Foundation pit risk reporting processing system and method based on AI language professional large model

Abstract

The invention discloses a foundation pit risk reporting and processing system based on an AI language professional large model, and the system comprises a man-machine interaction module which is used for issuing an instruction in a foundation pit risk reporting and processing process, and automatically generating a risk processing measure scheme and related texts; the data uploading module is used for quickly collecting and uploading photos of the risk points of the foundation pit; the data preprocessing module is used for preprocessing the foundation pit risk point photos; the risk identification module is used for carrying out risk identification on the preprocessed foundation pit risk point photos through a foundation pit risk identification model and generating a preliminary risk assessment report; and the risk reporting module is used for checking the preliminary risk assessment report for the first time, issuing risk points or potential hidden dangers which are not recognized to the AI language professional large model through the man-machine interaction module, and generating a first risk assessment report. The invention further discloses a foundation pit risk reporting processing method based on the AI language professional large model. Precise identification, rapid reporting and timely processing of the foundation pit risk are realized.

Description

Technical Field

[0001] The present invention relates to a foundation pit risk reporting processing system and method based on an AI language professional large model. Background Art

[0002] Foundation pit engineering is an important part of civil engineering. Its construction process is complex and extremely risky, involving a large amount of earth excavation, support structure construction, groundwater treatment and other processes. Each step requires strict construction management. Once an accident occurs in the foundation pit engineering, it will not only have a major impact on the project itself, but may also pose a serious threat to the surrounding environment and personal safety. At present, risk management in foundation pit construction mainly relies on the inspection and experience judgment of on-site construction workers. By observing and recording the exposed risk points and potential risk points on the construction site, they are reported to project managers for analysis and processing. Although the conventional management model guarantees the safety of construction to a certain extent, there are obvious deficiencies and defects.

[0003] First, this process is time-consuming. It takes a long time from the discovery of risk points to reporting and then to processing. As a result, some risk points may develop further without timely processing, increasing the possibility of accidents. Secondly, the efficiency of risk identification and processing is low, mainly relying on the personal experience of construction workers and the technical level of managers, and there is a large degree of subjectivity and uncertainty. Especially when managers have an inaccurate understanding of risks or lack experience, misjudgments may occur, resulting in the failure to deal with risk points in a timely and effective manner, leading to serious accidents. In addition, the existing risk management model lacks systematicness and standardization, and cannot achieve rapid identification and standard processing of risk points. It is difficult to meet the requirements of safety and efficiency for foundation pit construction. There is an urgent need for a new technical means to quickly and accurately identify and process risk points in foundation pit construction.

[0004] Therefore, a foundation pit risk reporting and processing system and method based on an AI language professional large model is provided. Summary of the invention

[0005] The purpose of the present invention is to overcome the existing defects and provide a foundation pit risk reporting and processing system and method based on an AI language professional large model, which realizes accurate identification, rapid reporting and timely processing of foundation pit risks.

[0006] The technical solution to achieve the above purpose is:

[0007] The foundation pit risk reporting and processing system based on the AI ​​language professional large model of the present invention includes:

[0008] The human-computer interaction module is used to report foundation pit risks and issue instructions during the processing process, and automatically generate risk treatment measures and related texts;

[0009] Data upload module, used for rapid collection and upload of photos of foundation pit risk points;

[0010] Data preprocessing module, used to preprocess photos of foundation pit risk points;

[0011] The risk identification module is used to identify the risks of the pre-processed photos of the risk points of the foundation pit through the foundation pit risk identification model and generate a preliminary risk assessment report;

[0012] The risk reporting module is used to conduct the first review of the preliminary risk assessment report, and to issue the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module to generate the first risk assessment report;

[0013] The initial risk processing module is used to review whether the risk points in the first risk assessment report are accurate, and to issue the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module to generate a second risk assessment report;

[0014] The risk reprocessing module is used to review whether the treatment measures in the second risk assessment report are reasonable, generate a final risk assessment report based on the review results, and issue risk processing instructions through the human-computer interaction module;

[0015] The processing feedback module is used to direct on-site construction personnel to handle risks according to the final risk assessment report. After all risk points are handled, the construction inspector issues a processing completion instruction through the human-computer interaction module, and the AI ​​language professional large model automatically generates a risk disposal report;

[0016] Information release module, used for the circulation of risk assessment reports, risk disposal reports and expert consultation reports in the process of identifying, reporting, processing and feedback of foundation pit risk points;

[0017] The personnel management module is used to clarify the responsibilities of personnel in each link of the foundation pit risk reporting and handling process.

