error of the camera device, and the first period needs to be reduced to increase the acquisition frequency of the distance measurement data. Obtain the distance measurement data of I groups of positioning devices at the same time, select three groups of positioning devices to generate a positioning device combination, calculate the candidate coordinates of the construction personnel according to the three-sided positioning method and the positioning device combination, generate multiple groups of candidate coordinates, calculate the difference value Δ2 between the multiple groups of candidate coordinates, and adjust the second period according to the difference value Δ2. The larger the difference value Δ2, the larger the error of the positioning device, and the second period needs to be reduced to increase the acquisition frequency of the image data. Obtain the first coordinate and the second coordinate at the same time, calculate the difference value Δ3 between the first coordinate and the second coordinate, and adjust the third period according to the difference value Δ3. The specific adjustment method of the first period, the second period and the third period refers to Example 3.

[0026] Step 7: The local monitoring device reports the ranging data and image data every third period, and the remote monitoring device updates the signal processing model and the image processing model and returns to step 2. Calculate the theoretical signal propagation time between each positioning device and the reference positioning device in the ranging data. The actual signal propagation time is the signal propagation time recorded in the ranging data, and the theoretical signal propagation time is calculated based on the installation distance between each positioning device and the reference positioning device. Calculate the time difference between the actual signal propagation time and the theoretical signal propagation time, and use the mean of the time difference as the time difference error coefficient δ. According to the image data and the internal parameter matrix K of each camera device j and the external parameter matrix M ext,j The image contour of the construction workers is reprojected, and the reprojection formula is: , where (u' j ,v' j ) is the reprojected pixel coordinate of the image contour, and P2 is the second coordinate of the construction worker corresponding to the image contour. The perspective deviation (Euclidean distance) between the reprojected pixel coordinates and the pixel coordinates of the image contour of each camera device is calculated, the perspective deviation is normalized and the mean of the processed perspective deviation is calculated to obtain the perspective error coefficient γ. The remote monitoring device sends the time difference error coefficient δ and the perspective error coefficient γ, the local monitoring device updates the time difference error coefficient δ of the signal processing model, and updates the perspective weight (w xj , w yj , w zj ), that is, w xj = w xj / (γ+1),w yj = w yj / (γ+1),w zj = w zj / (γ+1). Embodiment 2

[0027] This embodiment discloses a preferred method for generating a first coordinate, a second coordinate, and a trajectory path.

[0028] Predict the first coordinates of the construction worker based on a signal processing model. The positioning device sends a ranging signal to the construction worker. After the construction worker's smart helmet terminal receives the ranging signal, a fixed delay will be generated due to the signal processing process. At the same time, factors such as construction equipment obstruction may cause an increase in signal propagation path loss, thereby affecting the signal propagation speed. These factors together lead to systematic deviations in the signal propagation time recorded in the ranging data, which need to be compensated and corrected through the signal processing model. Extract the I group of positioning devices that have collected the ranging data of the construction worker, and substitute the ranging data into the signal processing model to calculate the signal propagation time t of each positioning device, that is, t=(t-δ) / 2, where δ is the time difference error coefficient. Extract a reference positioning device from the I group of positioning devices, create the reference coordinates (x0, y0, z0) of the reference positioning device, and the signal propagation time t0 and the coordinates (x0, y0, z0) of any non-reference positioning device. i ,y i ,z i ) and the signal propagation time t i An overdetermined system of equations, i=1,2,...,I-1.

[0029] The overdetermined system of equations is: , i=1,2,...,I-1. The overdetermined equations are solved by, for example, the least square method or the minimum value method, and the solution of the overdetermined equations is the first coordinate of the construction worker, wherein c is the speed of light.

[0030] The second coordinate of the construction worker is predicted based on an image processing model. The image processing model includes a monocular sub-model and a multi-eye sub-model, extracts J groups of camera devices that have collected the image data of the construction worker, and extracts the image contour of the construction worker in combination with the contour size characteristics of the construction worker. The monocular sub-model is based on the intrinsic parameter matrix K of the camera device. j (including focal length and principal point offset) Convert the image contour from pixel coordinates to normalized camera coordinates. The conversion formula is: , where (x j ,y j ) is the normalized camera coordinate of the image contour, (u j ,v j ) is the pixel coordinate of the image contour. The depth z is estimated by combining the contour size features of the construction workers j , and then get the observation coordinates of the image contour (x j ,y j ,z j The multi-eye sub-model calculates the angles of the construction workers in three directions (horizontal, vertical and depth directions) relative to the optical axis of the camera device based on the external parameter matrix of the camera device (including the position coordinates and optical axis direction of the camera device). The smaller the angle, the higher the reliability of the corresponding direction and the higher the corresponding viewing angle weight. The viewing angle weight (wxj , w yj , w zj ), substitute the observation coordinates and viewing angle weights into the bundle adjustment equation, which is simplified to: , solve the bundle adjustment equation, the solution of the bundle adjustment equation is the second coordinate of the construction worker, j=1,2,...,J.

