Construction real-name order receiving system and compliance auditing mechanism

Abstract

The invention belongs to the technical field of information management, and discloses a construction real-name order receiving system and a compliance auditing mechanism. Collecting data and cleaning the data to obtain accurate text registration information and high-definition data image data; performing field identification to obtain a certificate field data set; integrating the certificate field data set and the accurate text registration information into a structured data set; screening valid data to form a valid data set; performing construction personnel credit state evaluation, and marking an evaluation result to obtain a user marking data set; task conflict detection is carried out, and a corrected data set is output; insurance information verification is carried out, an insurance state label is output to mark the correction data set, and an information improvement data set is obtained; the information perfection data set is sent to a preset auditing terminal for overall judgment, and an auditing filing result is generated and stored in a preset auditing result database; and the auditing capability in the aspects of information authenticity and task rationality is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information management, more particularly, the present application relates to a construction real-name system and a compliance audit mechanism. BACKGROUND

[0002] In the field of industrial construction, labor management is an important means for ensuring construction safety compliance and improving engineering management efficiency. With the development of information construction platform, more and more construction units gradually establish personnel real-name system and require construction personnel to submit relevant identity documents and certificates and other materials, which can participate in construction tasks after passing the audit. This order receiving mechanism can effectively improve the efficiency of site management in practical application, but the existing audit mechanism still has many technical shortcomings in actual operation. Regarding the information provided by the construction personnel, the relevant certificates may be manually filled in by the user or the issuing unit, which may lead to the risk of misfilling or tampering. Therefore, it is necessary to more accurately identify the field content information in the image type data. However, the existing audit mechanism is not accurate enough in identifying the text in the image, which leads to distorted audit results. At the same time, the existing audit mechanism may ignore the validity of the relevant certificates, only determine whether the user has uploaded the relevant information, but does not determine whether the certificate is valid, which affects the actual audit and allows some users to still receive orders in the case of invalid or illegal certificates. On the other hand, the existing audit mechanism often lacks accurate identification of the user's historical construction records and related default records, which leads to some users being listed in the blacklist due to malicious off-duty or poor work attitude, and still receiving new orders, thereby causing construction risk hazards. In the existing society, some construction personnel may repeat the order receiving process in different construction areas to create false work, which leads to distorted project work time statistics and affects the platform's salary settlement. The existing audit mechanism often ignores this aspect, which affects the fairness and transparency of order receiving. In addition, the existing audit mechanism often only performs the most basic examination of the insurance information of the construction personnel, and does not perform in-depth verification of the effectiveness and specific protection content of the insurance, which leads to the construction personnel may work outside the insurance protection range, thereby damaging the legitimate rights and interests of the construction personnel.

[0003] Therefore, the present application provides a construction real-name system and a compliance audit mechanism to solve the above problems. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a construction real-name system and a compliance audit mechanism, comprising: S1. Collecting text registration information and image data of construction personnel, and performing data cleaning to obtain accurate text registration information and high-definition image data; S2. Field recognition is performed on the high-definition document image data to obtain a document field data set; and the document field data set is integrated with accurate text registration information to form a structured document data set; S3. Validity judgment is performed on the structured document data set to filter effective data to form an effective document data set; S4. Construction personnel credit status evaluation is performed on the effective document data set to mark the evaluation result to obtain a user marked document data set; task conflict detection is performed on the user marked document data set to output a corrected document data set; S5. Insurance information verification is performed on the corrected document data set to output an insurance status label to mark the corrected document data set to obtain an information perfect data set; S6. The information perfect data set is sent to a preset audit terminal for overall judgment to generate an audit record result and store the audit record result in a preset audit result database.

[0005] Further, the field recognition manner comprises: A field area containing text in each high-definition document image data is recognized, a preset size grid is constructed to divide the field area into image blocks to obtain a field area image block set; each column of pixels in any one field area image block is traversed, and the number of pixels higher than a preset gray threshold in each column of pixels is counted as a vertical projection array; if the vertical projection arrays with a continuous occurrence are all greater than a number threshold, the number of the continuous vertical projection arrays is taken as a continuous pixel bandwidth; and each continuous pixel bandwidth is taken as a segmentation basis to divide the field area into candidate character image units; Each candidate character image unit is subjected to gray normalization processing to construct a standard character image matrix; each standard character image matrix is sorted based on a center image coordinate to obtain an initial field string; the position coordinates of the initial field string are matched with a preset document information distribution template to recognize the character type of each character in the initial field string; the region where each character in the initial field string is located is subjected to contour enhancement, the enhanced character contour is matched with a dictionary database based on the character type to output actual field text, and all characters in the initial field string are converted into actual field text to obtain the document field data set.

