A monitoring method for dangerous intrusion of lifting equipment based on cloud-edge collaboration

Abstract

A monitoring method for dangerous intrusions in lifting and hoisting based on cloud-edge collaboration, including deploying monitoring equipment and alarms to calibrate dangerous areas in lifting and hoisting; collecting monitoring image information, crane operation status information and hook height information, constructing a monitoring data group after multi-data fusion, and sending it in real time; processing the monitoring data group and judging dangerous intrusions, and providing alarm feedback to the lifting and hoisting site. The method utilizes the feasibility of time-sharing multiplexing of hardware resources, multi-source data fusion monitoring, and effectively monitors dangerous intrusions in a cloud-edge collaborative manner.

Description

Technical Field

[0001] The present invention relates to the field of monitoring, and in particular to a method for monitoring dangerous intrusions of lifting equipment based on cloud-edge collaboration. Background Art

[0002] Urban infrastructure is usually large-scale, and there is a need for on-site assembly of large prefabricated components according to the construction plan. Prefabricated components usually need to be hoisted and transported according to the on-site construction conditions, and assembled on-site at the planned deployment location. Large lifting equipment usually works in outdoor scenes and is affected by the weather. During the hoisting and transportation of prefabricated components, falling objects from high altitudes may injure people. To avoid risks, signal workers are equipped to conduct manual on-site inspections when the on-site crane is working, supervise unsafe behaviors such as entering dangerous areas, and communicate on-site conditions with the crane operator in the cab through walkie-talkies.

[0003] Among large lifting equipment, bridge / gantry cranes are mainly used in the lifting and transportation of large materials in the field of engineering and construction. Due to the height gradient difference of the construction plane or the insufficient light entering the bottom of the foundation pit caused by the underground construction scene, there is a blind spot in the crane cab. However, in order to improve the efficiency of on-site construction, personnel are allowed to pass under the crane when the crane is not in operation. "Intrusion" occurs frequently but is not a dangerous behavior. When the crane is in operation, it is a compliant behavior to enter the blind spot of the cab under the crane when the rigger fixes the large materials that need to be transferred or the auxiliary prefabricated components are stably landed. It can be seen that there is a risk of falling objects from high altitude when the crane is working, and it is necessary to supervise personnel who break into the area under the hook and are not special workers; when the crane is not working, personnel are allowed to pass through the area under the hook to improve construction efficiency; when the crane is working, special workers are allowed to break into the hook in compliance with regulations to perform the task of fixing construction materials or auxiliary materials landing. In view of the above complex working conditions, the safety supervision of the crane working status is carried out manually on site.

[0004] The existing monitoring method usually sets a fixed-position camera in the hoisting area to monitor through images and acquisition processing. The construction site operation process is complicated and the safety management relies heavily on manual labor. Although the camera can adjust the angle and some positions to expand the monitoring field of view, the timeliness and effectiveness of the monitoring are low, and the response speed of the monitoring feedback is slow and there are many blind spots. In the existing monitoring method, there is also a way to use sensor equipment for data collection and processing to assist in monitoring, but the sensor data is applied independently, the data collection integration is poor, and there is a lack of systematic multi-data fusion monitoring. It is impossible to effectively determine whether the person passing under the crane hook is a "dangerous" behavior, thereby achieving effective monitoring of dangerous intrusions into lifting and hoisting. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a monitoring method for dangerous intrusions in lifting and hoisting based on cloud-edge collaboration, which utilizes the feasibility of time-sharing reuse of hardware resources and multi-source data fusion monitoring, and processes and analyzes the operating status data through a cloud server in a cloud-edge collaborative manner, thereby reducing the scale of hardware deployment while improving the utilization efficiency of computing resources, and ultimately achieving effective monitoring of dangerous intrusions in lifting and hoisting scenarios.