[0031] The trajectory prediction model is adjusted based on the first cycle and the second cycle. The noise covariance matrix R1 of the ranging data and the noise covariance matrix R2 of the image data are obtained. The noise covariance matrices R1 and R2 are obtained by statistical analysis of the historical errors of the ranging data and the image data, respectively. R1 reflects the ranging fluctuation characteristics of the positioning device, and R2 represents the imaging error distribution of the camera device. R1 is adjusted according to the first cycle T1, R1=αT1, and R2 is adjusted according to the second cycle T2, R2=βT2. The weight w1 of the ranging data and the weight w2 of the image data are calculated based on the adjusted noise covariance matrix, w1=1 / (|R1|+ε), w2=1 / (|R2|+ε). Among them, ε is the regularization constant term, and α and β are adjustment coefficients. The first cycle is shorter than the second cycle, and its real-time nature is more valuable in the trajectory prediction model. Therefore, the weight of the ranging data is higher, and α<β is usually set.

[0032] Reference Figure 6 to Figure 7 , multiple sets of first coordinates and second coordinates are input into the trajectory prediction model to obtain the trajectory path of the construction personnel. The trajectory path generates trajectory coordinate points according to the trajectory generation cycle, and the trajectory generation cycle is consistent with the first cycle. If the current moment t' is an integer multiple of the second cycle, a temporary coordinate point p is generated according to the first coordinate P1 and the second coordinate P2, that is, p=(w1P1+ w2P2) / (w1+ w2). Otherwise, the first coordinate P1 is directly used as the temporary coordinate point p, that is, p=P1. In this embodiment, the temporary coordinate point is used as the trajectory coordinate point.

[0033] In a more preferred embodiment, the Kalman filter generates the predicted value g2 at the current moment based on the state vector of the construction worker at the previous moment (including the trajectory coordinate point and movement speed at the previous moment), takes the temporary coordinate point p as the observation value g1, and then fuses the observation value g1 and the predicted value g2 after mapping through the observation matrix H, that is, g3= g2+K(g1-Hg2), and obtains the trajectory coordinate point at the current moment according to g3. Among them, the observation matrix H is usually , which means that only the trajectory coordinate points are observed without observing the movement speed. The Kalman gain F automatically adjusts the weight ratio of the observed value and the predicted value. When the observed value g1 deviates due to occlusion, the Kalman gain automatically reduces its weight to ensure the continuity of the trajectory coordinate points. Embodiment 3

[0034] This embodiment discloses a preferred method for adjusting the first period, the second period, and the third period.

[0035] The first cycle is adjusted according to the image data of different camera devices at the same time. J groups of image data of different camera devices at the same time are obtained, and then J groups of observation coordinates (x j ,y j ,z j ) Calculate the difference between J groups of observation coordinates ,in, is the centroid of the J groups of observation coordinates, which is the arithmetic mean of the J groups of observation coordinates. If the difference value Δ1 exceeds the first difference threshold, it indicates that the image data is inconsistent with the observation coordinates of multiple camera devices due to noise interference or occlusion effect, and the credibility of the image data is low, then the first period is reduced and the acquisition frequency of the ranging data is increased. The first difference threshold is usually 0.3-0.5 meters.

[0036] The second cycle is adjusted according to the distance measurement data of different positioning devices at the same time. The distance measurement data of I groups of positioning devices at the same time are obtained, three groups of positioning devices are selected to generate a positioning device combination, and the candidate coordinates of the construction personnel are calculated according to the three-sided positioning method and the positioning device combination to generate multiple Set candidate coordinates, and calculate multiple The difference value Δ2 between the candidate coordinates of the group, if the difference value Δ2 exceeds the second difference threshold, it means that the positioning noise of the ranging data at this time is large or there is occlusion, and high-frequency image data is needed to assist positioning, then the second period is reduced. The difference value Δ2 between the candidate coordinates is usually less than 0.4 meters. If it exceeds 1.0 meters, there may be serious occlusion, and the second difference threshold is usually 0.4-1 meters. Preferably, if I>5, 15 groups of positioning devices are randomly selected to calculate the difference value Δ2 to reduce the amount of calculation.