[0006] Further, the validity judgment manner comprises: The document time attribute fields in the structured data set are extracted based on character types to form a document time field set; the document time field set is grouped based on the original positions of the document time attribute fields to obtain a document time field group; symbols in each document time field group are recognized and replaced with a date separator to output a uniform time format string; the uniform time format string is divided into an issuance time field and an expiration time field based on a Chinese prompt word, and a current date is used as a time reference to compare a time interval between the time reference and the issuance time field and a time interval between the time reference and the expiration time field; if the issuance time field is later than the time reference or the expiration time is earlier than the time reference, the corresponding uniform time format string is determined as an invalid string; issuance information corresponding to a string other than the invalid string is recognized, and it is determined whether an issuance unit in the issuance information exists and whether an issuance number is normal by querying a preset external database; if any one of the issuance unit not existing or the issuance number being unable to be queried occurs, the corresponding string is also determined as an invalid string; user-related data corresponding to the invalid string is removed from the structured data set and added to an invalid log, and the invalid log and the remaining valid data are integrated into a valid data set.

[0007] Further, the method for evaluating the credit status of the construction personnel includes: A unique identity code of each user in the valid data set is recognized, and a historical construction behavior data set of the corresponding user is obtained based on the unique identity code; each historical construction record in the historical construction behavior data set is converted into a vector form and arranged in a time sequence as a construction behavior vector sequence of the corresponding user; principal component analysis is performed on the construction behavior vector sequence to output a principal component behavior vector and construct a change trend curve based on the principal component behavior vector; Known high-risk behavior information is obtained, and a risk behavior change curve is constructed based on the known high-risk behavior information; a behavior similarity function is constructed to calculate the behavior similarity between the change trend curve and the risk behavior change curve; if the behavior similarity is higher than a preset risk behavior similarity threshold, each dimension parameter in the construction behavior vector sequence corresponding to the change trend curve is compared with a preset standard behavior degree, a dimension parameter with a parameter value higher than an abnormal degree threshold is recognized, and an abnormal label is added to the corresponding construction behavior vector; an abnormal index is obtained by calculating the abnormal label density in the construction behavior vector sequence, the abnormal index and the behavior similarity are weighted and summed to obtain a risk evaluation factor, and the risk evaluation factor is matched with a historical risk evaluation level template to output an evaluation result of the corresponding user.

[0008] Further, the method for obtaining the historical construction behavior data set of the corresponding user includes: By matching the unique identity code with the historical record of the corresponding user, the contractor unit information field, the construction task start and end time field, the construction task completion condition field, the sign-in attendance rate information and the construction project feedback information are extracted to form a historical construction behavior data set.

[0009] Further, the task conflict detection method comprises: Based on the unique identity code, the execution task information of the corresponding user in the user mark data set is identified, and the execution task information includes the task time start and end field and the engineering geographic location field. The task time axis of the corresponding user is constructed based on the execution task information. The start and end time interval corresponding to each task record in the task time axis is determined, and it is identified whether there is an intersection of the start and end time interval of any two adjacent task records. If there is, the intersection time length is counted. The ground distance of the engineering geographic location of the adjacent task record is calculated. If the intersection time length is higher than the preset overlap time threshold or the ground distance is greater than the reasonable distance threshold, it is determined that the task time axis of the user has task conflict. The two or more task records with task conflict are revised to obtain state updated task records. The execution task information of the corresponding user in the user mark data set is covered by using the state updated task records to obtain a revised material data set.

[0010] Further, the task record revision method comprises: The field of the task record with task conflict is adjusted, the corresponding task state field is adjusted to a frozen state field, and the task record is marked as a non-compliant task. The task working hours field of the non-compliant task is extracted and cleared. The order receiving platform number of the non-compliant task is identified. The adjusted non-compliant task related information and the order receiving platform number are integrated into a conflict processing log. The unique identity code and the real name of the user are matched with the conflict processing log to obtain state updated task records.

[0011] Further, the insurance information verification method comprises: Obtain the insurance record of each user, including the policy number, the insurance state label and the insurance guarantee item entry. Identify the construction task type and the risk level identifier taken by the corresponding user in the revised material data set, and match with the insurance guarantee item entry. If the construction task type is not included in the insurance guarantee item entry or the insurance guarantee level does not match the risk level identifier, the corresponding insurance record is marked as not applicable clause. The insurance state label is detected in real time. If the insurance record is outside the insurance effective time interval, the insurance record is also marked as not applicable clause. Based on the specific invalidation condition of the not applicable clause, the insurance state field is generated, and the insurance state field and the corresponding policy number are integrated into the insurance state label.

[0012] Furthermore, the method for performing the overall discrimination includes: Extract the document fields, credit status assessment results, and insurance status labels from the information completion dataset separately. Then, compare them with the failure log and conflict handling log with the preset compliance review template. If all conditions of the preset compliance review template are met, the corresponding user is marked as an order-accepting user; otherwise, the user is marked as an order-unaccepting user. Integrate the relevant data and order-accepting status labels of the corresponding user from the information completion dataset into the review and filing results.

[0013] A construction order acceptance compliance review system, used to implement a construction order acceptance compliance review mechanism, characterized in that it includes: The data acquisition module is used to collect text registration information and data image data of construction personnel, and to perform data cleaning to obtain accurate text registration information and high-definition data image data. The field recognition module is used to perform field recognition on high-definition data image data to obtain a document field dataset; the document field dataset is then integrated with precise text registration information to form a structured data dataset. The validity assessment module is used to assess the validity of structured data datasets and filter valid data to form a valid data dataset. The information detection module is used to assess the credit status of construction workers on the valid data dataset, and to label the assessment results to obtain a user-labeled data dataset; it also performs task conflict detection on the user-labeled data dataset and outputs a corrected data dataset. The data correction module is used to verify insurance information on the corrected data dataset, output insurance status labels to mark the corrected data dataset, and obtain the information-complete dataset. The review and judgment module is used to send the complete information dataset to the preset review terminal for overall judgment, generate the review and filing results and store them in the preset review result database; the modules are connected to each other via wired and / or wireless means.