[0006] The present invention provides a method for monitoring dangerous intrusion of lifting equipment based on cloud-edge collaboration, comprising the following steps performed in sequence:

[0007] (1) Deploy monitoring equipment and alarms at the lifting site, where the monitoring equipment includes image acquisition devices deployed on the main shafts of multiple cranes, sensor devices deployed on the fixed part of the hook cable, and height absolute encoders deployed at the position of the hook lifting limiter;

[0008] (2) Place the crane's main hook bucket horizontally on the ground and calibrate the lifting and hoisting danger zone based on an area twice the bucket area;

[0009] (3) Periodically collect monitoring image information, crane operation status information, and hook height information, perform multi-data fusion to construct a monitoring data group, and send it to the cloud server in real time;

[0010] (4) Processing the monitoring data group, judging the dangerous intrusion of the lifting equipment, and sending the alarm information to the control background and feedback to the alarm when it is judged that there is a dangerous intrusion;

[0011] (5) Use the alarm to play alarm information and provide alarm feedback to the lifting and hoisting site.

[0012] Among them, in the step (1), the image acquisition device is a camera, which is used to acquire images of the monitored area; the sensing device is a gravity sensor, which is used to measure whether the hook is hoisting materials and gravity data; and the height absolute value encoder is used to measure the hook height information.

[0013] Wherein, the step (1) further includes the step of: using a cloud server to analyze the monitoring image information, and processing the crane operation status information to identify the crane operation status.

[0014] Wherein, in said step (3), monitoring image information, crane operation status information and hook height information are periodically collected respectively, and a monitoring data group is constructed after multi-data fusion, which specifically includes the following steps:

[0015] (3.1) Collect monitoring image information, crane operation status information and hook height information respectively according to a certain period;

[0016] (3.2) Taking the acquisition time point as the benchmark, the acquired monitoring image information, crane operation status information and hook height information are bound together to construct a data group;

[0017] (3.3) Compare the data set constructed later with the data set constructed in the previous cycle. If the data has changed, the data set constructed later is sent to the cloud server; if the data has not changed, the acquisition time point in the data set constructed later is kept unchanged, and the acquired monitoring image information, crane operation status information and hook height information are recorded as 0, and then the data set constructed later is updated and sent to the cloud server;

[0018] Wherein, in said step (4), the intrusion danger of the lifting and hoisting equipment is judged, including an intrusion danger judgment step and an estimated danger judgment step.

[0019] The intrusion risk determination step specifically includes the following steps:

[0020] (4.1.1) Determine whether the image recognition finds a dangerous person breaking in. If so, proceed to the next step;

[0021] (4.1.2) Determine whether the hook weight is greater than the weight threshold, if yes, proceed to the next step; if no, return to step (a);

[0022] (4.1.3) Determine whether the hook height is greater than the height threshold, if yes, proceed to the next step; if no, return to step (a);

[0023] (4.1.4) Determine if there is a risk of intrusion and send an alarm message.

[0024] The step of estimating risk and judging the risk specifically includes the following steps:

[0025] (4.2.1) Compare the changes in the corresponding data in the two groups of monitoring data before and after, identify and obtain the position change and change rate based on the monitoring image information and hook height information, and obtain the gravity data change and change rate based on the crane operation status information;

[0026] (4.2.2) respectively setting a position change rate threshold space and a gravity change rate threshold space, and comparing the position change rate and gravity data change rate obtained in step (4.2.1) with the position change rate threshold space and the gravity change rate threshold space respectively;

[0027] (4.2.3) When the position change rate and gravity data change rate are out of their corresponding threshold space, the prediction of the danger level of dangerous intrusion is initiated;

[0028] (4.2.4) Determine whether the image recognition finds a dangerous person breaking in, whether the weight of the hook is greater than the weight threshold, and whether the height of the hook is greater than the height threshold. If any of the results is yes, it is determined that there is a danger of intrusion and an alarm message is sent;

[0029] (4.2.5) The position change and rate of change and the gravity data change and rate of change are stored for subsequent analysis.

[0030] Among them, the processing of the monitoring data group in the step (4) also includes setting a dangerous intrusion monitoring algorithm for the crane operation state, and the algorithm includes the steps of image recognition of passing personnel under vertical pitch conditions, crane weight classification threshold design and crane height parameter threshold design.

