Penetration type management method for railway construction engineering operators and application

By adopting a penetrating management method in railway construction projects and using intelligent terminals and visual feature algorithms for automated video inspection, the problems of low management efficiency and frequent safety accidents are solved, and efficient inspection and safety management are achieved, reducing the number of The accident rate.

CN118942002APending Publication Date: 2024-11-12INST OF COMPUTING TECH CHINA ACAD OF RAILWAY SCI +2
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410965375.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Railway construction projects have low management efficiency, difficult supervision, frequent safety accidents, large personnel mobility, uneven skill levels, and insufficient emergency response capabilities.

Method used

Using the penetrating management method, the inspection number is generated based on the random number algorithm at every set time through the management center server, the video inspection is carried out using the intelligent terminal, and the necessary feature items of the operator and the work point are extracted and compared based on the visual feature algorithm. Achieve automated inspections, reduce human interference, and improve inspection efficiency.

Benefits of technology

It improves management efficiency, reduces the incidence of safe construction, improves the safety of work sites, enhances emergency response capabilities, and reduces the incidence of irregular operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942002A_ABST
    Figure CN118942002A_ABST
Patent Text Reader

Abstract

The invention discloses a penetration type management method for railway construction engineering operators and application, and relates to the technical field of railway construction engineering operator management.The penetration type management method can adopt a penetration type management mode, a management center server can directly conduct periodic video inspection on the operators, and the management efficiency is improved. A polling object is generated based on a random number algorithm, greeting type polling is avoided, polling is completely and autonomously carried out based on the algorithm, and polling content is collected based on an intelligent terminal carried by an operator; the management center server can be in communication connection with the intelligent terminal and then guide workers to complete penetration type visual inspection through the intelligent terminal, correctness inspection of on-duty personnel, inspection of equipment states and inspection of safety facilities can be achieved, the number of manual interference factors of inspection is small, and the inspection efficiency is high; the working enthusiasm of workers can be improved, the working state of meeting inspection can be kept all the time, the incidence rate of non-standard operation can be greatly reduced, the safety of work points is improved, and the accident rate is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of railway construction engineering operator management, and in particular to a railway construction engineering operator penetration management method and application. Background Art

[0002] The management of railway construction project operators is a complex process involving multiple aspects, including personnel scheduling, safety supervision, and efficiency improvement.

[0003] The existing management methods for railway construction engineering personnel usually include the following aspects:

[0004] 1. Hierarchical management system. Railway construction projects usually adopt a hierarchical management system, from the project headquarters to the construction team, and then to the specific work team, forming a hierarchical management structure. Each level has a clear person in charge who is responsible for managing and guiding the work of the next level.

[0005] 2. Quality and safety control. Railway construction projects emphasize quality and safety control. Usually, there will be a dedicated quality and safety department responsible for supervising the quality of the project and the safety of the construction site. This includes regular safety inspections, quality reviews, and the formulation and implementation of construction specifications.

[0006] 3. Progress management. Project progress is the key to railway construction, so there will be strict progress management. By formulating detailed construction plans and schedules, monitoring the progress of the project and ensuring it is completed on time.

[0007] 4. Technical training and assessment: Railway construction workers need to have certain professional skills, so there will be regular technical training and safety education. In addition, the skills of the operators will be assessed to ensure that they have the necessary operational capabilities.

[0008] 5. Labor protection and welfare. Taking into account the labor intensity and risks of railway construction projects, corresponding labor protection measures and welfare guarantees will be provided to the workers, such as safety protection equipment, medical insurance, etc.

[0009] However, the existing management methods also have some shortcomings:

[0010] 1. Inefficient management: The hierarchical management structure may lead to poor information transmission and slow decision-making process, thus affecting management efficiency.

[0011] 2. Supervision is difficult. Railway construction projects are often large-scale and cover a wide range of areas, making it difficult for supervisory departments to conduct effective and real-time supervision of each construction site.

[0012] 3. Safety accidents occur frequently. Despite a strict safety management system, safety accidents are still difficult to completely avoid due to the complexity and unpredictability of the construction environment.