[0014] The technical effects and advantages of the construction real-name order acceptance system and compliance review mechanism of this invention are as follows: By collecting user-submitted data and performing character recognition on image data, validating document and certificate information, assessing credit status based on historical construction records, detecting task conflicts, and verifying insurance information, a comprehensive data set is constructed and a holistic review process is conducted to achieve a compliance review mechanism for real-name order acceptance in construction projects. Compared to existing experience, this mechanism reduces misjudgments caused by information tampering or errors by accurately identifying the actual text content of key fields such as validity periods or qualification numbers. Furthermore, by combining document validity assessments, it further determines whether the submitted documents are valid, improving review accuracy. By querying users' historical construction tasks, this mechanism... Using users' historical task performance as a criterion, credit status is transformed into quantifiable indicators. Users exhibiting high-risk behavior are flagged, and historical behavior serves as one of the review criteria. Simultaneously, a timeline is constructed to determine if users have conflicting task execution situations, promptly identifying and freezing abnormal task records to prevent waste of platform resources. Insurance information is matched with specific construction task information to determine insurance validity, preventing high-risk workers from performing tasks without protection. In summary, this real-name construction order acceptance compliance review mechanism improves the review capability regarding information authenticity and task rationality, reduces the incidence of misjudgments, and ensures order acceptance security and credit transparency for construction workers. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of a construction real-name order acceptance compliance review mechanism according to the present invention; Figure 2 This is a schematic diagram of a construction real-name order acceptance compliance review system according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1 Please see Figure 1 As shown in this embodiment, a compliance review mechanism for real-name order acceptance in construction projects includes: S1. Collect text registration information and data image data of construction personnel, and perform data cleaning to obtain accurate text registration information and high-definition data image data; S2. Perform field recognition on the high-definition data image to obtain the document field dataset; integrate the document field dataset with the precise text registration information into a structured data dataset; S3. Determine the validity of the structured data dataset and select valid data to form a valid data dataset; S4. Evaluate the credit status of construction workers on the valid data dataset, and label the evaluation results to obtain the user-labeled data dataset; perform task conflict detection on the user-labeled data dataset and output the corrected data dataset; S5. Verify the insurance information on the corrected data dataset, output insurance status labels to mark the corrected data dataset, and obtain the information-completed dataset; S6. Send the complete dataset to the preset review terminal for overall judgment, generate the review and filing results, and store them in the preset review result database.

[0018] In this embodiment, the text registration information of construction workers is the document data submitted by the construction workers themselves, including personal information such as name, ID number, contact information and job title; the document image data includes image data of work documents such as qualification certificate images, vocational skill certificate images and health certificate images; by filling in missing fields in the text registration information of construction workers and filtering and denoising the document image data, data cleaning is achieved to obtain higher quality and more accurate text registration information and high-definition document image data.

[0019] Methods for field recognition include: The text-containing field regions in each high-definition image data are identified. A grid of preset size is constructed to divide the field regions into image blocks, resulting in a set of field region image blocks. First, the image regions containing continuous text image features in the high-definition image data are roughly identified as field regions. These regions are rectangular regions constructed based on the outermost pixel blocks belonging to the text. By setting the grid size and using several grids of the same size, the field region can be divided into several image blocks. The set of these image blocks is the field region image block set.

[0020] Iterate through each column of pixels in any field region image block, and count the number of pixels in each column that are higher than a preset grayscale threshold as a vertical projection array. In this embodiment, scan the pixel values ​​of pixels in each image block column by column, set the standard grayscale value of the character as the preset grayscale threshold, count the number of pixels in each column whose pixel value is higher than the preset grayscale threshold, and store the number of pixels in each column as the vertical projection array of the corresponding column of pixels. Since the brightness of the character spacing area is higher and the color is lighter, the more pixels that meet the above conditions, the more likely the column of pixels is to belong to the character spacing area.

[0021] If consecutive vertical projection arrays with projection values ​​all greater than the quantity threshold appear, the number of consecutive vertical projection arrays is taken as the width of the continuous pixel band. The projection value is the number of pixels in the corresponding column. The quantity threshold is set based on historical character judgment experience. If consecutive vertical projection arrays exceeding the quantity threshold appear, it means that the area formed by the corresponding consecutive columns is a bright column band, which meets the character spacing condition. Therefore, the number of vertical projection arrays is taken as the width of the continuous pixel band, which satisfies the character segmentation standard and provides a basis for subsequent character segmentation.

[0022] Based on the width of each consecutive pixel band as the segmentation criterion, the field region is divided into candidate character image units. The region where each consecutive pixel band width is located is used as the character segmentation interval, and the entire field region is divided into several independent candidate character image units, with each candidate character image unit corresponding to one character.