[0018] Preferably, in the human-computer interaction module, the human-computer dialogue interaction capability of the AI ​​language professional large model is used to identify the instructions issued by personnel at all levels and automatically execute them, including data reception before and after risk point processing, risk identification, risk adjustment, adjustment of treatment measures, expert processing, risk processing and completion of processing;

[0019] Utilize the integration capability of the professional knowledge base of large-scale foundation pit engineering in AI language to generate and adjust treatment measures for different types of risk points based on the risk types identified by the risk identification model;

[0020] Use AI language professional big models to integrate different information such as photos, texts, flow charts, etc. to generate risk assessment reports, risk disposal reports and expert consultation reports in text format;

[0021] In the data uploading module, the construction inspectors shall take a large number of photos at the foundation pit construction site on a regular basis. The photos taken shall cover all key locations of the construction site, including but not limited to the foundation pit slope, support structure, groundwater treatment area, and construction equipment and personnel activity area;

[0022] In the data preprocessing module, the uploaded photos are preprocessed to generate risk identification input data, and the preprocessing method includes:

[0023] Image denoising: The AI ​​language professional large model uses common image processing algorithms such as median filtering and mean filtering to denoise the uploaded photos and eliminate noise in the image;

[0024] Image enhancement: AI language professional large model uses common image enhancement methods such as histogram equalization and contrast stretching to adjust contrast and brightness;

[0025] Geometric correction, the AI ​​language professional large model calls methods such as affine transformation and perspective transformation to perform geometric correction on tilted or deformed images.

[0026] Preferably, the risk identification module includes:

[0027] The input unit automatically reads the risk identification input data through the trained foundation pit risk identification model, runs the model to identify the current risk point location and type, and marks it on the corresponding foundation pit risk point photo;

[0028] Feedback unit, used to feed back the photos of identified foundation pit risk points and risk types to the AI ​​language professional large model;

[0029] The preliminary plan generation unit automatically generates preliminary treatment measures for each risk point based on the type of risk point, and generates a preliminary risk assessment report, which is sent to the construction inspector. The preliminary treatment measures include specific treatment steps, flow charts and recommended measures. The preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measures.

[0030] The risk reporting module includes:

[0031] The first review unit is used by construction inspectors to conduct the first review of the risks identified in the preliminary risk assessment report, compare and analyze them with the actual situation on site, and determine whether the results of the risk identification model are accurate;

[0032] The first potential risk processing unit is used by construction inspectors to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module to the AI ​​language professional large model, adjust the preliminary risk assessment report, generate the first risk assessment report, and automatically submit it to the project leader for review.

[0033] Preferably, the risk initial processing module includes:

[0034] The second review unit is used by the project leader to review the risks identified in the first risk assessment report, compare and analyze them with the actual situation on site, and determine whether the results of the risk identification model are accurate;

[0035] The second potential risk processing unit is used by the project leader to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module to the AI ​​language professional large model, and adjust the first risk assessment report, generate a second risk assessment report, and automatically submit it to the project leader for review.

[0036] Preferably, in the risk reprocessing module, whether the treatment measures in the second risk assessment report are reasonable is reviewed, including three situations:

[0037] The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required;

[0038] Invalid solution: If the project leader believes that the treatment measure solution is invalid, the human-computer interaction module will issue a treatment measure adjustment instruction to the AI ​​language professional large model to modify the invalid part of the solution;

[0039] Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader;

[0040] Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations:

[0041] The plan is effective: the second risk assessment report is used as the final risk assessment report;

[0042] Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report;

[0043] Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

[0044] Preferably, in the processing feedback module, the risk treatment report includes each risk point type, comparison photos before and after treatment, treatment measures and treatment completion status;

[0045] In the information release module, when the AI ​​language professional model receives the report sending instruction issued by the human-computer interaction module, it will transfer the report to the next operator. According to the responsibilities of different personnel, it can be divided into:

[0046] When a risk is identified, the construction inspector issues a report circulation instruction through the human-computer interaction module, and the first risk assessment report will be sent to the project leader; when the risk is handled, the construction inspector issues a report circulation instruction through the human-computer interaction module, and the risk handling report will be sent to the project leader;