[0031] The process of image recognition of pedestrians under the vertical depression angle condition includes the following steps:

[0032] (A) grayscale processing is performed on the collected input image;

[0033] (B) Perform gamma correction on the image; suppose the pixel value is X, the gamma value of the reference camera image is 2.2, and the X conversion expression is:

[0034]

[0035] (C) Calculate the gradient size and direction of the pixel points in the image after gamma correction;

[0036] (D) Divide the image into 8×8 blocks, calculate its gradient histogram, and construct the descriptor of each block;

[0037] (E) After calculating the block processing, the block descriptors of the image are connected in series in order to obtain the HOG feature descriptor of the image;

[0038] Using the XGBoost model, using Taylor's second-order expansion, the expression is retained while retaining the first-order derivative and the second-order derivative:

[0039]

[0040] To minimize the objective function, we can set its derivative to 0, and get

[0041] in,

[0042] The steps of designing the crane weight classification threshold are as follows: for small cranes with a load less than 5 tons, 5% of the design load is selected as the threshold flag. M , the load is expressed in M ​​tons; for other standard cranes, refer to the bucket weight on site and select 2 tons as the threshold. The specific expression is as follows:

[0043]

[0044] The steps of designing the height parameter threshold of the crane are based on the height difference between the operator's height and the hook's height above the ground, and a height of 3 meters is selected as the height threshold flag. H , the expression is as follows: flag M =3m.

[0045] The cloud-edge collaborative crane and hoisting dangerous intrusion monitoring method of the present invention realizes an automated identification and disposal method to define and supervise whether the construction personnel's passing under the hook when the crane is working is a dangerous behavior. Compared with the prior art, it can achieve:

[0046] 1) Assisting on-site safety management in an automated manner. By deploying a camera on the main shaft of the crane, the area directly below the hook in the blind spot of the cab is monitored for intrusion by personnel. To distinguish the working status of the crane, an absolute value encoder is deployed on the lifting limiter of the hook to measure the height information of the hook, and a weight sensor is deployed next to the fixed end of the hook cable to measure the weight information of the hook. The working status of the crane is pattern recognized by combining the hook height and weight information.

[0047] 2) Comprehensively consider the feasibility of time-sharing reuse of hardware resources, adopt the design concept of cloud-edge collaboration, process and analyze the operation status data through cloud servers, solve the practical problem of long physical distances between cranes on the construction site, and use the feasibility of time-sharing reuse of hardware resources to reduce the scale of hardware deployment and improve the efficiency of computing resource utilization; deploy high-power sound amplification devices on the edge side to provide timely feedback on the results of comprehensive analysis of multi-source data on site. Comprehensively realize effective monitoring of dangerous intrusion behaviors in an automated manner when the crane is working.

[0048] 3) Optimize the design of the judgment of the dangerous intrusion of lifting and hoisting equipment, set up the intrusion danger judgment steps and the estimated danger judgment steps for comprehensive judgment, and improve the monitoring efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of the monitoring method for dangerous intrusions in lifting and hoisting based on cloud-edge collaboration. DETAILED DESCRIPTION

[0050] The specific implementation of the present invention is described in detail below. It is necessary to point out that the following implementation is only used for further explanation of the present invention and cannot be understood as limiting the scope of protection of the present invention. Some non-essential improvements and adjustments made to the present invention by persons skilled in the art based on the above-mentioned content of the present invention still fall within the scope of protection of the present invention.

[0051] The present invention provides a method for monitoring dangerous intrusion of lifting equipment based on cloud-edge collaboration, as shown in the attached Figure 1 As shown, it is a flow chart of the implementation of the monitoring method for dangerous intrusions in the operating status of a bridge / Gantry crane based on cloud-edge collaboration. It also schematically shows the application scenario of the monitoring method for dangerous intrusions in the operating status of a bridge / Gantry crane based on cloud-edge collaboration according to the present invention, which is described in detail below.

[0052] The present invention provides a method for monitoring dangerous intrusions of lifting and hoisting equipment based on cloud-edge collaboration, comprising the following steps performed in sequence.

[0053] First, monitoring equipment and alarms are deployed at the lifting site, wherein the monitoring equipment includes image acquisition devices respectively deployed on the main shaft parts of multiple cranes and sensor devices deployed on the hook cable fixing parts. The image acquisition devices and sensor devices are respectively used to collect monitoring image information and crane operation status information. The collected monitoring image information and crane operation status information can be normalized to the same time axis coordinate system, and the behavior of construction personnel breaking into the bottom of the crane hook can be finely pattern recognized to cope with the complex working conditions of the crane. The image acquisition device and the sensor device are respectively connected to the cloud server set up at the remote end through a dedicated line. In a preferred manner, the image acquisition device is a camera and the sensor device is a gravity sensor. Specifically, the camera is used to collect images of the monitoring area, and the weight sensor is used to measure whether the hook is hoisting materials. The cloud server can realize the analysis of the monitoring image information, and realize the recognition of the crane operation status by processing the crane operation status information. When construction is carried out at the lifting site, the alarm can provide on-site feedback on the analysis and recognition results and issue an alarm. In a preferred manner, the alarm selects a high-power amplification device, and the crane is selected as a bridge / gantry crane.