[0013] 4. There is a high turnover of personnel. Railway construction projects often require a large number of temporary workers and outsourced workers. The high turnover of personnel increases the difficulty of management.

[0014] 5. The skill levels of the operators vary, which may lead to unstable construction quality.

[0015] 6. Inadequate emergency response capabilities. In the event of an incident or emergency, the existing management system may not be able to respond quickly and effectively.

[0016] In order to solve these problems, it is necessary to continuously improve and perfect the management methods, introduce more advanced management technologies and methods, and improve the management efficiency and safety of railway construction projects. Summary of the invention

[0017] In order to solve the above technical problems, the present invention provides a method and application for penetrating management of railway construction project operators. The following technical solutions are adopted:

[0018] A penetrating management method for railway construction project operators, comprising the following steps: Step 1, archiving a plurality of operators and work sites respectively to obtain operator file data and work site file data, wherein the operator file data includes an operator necessary feature set, and the work site file data includes a work site necessary feature set;

[0019] Step 2: Multiple operators carry smart terminals to work, and the operators are numbered respectively;

[0020] Step 3: The management center server generates inspection numbers of a set proportion of several staff members based on a random number algorithm at set intervals. The management center server wirelessly communicates with the smart terminals carried by several operators in the order of the inspection numbers to perform penetrating video inspections. The inspection staff members hold smart terminals to record the inspection video data to be tested.

[0021] Step 4: The management center server extracts a plurality of necessary feature items of operators and a plurality of necessary feature items of corresponding work points from the inspection video data to be inspected based on a visual feature algorithm;

[0022] Based on the visual similarity comparison algorithm, multiple necessary feature items of operators are compared with the necessary feature set of operators, and multiple necessary feature items of work points are compared with the necessary feature set of work points. If the comparison is incomplete, the missing items of necessary features of operators and work points are output, and warnings are issued and manual re-inspection is performed.

[0023] By adopting the above technical solution and a penetrating management method, the management center server can directly conduct periodic video inspections on the staff, and the inspection objects are generated based on a random number algorithm to avoid greeting-style inspections. The inspections are completely autonomous based on the algorithm, and the collection of inspection content is based on the smart terminal carried by the operating personnel. The smart terminal can be a customized smart terminal with video interaction function and voice interaction function, or it can be replaced by a smart phone. The management center server can communicate with the smart terminal and then guide the staff to complete the penetrating visual inspection through the smart terminal, which can realize the correctness inspection of on-duty personnel, the inspection of equipment status and the inspection of safety facilities. The inspection content is complete, there are few human interference factors in the inspection, and the inspection efficiency is high. For those that cannot pass the autonomous visual inspection, manual review is required. The review can also be implemented through video calls based on smart terminals, realizing vertical penetrating management, improving management efficiency and reducing the incidence of safe construction.

[0024] Optionally, the operator's file data includes related information and a set of necessary features for the operators. The related information includes the operator's real-name information, file photo information, position information, contract information, health management information, work evaluation information and current construction site information; the set of necessary features for the operators includes visual feature items of multiple necessary tools carried by the operator and visual feature items of a selfie taken with a work certificate held in the operator's hand.

[0025] By adopting the above technical solution, the operating personnel file data can display the relevant information of the staff in detail, and the management personnel can query it on the management center server with authorization. The operating personnel's necessary feature set needs to include visual feature items of multiple necessary tools carried and visual feature items of selfies taken with work certificates. The visual feature items of multiple necessary tools carried can verify the tool carrying situation of the staff while on the job, and avoid the staff's irregular operations. The visual feature items of selfies taken with work certificates can verify whether the staff is holding the correct certificates to work.

[0026] Optionally, the work site file data includes information of all workers in the current shift and a necessary feature set of the work site, and the necessary feature set of the work site includes visual feature items of safety facilities and visual feature items of construction equipment.

[0027] By adopting the above technical solution, the work site archive data is used to provide complete inspection comparison items for the features obtained by the management center server through video analysis during video inspection, so that more detailed video inspection of work site facilities can be carried out.