[0047] When expert processing is required, the project leader issues a report circulation instruction through the human-computer interaction module, and the expert consultation report will be sent to the relevant experts; when risk processing is required, the project leader issues a report circulation instruction through the human-computer interaction module, and the final risk assessment report will be sent to the construction inspector;

[0048] When the expert proposes a treatment measure, the expert issues a report circulation instruction through the human-computer interaction module, and the expert consultation report containing feedback treatment measure suggestions will be sent to the project leader;

[0049] After each person receives a new report, it will be displayed in the system and a reminder message will be sent via SMS;

[0050] The personnel management module includes:

[0051] The project leader enters the information of each participant, including name, position, unit, and responsibilities;

[0052] Automatically create the corresponding login account and initial password for the participants. According to the division of responsibilities of different personnel, the corresponding permission functions are automatically opened, including the following three types:

[0053] Construction inspector: responsible for collecting and reporting risks, reviewing preliminary risk assessment reports, and directing construction personnel to implement treatment measures for each risk point;

[0054] Project leader: responsible for reviewing the first risk assessment report, reviewing the treatment measures plan of the second risk assessment report, and issuing treatment measures instructions;

[0055] Expert: Responsible for proposing more professional treatment measures for risk points that cannot be resolved on site.

[0056] The second method of the present invention for processing foundation pit risk reporting based on the AI ​​language professional large model includes:

[0057] Step S1, a construction inspector regularly takes a large number of photos at the foundation pit construction site, and the photos taken should cover various key locations of the construction site, where the photos taken include but are not limited to foundation pit slopes, support structures, groundwater treatment areas, and construction equipment and personnel activity areas;

[0058] Step S2, performing image denoising, image enhancement and geometric correction preprocessing operations on the photos of the risk points of the foundation pit to generate risk identification input data;

[0059] Step S3, using the foundation pit risk identification model to identify the risks of the pre-processed foundation pit risk point photos and generate a preliminary risk assessment report;

[0060] Step S4: The construction inspector conducts a first review of the preliminary risk assessment report and issues unidentified risk points or potential hazards to the AI ​​language professional large model to generate the first risk assessment report;

[0061] Step S5: The project leader reviews whether the risk points in the first risk assessment report are accurate, and issues unidentified risk points or potential risks to the AI ​​language professional large model to generate a second risk assessment report;

[0062] Step S6: The project leader reviews whether the treatment measures in the second risk assessment report are reasonable, generates a final risk assessment report based on the review results, and issues risk treatment instructions;

[0063] In step S7, the construction inspector directs the on-site construction personnel to handle the risks according to the final risk assessment report. After all risk points have been handled, the construction inspector issues a completion instruction, and the AI ​​language professional large model automatically generates a risk handling report, in which the risk handling report includes the type of each risk point, comparison photos before and after the treatment, the treatment measures plan and the completion status of the treatment.

[0064] Preferably, step S3 comprises:

[0065] Step S31, automatically reading risk identification input data through the trained foundation pit risk identification model, running the model to identify the current risk point location and type, and marking them on the corresponding foundation pit risk point photo;

[0066] Step S32, feeding back the identified foundation pit risk point photos and risk types to the AI ​​language professional large model;

[0067] Step S33, according to the type of risk point, the AI ​​language professional large model automatically generates a preliminary treatment measure plan for each risk point, and generates a preliminary risk assessment report, which is sent to the construction inspectors. The preliminary treatment measure plan includes specific processing steps, flow charts and recommended measures, and the preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measures.

[0068] Preferably, step S4 comprises:

[0069] Step S41, the construction inspector conducts a first review of the risks identified in the preliminary risk assessment report, compares and analyzes the risks with the actual situation on site, and determines whether the results of the risk identification model are accurate;

[0070] Step S42: The construction inspector sends the unidentified risk points or potential hazards to the AI ​​language professional big model, and the AI ​​language professional big model adjusts the preliminary risk assessment report, generates the first risk assessment report, and automatically submits it to the project leader for review;

[0071] The step S5 comprises:

[0072] Step S51, the project leader reviews the risks identified in the first risk assessment report, compares and analyzes them with the actual situation on site, and determines whether the results of the risk identification model are accurate;

[0073] In step S52, the project leader sends the unidentified risk points or potential hidden dangers to the AI ​​language professional big model. The AI ​​language professional big model adjusts the first risk assessment report, generates a second risk assessment report, and automatically submits it to the project leader for review.