[0054] It should be noted that, according to the actual situation on site, the work of multiple cranes is usually discontinuous in time, and the physical distribution distance between multiple cranes is usually far. Therefore, cameras are deployed on the main shaft of the crane, and connected to the cloud server through a dedicated line. The cloud server processes and analyzes the image data, which can realize the intrusion monitoring of personnel in the blind spots of the vision of multiple cranes. At the same time, the cloud server is used to analyze the monitoring image information, process the crane operation status information, and identify the crane operation status.

[0055] When the crane is not in operation, construction workers are allowed to pass through the area under the crane. There are obvious signs to indicate whether the crane is in operation, that is, it is in operation when the crane is hoisting materials, and it is in non-operating state when the crane is not hoisting materials. Therefore, by deploying a weight sensor near the fixed end of the hook cable, it is possible to measure and judge whether the hook is hoisting materials, and the measurement results of the weight sensor can be sent to the cloud server for data processing and periodic updates. It should be emphasized that, considering the height of the rope operator and the height difference between the hook and the ground, a height absolute encoder is deployed at the hook lifting limiter position to measure the hook height information, and the measurement results of the height absolute encoder are sent to the cloud server for data processing and periodic updates.

[0056] The height gradient difference will bring about the problem of beyond-horizon. The positional relationship between pixels and physical distance changes with the rise and fall of the main hook. When the crane is not in operation, operators are allowed to pass through the area below the crane. When the crane is in operation, the rigging workers are allowed to support the building materials to land steadily. The above situations can cause false alarms in the personnel intrusion monitoring. In order to distinguish the working state of the crane, a height absolute encoder is deployed on the lifting limiter of the hook to measure the height information of the hook, and a weight sensor is deployed next to the fixed end of the hook cable to measure the weight information of the hook. The working state of the crane is pattern recognized by combining the hook height and weight information.

[0057] In order to eliminate the influence of inaccurate monitoring data caused by the change of pixels and physical distance of the camera as the crane rises and falls, the present invention places the main hook bucket horizontally on the ground and calibrates the dangerous area with an area twice the area of ​​the bucket. When the crane rises, the scope of the dangerous area expands accordingly, which can effectively deal with the problem of offset and falling caused by unbalanced force of materials hoisted in the air.

[0058] In the process of monitoring dangerous intrusion, monitoring image information, crane operation status information and hook height information are collected periodically, and a monitoring data group is constructed after multi-data fusion. Specifically, monitoring image information, crane operation status information and hook height information are collected respectively according to a certain period. Among them, each collection cycle has a corresponding collection time point, so the collection time point is used as a benchmark to bind the collected monitoring image information, crane operation status information and hook height information to construct a data group. The construction form of the array can be constructed by arranging multiple data in sequence after base conversion. Since some collected data does not change in some states, if the completed data group is stored every time, storage space will be wasted and the transmission speed will also be affected. Therefore, after the data group is constructed, the later constructed data group is compared with the data group constructed in the previous cycle. If the data changes, the later constructed data group is sent to the cloud server; if there is no change, the collection time point in the later constructed data group remains unchanged, and the collected monitoring image information, crane operation status information and hook height information are recorded as 0, and then the later constructed data group is updated and sent to the cloud server;

[0059] It should be noted that the present application focuses on specific settings for the method of judging dangerous intrusion situations, and its judgment efficiency is higher. Specifically, when judging the dangerous intrusion situation of lifting and hoisting, an intrusion danger judgment step and an estimated danger judgment step are set.

[0060] The intrusion danger judgment step specifically includes the following steps: judging whether the image recognition finds a dangerous person intruding, if yes, proceeding to the next step; judging whether the weight of the hook is greater than the weight threshold, if yes, proceeding to the next step; if no, returning to the recognition step; judging whether the height of the hook is greater than the height threshold, if yes, proceeding to the next step; if no, returning to the recognition step; judging that there is an intrusion danger, sending an alarm message. When judging, this threshold judgment method cannot confirm the danger alone, so it is realized through multi-threshold progressive judgment.