[0028] Optionally, the specific steps of penetrating video inspection in step 3 are:

[0029] The management center server initiates an interaction request signal to the smart terminal carried by the staff member to be penetrating video inspection. The smart terminal rings after receiving the interaction request signal. The video inspection staff member operates the smart terminal to accept the interaction request. The management center server and the smart terminal are interactively connected, and the video inspection step instructions are displayed on the smart terminal. The staff member to be penetrating video inspection uses the front and rear cameras of the smart terminal to shoot multiple inspection videos to be inspected according to the video inspection step instructions, forms the inspection video data to be inspected and uploads it to the management center server.

[0030] Optionally, the steps of the visual feature algorithm in step 4 are:

[0031] De-jitter the inspection video data, correct the light changes in the video, adjust the video resolution to the set size, and use the deep learning model for real-time target detection to distinguish between people and equipment;

[0032] Character feature extraction: Extract multiple video frames containing work ID and facial features, and select the frame with the most correct angle as the visual feature comparison item for the selfie of holding the work ID;

[0033] Extracting multiple video frames containing tools as necessary tool visual feature comparison items;

[0034] Equipment feature extraction: Extract multiple video frames containing the complete front of the equipment, and select the multiple video frames with the most positive angles as the construction equipment visual feature comparison items;

[0035] Safety facility feature extraction: Extract multiple video frames containing complete safety facilities, and select multiple video frames with the most positive angles as safety facility visual feature comparison items.

[0036] In step 4, the steps of comparing based on the visual similarity comparison algorithm are:

[0037] Select a similarity measurement method to calculate the matching of each feature item with the feature items in the project set. The similarity measurement method is one of Euclidean distance, cosine similarity or Jaccard index;

[0038] Each feature item is matched with the feature items in the necessary feature set, and the similarity is calculated. According to the similarity measurement result, a threshold is set to determine whether the feature item is matched successfully. If the similarity is lower than the threshold, it is determined that the feature item is missing in the necessary feature set.

[0039] By adopting the above technical solution, the visual feature comparison items of the self-portrait of the handheld work certificate and the necessary tool visual feature comparison items of the inspection video data to be inspected are first obtained through the visual feature algorithm to form multiple necessary feature items of the operators, and the visual feature comparison items of the construction equipment and the safety facility visual feature comparison items of the inspection video data to be inspected are obtained through the visual feature algorithm to form multiple necessary feature items of the corresponding work points.

[0040] Then, multiple necessary feature items of operators are compared with the necessary feature set of operators based on the visual similarity comparison algorithm, and multiple necessary feature items of corresponding work points are compared with the necessary feature set of work points. Any items that are completely missing or the comparison fails are regarded as missing items. Finally, the results of missing items of necessary features of operators and missing items of necessary features of work points are output. If the comparison is successful, the result of the end of the inspection is output, thereby efficiently realizing penetrating automatic visual inspection with high inspection efficiency and complete inspection content.

[0041] Optionally, in step 3, inspection numbers of 2%-5% of the staff members are generated every thirty minutes based on a random number algorithm.

[0042] By adopting the above technical solution, the inspection interval and ratio are selected according to the number of staff. Generally, when the number of staff is less than 500, the inspection numbers of 2%-5% of the staff can be generated every 30 minutes based on the random number algorithm, which can achieve better inspection results.

[0043] Optionally, a random number algorithm is used in Python to generate random numbers using the random module.

[0044] A railway construction project operator penetration management system includes a management center server, multiple intelligent terminals and a project area wireless network module. The multiple intelligent terminals are wirelessly connected to the management center server through the project area wireless network module respectively. The management center server is pre-installed with a management program designed using a railway construction project operator penetration management method. The management program is run to output the missing items of necessary characteristics of the operators and the missing items of necessary characteristics of the work sites.

[0045] Optionally, it can be used for penetrating management of workers in railway construction projects.