[0023] For each candidate character image unit, grayscale normalization is performed to construct a standard character image matrix. In this matrix, the pixel values ​​of each pixel block in each candidate character image unit are normalized. The standard character image matrix is ​​constructed based on the pixel block distribution of the candidate character image unit. Each element is the adjusted pixel value in the candidate character image unit, and the element with no pixel block has a value of 0. The row and column dimensions of this matrix can exactly contain the corresponding complete single candidate character image unit.

[0024] The initial field string is obtained by sorting the center image coordinates of each standard character image matrix. The position coordinates of the center position of each standard character image matrix in the original candidate character image unit are identified. All standard character image matrices are sorted according to the direction in which the coordinate values ​​gradually increase, and the initial field string is constructed.

[0025] The position coordinates of the initial field string are matched with the preset document information distribution template to identify the character type of each character in the initial field string. Based on the position information of the initial field string in the complete high-definition data image, it is matched within the entire image area. Furthermore, the preset document information distribution template of the corresponding document is used as the matching basis to determine the character type of the characters in the initial field string, such as date type and encoding type, thereby enhancing the semantic interpretability of the identified characters.

[0026] Contour enhancement is performed on the region containing each character in the initial field string. Based on the character type, the enhanced character contours are matched with the dictionary database to output the actual field text. All characters in the initial field string are converted into actual field text to obtain the document field dataset. In this embodiment, an edge detection algorithm is used to enhance the contour of the region containing each character to improve the clarity of the character edge contours. The enhanced character contours are matched with the existing dictionary database to determine the actual precise field text. The same processing is performed on all initial field strings to obtain the document field dataset composed of document information-related text.

[0027] Methods for determining validity include: Based on the character type, the document time attribute field is extracted from the structured data dataset to form a document time field set. The document time attribute field refers to the field used to represent time information in the document-related data. In this embodiment, it refers to the field information containing year, month and day.

[0028] The document time field set is grouped based on the original position of the document time attribute field to obtain document time field groups. The document time field groups are divided into different groups based on the position of the document time attribute field in the original document. Each position corresponds to a document time field group, providing a data foundation for subsequent judgment.

[0029] Identify and replace the symbols in each document time field group with date separators, and output a uniform time format string. In order to prevent garbled characters or illegal characters, the separators of year, month and day are all replaced with "-" or " / ", which are converted into a uniform date display format, and a uniform time format string is obtained.

[0030] The uniform time format string is divided into an issuance time field and an expiration time field based on Chinese prompt words. Using the current date as the time base, the time interval between this time base and the issuance time field and the time interval between this time base and the expiration time field are compared respectively. The Chinese prompt words refer to prompt words such as "issuance date" or "effective date" that appear before the uniform time format string, and are used as the basis for subsequent judgment. In this embodiment, the Chinese prompt word is "issuance date". The first part of the field in the uniform time format string is determined as the issuance time field, and the second part of the field is determined as the expiration time field. The current date is compared with the issuance time field and the expiration time field respectively to calculate the time interval.

[0031] If the issuance time field is later than the time base, or the expiration time is earlier than the time base, the corresponding unified time format string will be determined as an invalid string. If the issuance time is later than the current time base, it means that the document has not been effective. If the expiration time is earlier than the time base, it means that the document has expired. Either of the above situations can determine the unified time format string as an invalid string.

[0032] The system identifies the issuance information corresponding to strings other than invalid strings. It queries a pre-set external database to determine if the issuing authority exists and if the issuance number is valid. If either the issuing authority is non-existent or the issuance number cannot be found, the corresponding string is also considered invalid. For strings other than invalid strings, the system queries issuance information, including information such as the issuing authority name and issuance number. It checks the validity of the issuance information for these strings in a pre-set external database; if invalid, the corresponding string is also considered invalid. This indirectly enhances the credibility of the data and the accuracy of the review process.

[0033] The user-related data corresponding to the invalid string is removed from the structured data dataset and an invalidity log is added. The invalidity log is then integrated with the remaining valid data to form a valid data dataset. Since some user identification documents are invalid, the user-related data corresponding to the invalid string is removed. At the same time, the field content of the invalid string, the reason for removal, and the corresponding user information are combined into an invalidity log. The invalidity log is then integrated with the remaining valid data in the structured data dataset to form a valid data dataset. The invalidity log does not participate in data processing; it is only used as tag information to reflect the relevant users who are removed during the validity judgment process.

[0034] Methods for assessing the credit status of construction workers include: The unique identification code of each user in the valid data dataset is identified, and the historical construction behavior dataset of the corresponding user is obtained based on the unique identification code. In this embodiment, the unique identification code of the user refers to the ID card number provided by the user. Using the user's ID card number as an index, the historical records bound to the user's ID card number are extracted from the historical database. By establishing a strong correlation between the unique identification code and the historical construction records, the situation of switching identities to circumvent the review is prevented.