[0061] For the estimated danger judgment step, it specifically includes the following steps: compare the changes in the corresponding data in the two groups of monitoring data before and after, obtain the position change and change rate based on the monitoring image information and the hook height information, and obtain the gravity data change and change rate based on the crane operation status information; set the position change rate threshold space and the gravity change rate threshold space respectively, and compare the position change rate and the gravity data change rate obtained in the previous step with the position change rate threshold space and the gravity change rate threshold space respectively; when any of the position change rate and the gravity data change rate is not in its corresponding threshold space, start the prediction of the degree of danger of dangerous intrusion; judge whether the image recognition finds that a dangerous person has intruded, whether the hook weight is greater than the weight threshold, and whether the hook height is greater than the height threshold. When any of the results is yes, it is judged that there is an intrusion danger and an alarm message is sent; the position change and change rate and the gravity data change and change rate are stored for subsequent analysis.

[0062] In addition, since image recognition, parameter feedback and update, and on-site alarms (high-power sound amplification equipment) all have a certain degree of time lag, in order to reduce the probability of false alarms and increase response speed, the following flowchart can be used for collaborative information processing and on-site feedback.

[0063] Specifically, a dangerous intrusion monitoring algorithm for the crane operation status is set up, including image recognition of people passing by under vertical pitch conditions, crane weight classification threshold design and crane height parameter threshold design.

[0064] In a specific implementation, the on-site background environment is usually highly complex, and there are fewer human features at vertical depression angles. The height of the crane above the ground is about 8-15 meters. If there is underground construction, the height underground may even reach more than 40 meters. The recognition effect of the personnel recognition algorithm under vertical depression conditions directly affects the dangerous intrusion monitoring of the crane operation status. Therefore, the HOG+XGBoost model is selected to monitor the intrusion of personnel in the area directly below the crane hook. The HOG extraction algorithm screens and clusters image features through gradient histograms, and can effectively perform edge recognition based on gradient descent.

[0065] Specifically, the image recognition process of a person passing by under vertical pitch angle conditions is as follows:

[0066] (1) grayscale the collected input image;

[0067] (2) Perform gamma correction on the image; suppose the pixel value is X, and the gamma value of the image captured by the reference camera is 2.2, then the X conversion expression is:

[0068]

[0069] (3) Calculate the gradient size and direction of the pixel points in the image after gamma correction;

[0070] (4) Divide the image into 8×8 blocks, calculate its gradient histogram, and construct the descriptor of each block;

[0071] (5) After the block processing, the descriptors of each block of the image are connected in series to obtain the HOG feature descriptor of the image.

[0072] The XGBoost model comes with a regularization function and uses Taylor's second-order expansion to retain richer image edge features. The Taylor expansion expression of the XGBoost model retaining the first-order derivative and the second-order derivative is:

[0073]

[0074] To minimize the objective function, we can set its derivative to 0, and get

[0075] in,

[0076] Parameters in the XGBoost model, such as the depth of a single learner, the learning rate of a weak learner, and the maximum number of weak learners, need to be gradually set and optimized based on the real-life image library.

[0077] In the design of crane weight classification threshold, for small cranes with a design load of less than 5 tons (design load is expressed in M ​​tons), considering the sensitivity and resolution of the weight sensor, 5% of the design load can be selected as the threshold flag. M For other standard cranes, refer to the bucket weight on site and select 2 tons as the threshold. The expression is as follows:

[0078]

[0079] When designing the height parameter threshold of the crane, the height difference between the operator and the hook height above the ground is considered, and a height of 3 meters is selected as the height threshold flag. H , the expression is as follows: flag M =3m.

[0080] Although exemplary embodiments of the present invention have been described for the purpose of illustration, it will be understood by those skilled in the art that various modifications, additions and substitutions may be made in form and detail without departing from the scope and spirit of the invention disclosed in the appended claims, and all such changes shall fall within the scope of protection of the appended claims of the present invention, and the various steps in the various departments and methods of the products claimed by the present invention may be combined together in any combination. Therefore, the description of the embodiments disclosed in the present invention is not intended to limit the scope of the present invention, but is used to describe the present invention. Accordingly, the scope of the present invention is not limited by the above embodiments, but is defined by the claims or their equivalents.