[0046] By adopting the above technical solution, a railway construction project operator penetration management system is matched with a railway construction project operator penetration management method, and applied to the management of railway construction project operators, it can enhance the work enthusiasm of the staff, always keep them in a working state to welcome inspections, and can significantly reduce the incidence of irregular operations, improve the safety of work sites, and reduce the accident rate.

[0047] In summary, the present invention includes at least one of the following beneficial technical effects:

[0048] The present invention can provide a penetrating management method and application for railway construction project operators. With the penetrating management mode, the management center server can directly perform periodic video inspections on the operators, and the inspection objects are generated based on a random number algorithm to avoid greeting inspections. The inspections are completely based on the algorithm and run autonomously. The collection of inspection content is based on the intelligent terminal carried by the operators. The management center server can achieve communication connection with the intelligent terminal and guide the operators to complete the penetrating visual inspection through the intelligent terminal. The inspection of the correctness of the on-the-job personnel, the inspection of the equipment status and the inspection of the safety facilities can be realized. The inspection content is complete, the human interference factors of the inspection are small, and the inspection efficiency is high.

[0049] When applied to the management of operating personnel in railway construction projects, it can enhance the working enthusiasm of the staff and keep them in a working state ready for inspection at all times. It can significantly reduce the incidence of irregular operations, improve the safety of work sites, and reduce the accident rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flow chart of a method for penetrating management of railway construction project workers according to the present invention;

[0051] Figure 2 It is a schematic diagram of the component connection principle of a railway construction engineering worker penetration management system of the present invention.

[0052] Description of reference numerals: 100, intelligent terminal; 101, management center server; 102, engineering area wireless network module. DETAILED DESCRIPTION

[0053] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0054] The embodiment of the present invention discloses a penetrating management method for railway construction project operators and its application.

[0055] Reference Figure 1 and Figure 2 , Embodiment 1, a method for penetrating management of railway construction project workers, comprising the following steps: Step 1, respectively archiving a plurality of workers and work sites to obtain worker file data and work site file data, the worker file data including a necessary feature set of the workers, and the work site file data including a necessary feature set of the work sites;

[0056] Step 2: Multiple operators carry the intelligent terminal 100 to work, and the multiple operators are numbered respectively;

[0057] Step 3: The management center server 101 generates inspection numbers of a set proportion of several staff members based on a random number algorithm at set intervals. The management center server 101 wirelessly communicates with the smart terminals 100 carried by several workers in the order of the inspection numbers to perform penetrating video inspections. The inspection staff holds the smart terminal 100 to record the inspection video data to be tested.

[0058] Step 4, the management center server 101 extracts a plurality of necessary feature items of operators and a plurality of necessary feature items of corresponding work points of the inspection video data to be inspected based on a visual feature algorithm;

[0059] Based on the visual similarity comparison algorithm, multiple necessary feature items of operators are compared with the necessary feature set of operators, and multiple necessary feature items of work points are compared with the necessary feature set of work points. If the comparison is incomplete, the missing items of necessary features of operators and work points are output, and warnings are issued and manual re-inspection is performed.

[0060] By adopting a penetrating management method, the management center server 101 can directly conduct periodic video inspections on the staff, and the inspection objects are generated based on a random number algorithm to avoid greeting-style inspections. The inspections are completely based on the algorithm and run autonomously. The collection of inspection content is based on the smart terminal 100 carried by the operator. The smart terminal 100 can be a customized smart terminal with video interaction function and voice interaction function, or it can be replaced by a smart phone. The management center server 101 can communicate with the smart terminal 100 and guide the staff to complete the penetrating visual inspection through the smart terminal 100. It can realize the inspection of the correctness of on-duty personnel, the inspection of equipment status and the inspection of safety facilities. The inspection content is complete, there are few human interference factors in the inspection, and the inspection efficiency is high. For those that cannot pass the autonomous visual inspection, manual review is required. The review can also be implemented based on the video call of the smart terminal 100 to realize vertical penetrating management, improve management efficiency, and reduce the incidence of safe construction.