[0037] Adjust the third period according to the first coordinate and the second coordinate at the same time. Get the first coordinate (x1, y1, z1) and the second coordinate (x2, y2, z2) at the same time, and calculate the difference value The difference value Δ3 reflects the consistency of the ranging data and the image data. If Δ3 exceeds the third difference threshold, the third period needs to be reduced and the high frequency reporting is performed to remotely correct the signal processing model and the image processing model. The construction road belongs to the general construction area, and the third difference threshold is usually 0.5-1 meter. In this embodiment, the third difference threshold is, for example, 0.5 meter. Embodiment 4

[0038] Reference Figure 8This embodiment discloses a system for implementing the highway construction personnel management method based on trajectory positioning, including: a monitoring device, a throttling device, a camera device, a positioning device, a local monitoring device, an early warning device, a remote monitoring device, a wireless transceiver device and a database.

[0039] The monitoring device is configured to predict the traffic flow in the buffer area. The throttling device is configured to adjust the traffic flow in the buffer area. The positioning device is configured to collect the distance measurement data of the construction personnel. The camera device is configured to collect the image data of the construction personnel. The local monitoring device is configured to predict the first coordinate and the second coordinate of the construction personnel. The early warning device is configured to predict the trajectory path of the construction personnel and output an early warning signal. The remote monitoring device is configured to update the signal processing model and the image processing model. The positioning device reports the distance measurement data every first period, the camera device reports the image data every second period, the local monitoring device reports the distance measurement data and the image data every third period, and the local monitoring device adjusts the first period according to the image data of different camera devices at the same time, adjusts the second period according to the distance measurement data of different positioning devices at the same time, and adjusts the third period according to the first coordinate and the second coordinate at the same time. The wireless transceiver is used to communicate with the positioning device and feed back the distance measurement signal to the positioning device. The wireless transceiver can be installed in the smart helmet of the construction personnel. The database is used to store the signal processing model and the image processing model.

[0040] Reference Fig. 9 , the wireless transceiver may include a mobile positioning module, a wireless communication module, an alarm module, a clock module, a power supply, an environmental perception module, and a microprocessor. The mobile positioning module is a core component, which is used to feed back the ranging signal to the positioning device. The wireless communication module is used to receive the clock calibration instruction or the instruction to adjust the first cycle from the local monitoring device. The alarm module is, for example, a warning light or a buzzer. After receiving the early warning signal, the alarm module turns on the warning light or the buzzer. The clock module is used to record the moment when the mobile positioning module receives the ranging signal. The power supply is used to supply power to the mobile positioning module, the alarm module, etc. The environmental perception module increases auxiliary energy and can be used to measure environmental PM2.5 or noise. The microprocessor is used to coordinate the work of the mobile positioning module, the wireless communication module, the clock module, etc. Reference Fig.10 The smart helmet includes a cap body 10 and a brim 20. The outer side of the cap body is provided with a waterproof box body 30. The mobile positioning module, wireless communication module, clock module, microprocessor, power supply and environment sensing module of the wireless transceiver are fixed in the box body. The warning module 31 is located on the surface of the box body, and the charging interface 32 of the power supply is also provided on the surface of the box body.

[0041] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A highway construction personnel management method based on trajectory positioning, characterized in that: The following steps are involved: Step 1: Generate an electronic fence for the construction road, set a buffer area and a pass area on the construction road based on the electronic fence, set a monitoring device in the pass area, and set a throttling device in the buffer area; Step 2: The monitoring device predicts the traffic flow in the buffer area, the throttling device adjusts the traffic flow in the buffer area, the positioning device periodically collects the distance measurement data of the construction personnel, and the camera device periodically collects the image data of the construction personnel; Step 3: The positioning device reports the distance measurement data at a first period interval, the camera device reports the image data at a second period interval, and the local monitoring device predicts the first coordinates of the construction personnel based on a signal processing model and predicts the second coordinates of the construction personnel based on an image processing model; Step 4: The early warning device adjusts the trajectory prediction model based on the first cycle and the second cycle, inputs multiple sets of first coordinates and second coordinates into the trajectory prediction model to obtain the trajectory path of the construction personnel, and outputs an early warning signal if at least one trajectory coordinate point of the trajectory path is outside the electronic fence; Step 5: extracting the extreme value distance between the trajectory path and the electronic fence, and updating the signal period of the throttling device according to the traffic flow and multiple sets of extreme value distances; Step 6: adjusting the first period according to the image data of different camera devices at the same time, adjusting the second period according to the ranging data of different positioning devices at the same time, and adjusting the third period according to the first coordinates and the second coordinates at the same time; Step 7: The local monitoring device reports the distance measurement data and image data every third period, and the remote monitoring device updates the signal processing model and the image processing model, and returns to step 2.

2. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 3, extract the I group of positioning devices that have collected the distance measurement data of the construction worker, substitute the distance measurement data into the signal processing model to calculate the signal propagation time of each positioning device, extract a reference positioning device from the I group of positioning devices, create the reference coordinates (x0, y0, z0) of the reference positioning device and the signal propagation time t0 and the coordinates (x0, y0, z0) of any non-reference positioning device. i ,y i ,z i ) and the signal propagation time t i The overdetermined set of equations is used to calculate the first coordinate of the construction worker, i=1,2,...,I-1.

3. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 3, the image processing model includes a monocular sub-model and a multi-eye sub-model, which extracts J groups of camera devices that have collected image data of the construction worker, extracts the image contour of the construction worker in the image data, and the monocular sub-model predicts the observation coordinates of the image contour based on the intrinsic parameter matrix of the camera device. The multi-eye sub-model predicts the viewing angle weight of the image contour based on the extrinsic parameter matrix of the camera device, and then calculates the second coordinate of the construction worker based on multiple observation coordinates and corresponding viewing angle weights.

4. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 4, the noise covariance matrix of the ranging data is adjusted according to the first cycle, and the noise covariance matrix of the image data is adjusted according to the second cycle. The weight w1 of the ranging data and the weight w2 of the image data in the trajectory prediction model are respectively calculated based on the adjusted noise covariance matrix.

5. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 4, multiple sets of first coordinates and second coordinates are input into the trajectory prediction model. If the current time is an integer multiple of the second period, the trajectory coordinate point is generated according to the first coordinate and the second coordinate. Otherwise, the first coordinate is directly used as the trajectory coordinate point.

6. The highway construction personnel management method based on trajectory positioning according to claim 3 is characterized in that: In step 6, J groups of image data from different camera devices at the same time are acquired, and then J groups of observation coordinates are calculated based on the image data, the difference value Δ1 between the J groups of observation coordinates is calculated, and the first period is adjusted based on the difference value Δ1.

7. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 6, the ranging data of I groups of positioning devices at the same time are obtained, three groups of positioning devices in the ranging data are selected to generate a positioning device combination, candidate coordinates are calculated according to the three-sided positioning method and the positioning device combination, multiple groups of candidate coordinates are generated, the difference value Δ2 between the multiple groups of candidate coordinates is calculated, and the second period is adjusted according to the difference value Δ2.

8. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 6, the first coordinate and the second coordinate at the same time are obtained, the difference value Δ3 between the first coordinate and the second coordinate is calculated, and the third period is adjusted according to the difference value Δ3.

9. The highway construction personnel management method based on trajectory positioning according to claim 1 is characterized in that: In step 7, the remote monitoring device calculates the theoretical signal propagation time between any non-reference positioning device and the reference positioning device based on the ranging data, calculates the time difference between the actual signal propagation time and the theoretical signal propagation time, takes the mean of the time difference as the time difference error coefficient δ, reprojects the image data, calculates the viewing angle deviation between the reprojected pixel coordinates and the pixel coordinates of the image contour of each camera device, takes the mean of the viewing angle deviation as the viewing angle error coefficient γ, and the local detection device updates the signal processing model and the image processing model according to δ and γ, respectively.

10. A system for implementing the highway construction personnel management method based on trajectory positioning according to claim 1, characterized in that: include: a monitoring device, the monitoring device being configured to predict traffic flow in the buffer area; a throttling device, the throttling device being configured to adjust the flow of traffic in the buffer area; A positioning device, the positioning device is configured to collect distance measurement data of construction personnel; A camera device, the camera device is configured to collect image data of construction personnel; A local monitoring device, the local monitoring device is configured to predict a first coordinate and a second coordinate of a construction worker; An early warning device, the early warning device is configured to predict the trajectory path of the construction personnel and output an early warning signal; A remote monitoring device, the remote monitoring device is configured to update the signal processing model and the image processing model, wherein: The positioning device reports the distance measurement data at intervals of a first cycle, the camera device reports the image data at intervals of a second cycle, and the local monitoring device reports the distance measurement data and the image data at intervals of a third cycle. The local monitoring device adjusts the first period according to the image data of different camera devices at the same time, adjusts the second period according to the ranging data of different positioning devices at the same time, and adjusts the third period according to the first coordinates and the second coordinates at the same time.

Images