[0035] The historical construction behavior dataset is transformed into vector form for each historical construction record and arranged chronologically to form a sequence of construction behavior vectors for the corresponding user. This is achieved by concatenating the following data: contractor information, construction task start and end time, construction task completion status, attendance rate, and project feedback information. The contractor information field uses a code representing the contractor's name; the construction task start and end time field uses a specific time interval value; the construction task completion status field uses 1 for completion, 0 for withdrawal, and 2 for user removal; the attendance rate is a numerical value; and the project feedback information is represented by a score associated with the project's specific evaluation criteria. The construction behavior vector sequence is obtained by sorting the historical construction records according to their timestamps.

[0036] Principal component analysis is performed on the construction behavior vector sequence to output principal component behavior vectors and construct a trend curve based on these principal component behavior vectors. In this embodiment, the principal component analysis algorithm is used to perform principal component analysis on the user's construction behavior vector sequence. The principal component behavior vectors that best reflect the behavior characteristics are extracted through dimensionality reduction analysis, and then a trend curve is constructed according to the numerical changes of these principal component behavior vectors over time.

[0037] Information on known high-risk behaviors is obtained, and risk behavior change curves are constructed based on this information. The known high-risk behavior information refers to high-risk behavior templates constructed based on historical behavior data samples of known high-risk construction personnel, including information such as breach of contract, absence from duty, high project termination rate, and poor project completion. The known high-risk behavior information is transformed into risk behavior change curves using unified vector modeling and principal component analysis algorithms, which serve as a reference for determining risk behavior profiles.

[0038] A behavioral similarity function is constructed to calculate the behavioral similarity between the trend curve and the wind direction behavior curve. The formula for calculating the behavioral similarity function is as follows: ;in, This represents the sequence corresponding to the principal component behavior vector; A vector sequence composed of principal components representing known high-risk behavioral information; This indicates the similarity in behavior between the trend curve and the wind direction behavior curve. This represents the number of vectors in the sequence of principal components of the two curves. and They are the same length.

[0039] If the behavioral similarity is higher than the preset risk behavior similarity threshold, the parameters of each dimension in the construction behavior vector sequence corresponding to the trend curve are compared with the preset standard behavior level. Dimension parameters with values ​​higher than the anomaly threshold are identified, and anomaly labels are added to the corresponding construction behavior vectors. The preset risk behavior similarity threshold is set based on historical judgment experience. If the behavioral similarity is higher than this threshold, it indicates that the current user's behavioral characteristics are similar to those of known high-risk behavioral information. At this time, the specific parameter values, labels, or numbers of each dimension in the construction behavior vector sequence are compared with the preset standard behavior level in the known construction project standards. If data in a dimension exceeds the anomaly threshold in the construction project standards, anomaly labels are added to the corresponding construction behavior vectors, such as high absenteeism rate and abnormal construction termination labels.

[0040] An anomaly index is obtained by calculating the anomaly label density in the construction behavior vector sequence. The anomaly index and behavior similarity are weighted and summed to obtain a risk assessment factor. The risk assessment factor is matched with a historical risk assessment level template to output the corresponding user's assessment result. The anomaly label density is obtained by dividing the number of anomaly labels by the total number of vectors in the construction behavior vector sequence, and this density is used as the anomaly index. The weights of the anomaly index and behavior similarity are set according to historical judgment experience to comprehensively consider both dimensions. The risk assessment factor is obtained by weighted summation. The risk assessment factor is matched with the interval in the historical risk assessment level template designed based on historical judgment experience to output the corresponding user's credit rating as the assessment result.

[0041] The methods for obtaining the historical construction behavior dataset of the corresponding user include: By matching the unique identification code with the corresponding user's historical records, the following fields are extracted: contractor information, construction task start and end time, construction task completion status, attendance rate, and construction project feedback information. These fields and information are used to construct the historical construction behavior dataset, which serves as the data foundation for assessing the credit status of construction personnel.

[0042] Methods for performing task conflict detection include: Based on the unique identification code, the user tag data is used to identify the execution task information of the corresponding user in the user tag data set. The execution task information includes the task start and end time field and the project geographical location field. By matching the unique identification code and the data of the corresponding user in the user tag data set, the execution task information of the user is extracted separately, including the task start and end time field and the project geographical location field for each construction task performed by the user, as the data basis for subsequent operations.

[0043] Based on the task execution information, a task timeline for the corresponding user is constructed. This is achieved by constructing a timeline-type data structure and filling the task start and end time fields and project geographical location fields corresponding to each task execution information into the timeline.

[0044] Traverse the task timeline to determine the start and end time intervals corresponding to each task record, identify whether there is an intersection of start and end time intervals between any two adjacent task records, and if so, count the length of the intersection time. In this process, by traversing the task timeline, determine whether there is an overlap between the start and end time intervals of any two adjacent task records. If so, further count the length of the overlapping part as one of the subsequent judgment conditions.

[0045] The surface distance between the engineering geographical locations of adjacent task records is calculated. Based on the intersection time length and surface distance, it is determined whether there is a task conflict on the user's task timeline. The surface distance between the engineering geographical locations corresponding to any two adjacent task records is used as one of the subsequent judgment conditions. It should be noted that the surface distance does not refer to the displacement distance, but to the shortest actual path length between two locations determined by the GPS positioning system. If the surface distance is far and there is an intersection time length, a task conflict is considered to exist. If the intersection time length is higher than a certain reasonable time overlap threshold, it is also determined to be a task conflict.