[0061] In Example 2, the operator's file data includes related information and a set of necessary features for the operators. The related information includes the operator's real-name information, file photo information, position information, contract information, health management information, work evaluation information and current construction site information; the set of necessary features for the operators includes visual feature items of multiple necessary tools carried by the operator and visual feature items of a selfie taken with a work certificate in hand.

[0062] The operator profile data can display the relevant information of the staff in detail. The management personnel can query it on the management center server 101 with authorization. The operator's necessary feature set needs to include visual feature items of multiple necessary tools carried and visual feature items of selfies taken with work certificates. The visual feature items of multiple necessary tools carried can verify the tool carrying situation of the staff while on the job to avoid irregular operations of the staff. The visual feature items of selfies taken with work certificates can verify whether the staff is holding the correct certificates to work.

[0063] In the third embodiment, the work site file data includes information of all workers in the current shift and a necessary feature set of the work site, and the necessary feature set of the work site includes visual feature items of safety facilities and visual feature items of construction equipment.

[0064] The worksite archive data is used to provide complete inspection comparison items for the features obtained by the management center server 101 through video analysis during video inspection, so as to enable more detailed video inspection of the worksite facilities.

[0065] In Example 4, the specific steps of penetrating video inspection in step 3 are:

[0066] The management center server 101 initiates an interaction request signal to the smart terminal 100 carried by the staff member to be subjected to the penetrating video inspection. The smart terminal 100 rings after receiving the interaction request signal. The video inspection staff member operates the smart terminal 100 to accept the interaction request. The management center server 101 and the smart terminal 100 realize an interactive connection, and the video inspection step instructions are displayed on the smart terminal 100. The staff member to be subjected to the penetrating video inspection uses the front and rear cameras of the smart terminal 100 to shoot multiple inspection videos to be inspected according to the video inspection step instructions, forms the inspection video data to be inspected and uploads it to the management center server 101.

[0067] In Example 5, the steps of the visual feature algorithm in step 4 are:

[0068] De-jitter the inspection video data, correct the light changes in the video, adjust the video resolution to the set size, and use the deep learning model for real-time target detection to distinguish between people and equipment;

[0069] Character feature extraction: Extract multiple video frames containing work ID and facial features, and select the frame with the most correct angle as the visual feature comparison item for the selfie of holding the work ID;

[0070] Extracting multiple video frames containing tools as necessary tool visual feature comparison items;

[0071] Equipment feature extraction: Extract multiple video frames containing the complete front of the equipment, and select the multiple video frames with the most positive angles as the construction equipment visual feature comparison items;

[0072] Safety facility feature extraction: Extract multiple video frames containing complete safety facilities, and select multiple video frames with the most positive angles as safety facility visual feature comparison items.

[0073] In Example 6, in step 4, the step of performing comparison based on the visual similarity comparison algorithm is:

[0074] Select a similarity measurement method to calculate the matching of each feature item with the feature items in the project set. The similarity measurement method is one of Euclidean distance, cosine similarity or Jaccard index;

[0075] Each feature item is matched with the feature items in the necessary feature set, and the similarity is calculated. According to the similarity measurement result, a threshold is set to determine whether the feature item is matched successfully. If the similarity is lower than the threshold, it is determined that the feature item is missing in the necessary feature set.

[0076] First, the visual feature comparison items of the selfie of the handheld work certificate and the necessary tool visual feature comparison items of the inspection video data to be inspected are obtained through the visual feature algorithm to form multiple necessary feature items of the operators. The visual feature comparison items of the construction equipment and the safety facility visual feature comparison items of the inspection video data to be inspected are obtained through the visual feature algorithm to form multiple necessary feature items of the corresponding work points.

[0077] Then, multiple necessary feature items of operators are compared with the necessary feature set of operators based on the visual similarity comparison algorithm, and multiple necessary feature items of corresponding work points are compared with the necessary feature set of work points. Any items that are completely missing or the comparison fails are regarded as missing items. Finally, the results of missing items of necessary features of operators and missing items of necessary features of work points are output. If the comparison is successful, the result of the end of the inspection is output, thereby efficiently realizing penetrating automatic visual inspection with high inspection efficiency and complete inspection content.