Claims

1. A method for monitoring dangerous intrusion of lifting equipment based on cloud-edge collaboration, characterized in that: The method comprises the following steps which are performed in order: (1) Deploy monitoring equipment and alarms at the lifting site, where the monitoring equipment includes image acquisition devices deployed on the main shafts of multiple cranes, sensor devices deployed on the fixed part of the hook cable, and height absolute encoders deployed at the position of the hook lifting limiter; (2) Place the crane's main hook bucket horizontally on the ground and calibrate the lifting and hoisting danger zone based on an area twice the bucket area; (3) Periodically collect monitoring image information, crane operation status information, and hook height information, perform multi-data fusion to construct a monitoring data group, and send it to the cloud server in real time; (4) Processing the monitoring data group, judging the dangerous intrusion of the lifting equipment, and sending the alarm information to the control background and feedback to the alarm when it is judged that there is a dangerous intrusion; (5) Use the alarm to play alarm information and provide alarm feedback to the lifting and hoisting site; Wherein, in said step (3), monitoring image information, crane operation status information and hook height information are periodically collected respectively, and a monitoring data group is constructed after multi-data fusion, which specifically includes the following steps: (3.1) Collect monitoring image information, crane operation status information and hook height information respectively according to a certain period; (3.2) Taking the acquisition time point as the benchmark, the acquired monitoring image information, crane operation status information and hook height information are bound together to construct a data group; (3.3) The data group constructed later is compared with the data group constructed in the previous cycle. If the data has changed, the data group constructed later is sent to the cloud server; if the data has not changed, the collection time point in the data group constructed later is kept unchanged, and the collected monitoring image information, crane operation status information and hook height information are recorded as 0. After that, the data group constructed later is updated and then sent to the cloud server.

2. The method according to claim 1, characterized in that: In the step (1), the image acquisition device is a camera for acquiring images of the monitored area; the sensing device is a gravity sensor for measuring whether the hook is hoisting materials and gravity data; Height absolute encoder is used to measure the hook height information.

3. The method according to claim 2, characterized in that: The step (1) also includes the step of analyzing the monitoring image information using a cloud server, and processing the crane operation status information to identify the crane operation status.

4. The method according to claim 1, characterized in that: In the step (4), the intrusion danger of the lifting and hoisting equipment is judged, including an intrusion danger judgment step and an estimated danger judgment step.

5. The method according to claim 4, characterized in that: The intrusion risk judgment step specifically includes the following steps: (4.1.1) Determine whether the image recognition finds a dangerous person breaking in. If so, proceed to the next step; (4.1.2) Determine whether the hook weight is greater than the weight threshold, if yes, proceed to the next step; if no, return to step (4.1.1); (4.1.3) Determine whether the hook height is greater than the height threshold, if yes, proceed to the next step; if no, return to step (4.1.1); (4.1.4) Determine if there is a risk of intrusion and send an alarm message.

6. The method according to claim 4 or 5, characterized in that: The step of estimating risk and judging specifically comprises the following steps: (4.2.1) Compare the changes in the corresponding data in the two groups of monitoring data before and after, identify and obtain the position change and change rate based on the monitoring image information and hook height information, and obtain the gravity data change and change rate based on the crane operation status information; (4.2.2) respectively setting a position change rate threshold space and a gravity change rate threshold space, and comparing the position change rate and gravity data change rate obtained in step (4.2.1) with the position change rate threshold space and the gravity change rate threshold space respectively; (4.2.3) When the position change rate and gravity data change rate are out of their corresponding threshold space, the prediction of the danger level of dangerous intrusion is initiated; (4.2.4) Determine whether the image recognition finds a dangerous person breaking in, whether the weight of the hook is greater than the weight threshold, and whether the height of the hook is greater than the height threshold. If any of the results is yes, it is determined that there is a danger of intrusion and an alarm message is sent; (4.2.5) The position change and rate of change and the gravity data change and rate of change are stored for subsequent analysis.

7. The method according to claim 6, characterized in that: The processing of the monitoring data group in step (4) also includes setting a crane operation status dangerous intrusion monitoring algorithm, the algorithm including the steps of image recognition of passers-by under vertical pitch angle conditions, crane weight classification threshold design and crane height parameter threshold design.

Images