[0046] Revise two or more task records with task conflicts to obtain status-updated task records. Use the status-updated task records to overwrite the execution task information of the corresponding user in the user-labeled dataset to obtain the corrected data dataset. The status-updated task records obtained by revising the task records with task conflicts are used to replace the relevant data of the original execution task information. This ensures that the task records in the processed dataset are all task records that have been checked and revised, which is the corrected data dataset.

[0047] Methods for revising task records include: For task records with task conflicts, adjust the fields by changing the corresponding task status field to a frozen status field and marking the task record as a non-compliant task. Specifically, when a task record with a task conflict is identified, the status field of the corresponding task in the task record is changed to a frozen status field, indicating that the task has logically lost its eligibility to be executed, and the entire task record is marked as a non-compliant task, indicating that the data involved in the corresponding user's task is not compliant.

[0048] Extract the task time field of the non-compliant task and clear it to zero. This is done by extracting the task time field of the corresponding non-compliant task and setting its value to 0 to prevent non-compliant tasks from still being included in the settlement process.

[0049] Identify the order-taking platform number of the non-compliant task, integrate the adjusted non-compliant task information with the order-taking platform number into a conflict resolution log, match the user's unique identification code and real name with the conflict resolution log to obtain the status update task record. The conflict resolution log is constructed by combining the relevant information of the frozen and cleared non-compliant tasks with the order-taking platform number corresponding to the non-compliant task, and then matching it with the user's unique identification code and real name as an index to generate a complete status update task record.

[0050] The methods for verifying insurance information include: The system retrieves each user's insurance records, including policy number, insurance status tags, and insurance coverage entries. Insurance records submitted by a specific user are retrieved using a unique identifier as an index. Insurance status tags include fields such as "activated," "inactive," "not effective," and "lapsed," reflecting the current activation or effectiveness of the insurance. Insurance coverage entries refer to the set of fields in a specific policy clause, expressing information related to the insurance function, such as applicable occupation information, coverage amount, and specific coverage details.

[0051] The system identifies the type and risk level of the construction task accepted by the user in the revised data dataset and matches it with the insurance coverage entries. If the construction task type is not included in the insurance coverage entries or the insurance coverage level does not match the risk level, the corresponding insurance record is marked as an inapplicable clause. Specifically, the system queries an external database to determine the task type and risk level of the construction task accepted by the user in the revised data dataset and matches it with the insurance coverage entries to determine whether the task type matches the type of work covered by the insurance and whether the insurance coverage level can cover the risk level of the current construction task. If the type of work corresponding to the construction task is not recorded in the insurance coverage entries, or if the coverage level does not match the risk level of the task, the corresponding insurance record is marked as an inapplicable clause.

[0052] The system monitors insurance status tags in real time. If an insurance record is outside the effective period, it is marked as an inapplicable clause. If a user's current insurance record is no longer within the effective period and the insurance status tag has expired, it is considered that there is no insurance coverage at this time, and the corresponding user's insurance record is also marked as an inapplicable clause.

[0053] An insurance status field is generated based on the specific invalidation of inapplicable terms. This insurance status field is then integrated with the corresponding policy number to form an insurance status label. The insurance status field reflects the user's specific insurance status. For insurance records with inapplicable terms, fields indicating the reason for invalidation, such as "Insurance Invalid" or "Coverage Mismatch," are used as the insurance status field. For normal insurance records, the "Insurance Status Normal" field is used as the insurance status field. The policy number is used as an index, and the insurance status field and the policy number are integrated to form an insurance status label, which more clearly displays the user's insurance status.

[0054] Methods for overall discrimination include: Information is extracted to improve the dataset's document field, credit status assessment result, and insurance status label. This data is then compared with the default logs and conflict resolution logs against a preset compliance review template. If all conditions of the preset compliance review template are met, the corresponding user is marked as available for order taking; otherwise, they are marked as unavailable. The document field is a string containing the document number identified during field recognition. The credit status assessment result refers to the user's credit rating obtained during the construction worker credit status assessment. The process involves matching the document field, assessment result, insurance status label, default logs, and conflict resolution logs against a preset compliance review template based on historical review experience. The specific values ​​or fields of each dimension are compared with the preset compliance review template. If the review conditions of the preset compliance review template are met, the corresponding user is marked as available for order taking; otherwise, they are marked as unavailable.

[0055] The relevant data of the corresponding user and the order acceptance status mark in the information completion dataset are integrated into the review and filing result. In this embodiment, the data submitted by each user are integrated, and the review and filing result is constructed and stored in the preset review result database as a record, using the order acceptance status mark as an index.