[0078] The following is a code example for calculating similarity based on cosine similarity comparison between video frames:

[0079] importcv2

[0080] importnumpyasnp

[0081] fromsklearn.metrics.pairwiseimportcosine_similarity

[0082] #Load the video

[0083] video_path='path_to_video.mp4'

[0084] cap=cv2.VideoCapture(video_path)

[0085] # Check if the video is loaded successfully

[0086] ifnotcap.isOpened():

[0087] print("Error:Couldnotopenvideo.")

[0088] exit()

[0089] #Read the first frame as a reference frame

[0090] ret,reference_frame=cap.read()

[0091] ifnotret:

[0092] print("Error:Couldnotreadthefirstframe.")

[0093] exit()

[0094] #Convert to grayscale image

[0095] reference_frame_gray=cv2.cvtColor(reference_frame,cv2.COLOR_BGR2GRAY)

[0096] #Feature extraction, such as using ORB

[0097] orb = cv2.ORB_create()

[0098] kp_reference,des_reference=orb.detectAndCompute(reference_frame_gray,None)

[0099] # Initialize the similarity list

[0100] similarities=[]

[0101] #Read subsequent frames and calculate similarity

[0102] whileTrue:

[0103] ret,current_frame=cap.read()

[0104] ifnotret:

[0105] break#End of video

[0106] current_frame_gray=cv2.cvtColor(current_frame,cv2.COLOR_BGR2GR AY)

[0107] kp_current,des_current=orb.detectAndCompute(current_frame_gray,None)

[0108] # Calculate cosine similarity

[0109] ifdes_currentisnotNone:

[0110] #Use BFMatcher for feature matching

[0111] bf=cv2.BFMatcher(cv2.NORM_HAMMING,crossCheck=True)

[0112] matches=bf.match(des_reference,des_current)

[0113] # Extract the coordinates of the matching points

[0114] points1=np.float32([kp_reference[m.queryIdx].ptforminmatches]).reshape(-1,1,2)

[0115] points2=np.float32([kp_current[m.trainIdx].ptforminmatches]).reshape(-1,1,2)

[0116] #Use RANSAC method to estimate the transformation matrix

[0117] matrix,mask=cv2.findHomography(points1,points2,cv2.RANSAC,5.0)

[0118] matches_mask=mask.ravel().tolist()

[0119] # Calculate similarity

[0120] similarity=cosine_similarity(des_reference,des_current)

[0121] similarities.append(similarity)

[0122] #Release video

[0123] cap.release()

[0124] # Output similarity

[0125] fori,similarityinenumerate(similarities):

[0126] print(f"Frame{i+2}similarity:{similarity}")

[0127] The similarity value range is set to 0-1. The larger the value, the higher the similarity. The similarity threshold is set to 0.9. During the video frame comparison process, if the similarity is greater than 0.9, it is considered a successful match.

[0128] In Example 7, in step 3, inspection numbers of 2%-5% of the staff are generated every thirty minutes based on a random number algorithm.

[0129] The inspection interval and ratio are selected according to the number of staff. Usually, when the number of staff is less than 500, the inspection numbers of 2%-5% of the staff can be generated every 30 minutes based on the random number algorithm, which can achieve better inspection results.

[0130] Example 8: The random number algorithm is in Python, using the random module to generate random numbers.

[0131] The following is a random number algorithm code example:

[0132] importrandom

[0133] defgenerate_ids_by_ratio(ratios):

[0134] ids=[]

[0135] total = sum(ratios.values())

[0136] random_value=random.uniform(1,total)

[0137] current_sum=1

[0138] forcategory,ratioinratios.items():

[0139] current_sum+=ratio

[0140] ifrandom_value<=current_sum:

[0141] ids.append(category)

[0142] break

[0143] returnids

[0144] #Example Usage

[0145] ratios={"A":0.3,"B":0.2,"C":0.5}

[0146] unique_id=generate_ids_by_ratio(ratios)

[0147] print(unique_id);

[0148] where ratios is a dictionary, the key is the category, and the value is the ratio of each category. We first calculate the total ratio, and then generate a random number between 1 and the total ratio.

[0149] Embodiment 9, a railway construction project worker penetration management system, including a management center server 101, multiple smart terminals 100 and a project area wireless network module 102, multiple smart terminals 100 are wirelessly connected to the management center server 101 through the project area wireless network module 102, the management center server 101 is pre-installed with a management program designed using a railway construction project worker penetration management method, and the management program is run to output the missing items of necessary characteristics of the workers and the missing items of necessary characteristics of the work sites.

[0150] The engineering area wireless network module 102 may be a 4G / 5G wireless network.

[0151] Example 10 is applied to the penetrating management of workers in railway construction projects.

[0152] A railway construction project operator penetration management system is provided based on a railway construction project operator penetration management method. It is applied to the management of railway construction project operators, which can enhance the work enthusiasm of the staff, always keep them in a working state to welcome inspections, significantly reduce the occurrence rate of irregular operations, improve the safety of work sites, and reduce the accident rate.

[0153] A specific inspection example is used below to illustrate the principle of the facility for the railway construction project workers' penetrating management method and application thereof of the present invention:

[0154] At a certain inspection time point, the management center server 101 uses a random number algorithm to generate 5 numbers of the 250 staff members to be inspected at a ratio of 2%, which are 15, 28, 156, 182, and 213 respectively;

[0155] The inspection process for staff member No. 28 is as follows:

[0156] The management center server 101 establishes wireless communication with the smart terminal 100 of the worker No. 28, first instructing the worker No. 28 to take a selfie holding the work permit to form a first video, instructing the worker No. 28 to record a video of the necessary tool project to obtain a second video, instructing the worker No. 28 to record a video of the work site equipment to obtain a third video, instructing the worker No. 28 to record an energy-saving video of the safety facilities to obtain a fourth video, and instructing the first, second, third, and fourth videos to form the inspection video data to be tested;

[0157] The management center server 101 extracts a plurality of necessary feature items of operators and a plurality of necessary feature items of corresponding work points from the inspection video data to be inspected based on a visual feature algorithm;

[0158] Based on the visual similarity comparison algorithm, multiple necessary feature items of operators are compared with the necessary feature sets of operators, and multiple necessary feature items of work points are compared with the necessary feature sets of work points. Finally, in the comparison of the visual feature comparison items of safety facilities with the necessary feature sets of work points, it is found that the comparison of the fire hose feature items is unsuccessful. The management center server 101 activates the alarm. The management personnel pays attention to the display content of the unsuccessful comparison, and directly communicates with the staff member No. 28 through the smart terminal 100 through the management end to ask about the fire hose. The staff member No. 28 said that due to the risk of fire, he had just used it and had not returned it. After understanding the situation, the management personnel believed that the fire had not spread and had been eliminated, and urged the fire hose to be returned as soon as possible, completing the inspection.

[0159] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A penetrating management method for railway construction project operators, characterized by: The following steps are involved: Step 1, respectively archiving multiple operators and work sites to obtain operator archival data and work site archival data, wherein the operator archival data includes an operator necessary feature set, and the work site archival data includes a work site necessary feature set; Step 2: Multiple operators carry smart terminals to work, and the operators are numbered respectively; Step 3: The management center server generates inspection numbers of a set proportion of several staff members based on a random number algorithm at set intervals. The management center server wirelessly communicates with the smart terminals carried by several operators in the order of the inspection numbers to perform penetrating video inspections. The inspection staff members hold smart terminals to record the inspection video data to be tested. Step 4: The management center server extracts a plurality of necessary feature items of operators and a plurality of necessary feature items of corresponding work points from the inspection video data to be inspected based on a visual feature algorithm; Based on the visual similarity comparison algorithm, multiple necessary feature items of operators are compared with the necessary feature set of operators, and multiple necessary feature items of work points are compared with the necessary feature set of work points. If the comparison is incomplete, the missing items of necessary features of operators and work points are output, and warnings are issued and manual re-inspection is performed.