[0056] This embodiment collects user-submitted data, performs character recognition on image data, assesses the validity of documents and certificates, evaluates credit status based on historical construction records, detects task conflicts, verifies insurance information, constructs a complete dataset, and conducts an overall review and judgment to achieve a compliance review mechanism for real-name construction order acceptance. Compared with existing experience, by accurately identifying the actual text content of key fields such as validity period or qualification number, it reduces misjudgments caused by information tampering or errors in filling out forms. Furthermore, by combining document validity judgment, it further determines whether the relevant documents submitted by the user are valid documents, improving review accuracy. By querying the user's historical construction tasks... By using a user's performance in historical tasks as a criterion, credit status is transformed into quantifiable indicators. Users exhibiting high-risk behaviors are flagged, and their historical behavior serves as one of the review criteria. Simultaneously, a timeline is constructed to determine if there are any conflicting tasks, allowing for the timely detection, freezing, and processing of abnormal task records to prevent waste of platform resources. Insurance information is matched with specific construction task information to determine the validity of insurance, preventing high-risk workers from performing tasks without protection. In summary, this real-name registration-based construction order compliance review mechanism enhances the review capabilities regarding information authenticity and task rationality, reduces the incidence of misjudgments, and ensures order security and credit transparency for construction workers.

[0057] Example 2 Please see Figure 2 As shown, for parts not described in detail in this embodiment, please refer to the description in Embodiment 1. A construction real-name order acceptance compliance review system is provided, including: The data acquisition module is used to collect text registration information and data image data of construction personnel, and to perform data cleaning to obtain accurate text registration information and high-definition data image data. The field recognition module is used to perform field recognition on high-definition data image data to obtain a document field dataset; the document field dataset is then integrated with precise text registration information to form a structured data dataset. The validity assessment module is used to assess the validity of structured data datasets and filter valid data to form a valid data dataset. The information detection module is used to assess the credit status of construction workers on the valid data dataset, and to label the assessment results to obtain a user-labeled data dataset; it also performs task conflict detection on the user-labeled data dataset and outputs a corrected data dataset. The data correction module is used to verify insurance information on the corrected data dataset, output insurance status labels to mark the corrected data dataset, and obtain the information-complete dataset. The review and judgment module is used to send the complete information dataset to the preset review terminal for overall judgment, generate the review and filing results and store them in the preset review result database; the modules are connected to each other via wired and / or wireless means.

[0058] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0059] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0060] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A compliance verification mechanism for construction order acceptance based on real-name registration, characterized in that, include: S1. Collect text registration information and data image data of construction personnel, and perform data cleaning to obtain accurate text registration information and high-definition data image data; S2. Perform field recognition on the high-definition data image to obtain the document field dataset; integrate the document field dataset with the precise text registration information into a structured data dataset; S3. Determine the validity of the structured data dataset and select valid data to form a valid data dataset; S4. Evaluate the credit status of construction workers on the valid data dataset, and label the evaluation results to obtain the user-labeled data dataset; perform task conflict detection on the user-labeled data dataset and output the corrected data dataset; S5. Verify the insurance information on the corrected data dataset, output insurance status labels to mark the corrected data dataset, and obtain the information-completed dataset; S6. Send the complete dataset to the preset review terminal for overall judgment, generate the review and filing results, and store them in the preset review result database.

2. The construction real-name order acceptance compliance review mechanism according to claim 1, characterized in that, The methods for performing field recognition include: Identify the text-containing field regions in each high-definition image data, construct a grid of preset size to divide the field region into image blocks, and obtain a set of field region image blocks; traverse each column of pixels in any field region image block, and count the number of pixels in each column that are higher than a preset grayscale threshold as a vertical projection array; if vertical projection arrays with consecutive projection values ​​all greater than the number threshold appear, then the number of consecutively appearing vertical projection arrays is taken as the width of the continuous pixel band; based on the width of each continuous pixel band, the field region is divided into candidate character image units; Each candidate character image unit is grayscale normalized to construct a standard character image matrix. The initial field strings are then sorted based on the center image coordinates of each standard character image matrix. The position coordinates of the initial field strings are matched with a preset document information distribution template to identify the character type of each character in the initial field strings. Contour enhancement is performed on the region containing each character in the initial field strings. Based on the character type, the enhanced character contours are matched with a dictionary database to output the actual field text. Finally, all characters in the initial field strings are converted into actual field text to obtain the document field dataset.

3. The compliance review mechanism for real-name order acceptance in construction projects according to claim 2, characterized in that, The methods for determining validity include: Extract document time attribute fields from the structured data dataset based on character type to form a document time field set. Group the document time field set based on the original position of the document time attribute fields to obtain document time field groups. Identify and replace symbols in each document time field group with date separators to output a unified time format string. Divide the unified time format string into an issuance time field and an expiration time field based on Chinese prompt words. Using the current date as the time base, compare the time interval between the time base and the issuance time field and the time interval between the time base and the expiration time field. If the issuance time field is later than the time base, or the expiration time is earlier than the time base, the corresponding unified time format string is determined to be an invalid string. Identify the issuance information corresponding to strings other than invalid strings. Determine whether the issuing unit exists and whether the issuance number is normal by querying a preset external database. If either the issuing unit does not exist or the issuance number cannot be queried, the corresponding string is also determined to be an invalid string. Remove the user-related data corresponding to the invalid string from the structured data dataset and add an invalid log. Integrate the invalid log with the remaining valid data into a valid data dataset.