2. A railway construction project worker penetration management method according to claim 1, characterized in that: The operator's archival data includes related information and a set of necessary features for operators. The related information includes the operator's real-name information, archival photo information, position information, contract information, health management information, work evaluation information and current construction site information; the set of necessary features for operators includes visual feature items of multiple necessary tools carried by the operator and visual feature items of selfies taken with work certificates held in their hands.

3. A railway construction project worker penetration management method according to claim 2, characterized in that: The work site file data includes the information of all staff members in the current shift and the necessary feature set of the work site. The necessary feature set of the work site includes the visual feature items of safety facilities and the visual feature items of construction equipment.

4. A method for penetrating management of railway construction workers according to claim 3, characterized in that: The specific steps of penetrating video inspection in step 3 are: The management center server initiates an interaction request signal to the smart terminal carried by the staff member to be penetrating video inspection. The smart terminal rings after receiving the interaction request signal. The video inspection staff member operates the smart terminal to accept the interaction request. The management center server and the smart terminal are interactively connected, and the video inspection step instructions are displayed on the smart terminal. The staff member to be penetrating video inspection uses the front and rear cameras of the smart terminal to shoot multiple inspection videos to be inspected according to the video inspection step instructions, forms the inspection video data to be inspected and uploads it to the management center server.

5. A railway construction project worker penetration management method according to claim 4, characterized in that: The steps of the visual feature algorithm in step 4 are: De-jitter the inspection video data, correct the light changes in the video, adjust the video resolution to the set size, and use the deep learning model for real-time target detection to distinguish between people and equipment; Character feature extraction: Extract multiple video frames containing work ID and facial features, and select the frame with the most correct angle as the visual feature comparison item for the selfie of holding the work ID; Extracting multiple video frames containing tools as necessary tool visual feature comparison items; Equipment feature extraction: Extract multiple video frames containing the complete front of the equipment, and select the multiple video frames with the most positive angles as the construction equipment visual feature comparison items; Safety facility feature extraction: Extract multiple video frames containing complete safety facilities, and select multiple video frames with the most positive angles as safety facility visual feature comparison items.

6. A method for penetrating management of railway construction workers according to claim 5, characterized in that: In step 4, the steps of comparing based on the visual similarity comparison algorithm are: Select a similarity measurement method to calculate the matching of each feature item with the feature items in the project set. The similarity measurement method is one of Euclidean distance, cosine similarity or Jaccard index; Each feature item is matched with the feature items in the necessary feature set, and the similarity is calculated. According to the similarity measurement result, a threshold is set to determine whether the feature item is matched successfully. If the similarity is lower than the threshold, it is determined that the feature item is missing in the necessary feature set.

7. A method for penetrating management of railway construction workers according to claim 6, characterized in that: In step 3, every thirty minutes, inspection numbers of 2%-5% of the staff are generated based on a random number algorithm.

8. A railway construction project worker penetration management method according to claim 7, characterized in that: The random number algorithm is in Python, using the random module to generate random numbers.

9. A railway construction project operator penetration management system, characterized in that: It includes a management center server, multiple smart terminals and a project area wireless network module. The multiple smart terminals are wirelessly connected to the management center server through the project area wireless network module respectively. The management center server is pre-installed with a management program designed using a penetrating management method for railway construction project workers as described in any one of claims 1 to 8. Running the management program outputs missing items of necessary characteristics of workers and missing items of necessary characteristics of work sites.

10. A railway construction project worker penetration management system according to claim 9, characterized in that: Applied to the penetrating management of workers in railway construction projects.

Citation Information

Patent Citations

  • A management system of a mobile intelligent terminal special for railways

    CN109559051A

  • Substation inspection system and control method thereof

    CN114301129A

  • Intelligent safety management platform and method for building engineering construction process

    CN114926155A

  • Robot random inspection method based on image recognition

    CN114995381A

  • Smart supervision and construction project management system for efficient construction management

    KR102489400B1