4. The compliance review mechanism for real-name order acceptance in construction projects according to claim 3, characterized in that, The methods for assessing the credit status of construction workers include: Identify the unique identification code of each user in the valid data dataset, and obtain the historical construction behavior dataset of the corresponding user based on the unique identification code; convert each historical construction record in the historical construction behavior dataset into vector form and arrange them in chronological order to form the construction behavior vector sequence of the corresponding user; perform principal component analysis on the construction behavior vector sequence, output the principal component behavior vector, and construct the change trend curve based on the principal component behavior vector; The system acquires known high-risk behavior information and constructs risk behavior change curves based on this information. It then constructs a behavior similarity function to calculate the behavior similarity between the trend curve and the wind direction behavior change curve. If the behavior similarity exceeds a preset risk behavior similarity threshold, it compares the parameters of each dimension in the construction behavior vector sequence corresponding to the trend curve with a preset standard behavior level, identifies the dimension parameters whose values ​​exceed the anomaly threshold, and adds anomaly labels to the corresponding construction behavior vectors. It calculates the anomaly label density in the construction behavior vector sequence to obtain an anomaly index, and performs a weighted sum of the anomaly index and behavior similarity to obtain a risk assessment factor. Finally, it matches the risk assessment factor with a historical risk assessment level template to output the corresponding user's assessment result.

5. The construction real-name order acceptance compliance review mechanism according to claim 4, characterized in that, The methods for obtaining the historical construction behavior dataset of the corresponding user include: By matching the unique identification code with the corresponding user's historical records, the following fields are extracted: contractor information, construction task start and end time, construction task completion status, attendance rate, and construction project feedback information, to form a historical construction behavior dataset.

6. The construction real-name order acceptance compliance review mechanism according to claim 5, characterized in that, The methods for performing task conflict detection include: The user's execution task information in the user tag dataset is identified based on a unique identity code. The execution task information includes a task start and end time field and a project geographical location field. A task timeline for the corresponding user is constructed based on the execution task information. The task timeline is traversed to determine the start and end time intervals corresponding to each task record. It is then identified whether any two adjacent task records have an intersection of start and end time intervals. If they do, the intersection time length is calculated. The surface distance of the project geographical location of the adjacent task records is calculated. If the intersection time length is higher than a preset overlap time threshold or the surface distance is greater than a reasonable distance threshold, it is determined that there is a task conflict in the user's task timeline. Revise the task records of two or more tasks that have task conflicts to obtain status-updated task records; use the status-updated task records to overwrite the execution task information of the corresponding user in the user-labeled dataset to obtain the corrected data dataset.

7. The construction real-name order acceptance compliance review mechanism according to claim 6, characterized in that, The methods for revising task records include: For task records with task conflicts, adjust the fields by changing the corresponding task status field to a frozen status field and marking the task record as a non-compliant task; extract the task time field of the non-compliant task and clear it; identify the order-taking platform number of the non-compliant task, integrate the adjusted non-compliant task information with the order-taking platform number into a conflict handling log, and match the user's unique identification code and real name with the conflict handling log to obtain the status-updated task record.

8. The compliance review mechanism for real-name order acceptance in construction projects according to claim 7, characterized in that, The methods for verifying insurance information include: Retrieve each user's insurance records, including policy number, insurance status label, and insurance coverage items; identify the construction task type and risk level identifier of the corresponding user in the correction data dataset and match them with the insurance coverage items. If the construction task type is not included in the insurance coverage items or the insurance coverage level does not match the risk level identifier, the corresponding insurance record is marked as an inapplicable clause; monitor the insurance status label in real time. If the insurance record is outside the insurance effective period, it is also marked as an inapplicable clause; generate an insurance status field based on the specific invalidation of the inapplicable clause, and integrate the insurance status field with the corresponding policy number to form an insurance status label.

9. The compliance review mechanism for real-name order acceptance in construction projects according to claim 8, characterized in that, The methods for performing overall discrimination include: Extract the document fields, credit status assessment results, and insurance status labels from the information completion dataset separately. Then, compare them with the failure log and conflict handling log with the preset compliance review template. If all conditions of the preset compliance review template are met, the corresponding user is marked as an order-accepting user; otherwise, the user is marked as an order-unaccepting user. Integrate the relevant data and order-accepting status labels of the corresponding user from the information completion dataset into the review and filing results.

10. A construction order acceptance compliance review system, used to implement the construction order acceptance compliance review mechanism as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to collect text registration information and data image data of construction personnel, and to perform data cleaning to obtain accurate text registration information and high-definition data image data. The field recognition module is used to perform field recognition on high-definition data image data to obtain a document field dataset; the document field dataset is then integrated with precise text registration information to form a structured data dataset. The validity assessment module is used to assess the validity of structured data datasets and filter valid data to form a valid data dataset. The information detection module is used to assess the credit status of construction workers on the valid data dataset, and to label the assessment results to obtain a user-labeled data dataset; it also performs task conflict detection on the user-labeled data dataset and outputs a corrected data dataset. The data correction module is used to verify insurance information on the corrected data dataset, output insurance status labels to mark the corrected data dataset, and obtain the information-complete dataset. The review and judgment module is used to send the complete information dataset to the preset review terminal for overall judgment, generate the review and filing results and store them in the preset review result database; the modules are connected to each other via wired and / or wireless means.

Images