[0074] Preferably, in step S6, whether the treatment measures in the second risk assessment report are reasonable is reviewed, including three situations:

[0075] The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required;

[0076] Invalid solution: If the project leader believes that the treatment measure solution is invalid, the human-computer interaction module will issue a treatment measure adjustment instruction to the AI ​​language professional large model to modify the invalid part of the solution;

[0077] Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader;

[0078] Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations:

[0079] The plan is effective: the second risk assessment report is used as the final risk assessment report;

[0080] Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report;

[0081] Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

[0082] The beneficial effects of the present invention are:

[0083] 1) The present invention aims at the slow identification speed of risk points. By introducing an image recognition algorithm, the real-time monitoring of the construction site and the rapid identification of risk points are achieved, which greatly shortens the time of risk identification;

[0084] 2) The present invention aims at the low accuracy of risk identification. By using a large professional model of AI language and combining a large amount of foundation pit engineering construction data and image data, deep learning and training are carried out to improve the accuracy of risk identification and more accurately judge potential risk points.

[0085] 3) The present invention aims at the problem of untimely risk reporting and processing. By constructing an integrated risk reporting and processing module, the present invention can realize real-time reporting and processing of risk points. After the risk point is identified, a report is immediately generated and transferred to the relevant management personnel. The management personnel can directly view the detailed information of the risk point through the system and quickly formulate processing measures according to the suggested solutions provided by the system.

[0086] 4) The present invention aims to solve the problem that risk treatment measures rely on the experience of managers. Through the AI ​​language professional large model, the present invention provides standardized and systematic treatment plan suggestions, reduces the dependence on personal experience, and ensures the scientificity and effectiveness of the treatment measures;

[0087] The present invention utilizes the trained foundation pit risk image recognition model to identify on-site foundation pit risks in real time, and utilizes the AI ​​language professional large model in the foundation pit field to automatically generate a treatment measure plan and submit it to the project leader. The project leader is responsible for reviewing whether the plan is effective. If the plan is effective, rectification will be carried out. If the plan is invalid, the plan will be adjusted. If the on-site risk cannot be resolved, expert processing will be selected. At this time, the AI ​​language professional large model automatically generates an expert consultation report and sends it to relevant experts, proposing more professional solutions, thereby realizing accurate identification, rapid reporting and timely processing of foundation pit risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 It is a module diagram of the foundation pit risk reporting and processing system based on the AI ​​language professional large model of the present invention;

[0089] Figure 2 It is a specific module diagram of the risk identification module in the present invention;

[0090] Figure 3 It is a specific module diagram of the risk reporting module in the present invention;

[0091] Figure 4 It is a specific module diagram of the risk initial processing module in the present invention;

[0092] Figure 5 It is a flow chart of the foundation pit risk reporting and processing method based on the AI ​​language professional large model of the present invention;

[0093] Figure 6 It is a specific flow chart of the present invention for performing risk identification on the pre-processed foundation pit risk point photos through the foundation pit risk identification model and generating a preliminary risk assessment report;

[0094] Figure 7 It is a specific flow chart in which the construction inspector in the present invention conducts the first review of the preliminary risk assessment report and issues the unidentified risk points or potential hazards to the AI ​​language professional large model to generate the first risk assessment report;

[0095] Figure 8 The project leader in the present invention reviews whether the risk points in the first risk assessment report are accurate, and issues unidentified risk points or potential hazards to the AI ​​language professional large model to generate a specific flow chart for the second risk assessment report. DETAILED DESCRIPTION

[0096] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0097] The present invention will be further described below in conjunction with the accompanying drawings.

[0098] like Figure 1 As shown, the foundation pit risk reporting and processing system based on the AI ​​language professional large model includes: human-computer interaction module 1, data upload module 2, data preprocessing module 3, risk identification module 4, risk reporting module 5, risk initial processing module 6, risk re-processing module 7, processing feedback module 8, information release module 9 and personnel management module 10.

[0099] The human-computer interaction module 1 is used to report foundation pit risks and issue instructions during the processing process, and automatically generate risk treatment measures and related texts.

[0100] In the embodiment, the human-computer dialogue interaction capability of the AI ​​language professional large model is used to identify the instructions issued by personnel at all levels and automatically execute them, including data reception before and after risk point processing, risk identification, risk adjustment, adjustment of treatment measures, expert processing, risk processing and completion of processing;

[0101] Utilize the integration capability of the professional knowledge base of large-scale foundation pit engineering in AI language to generate and adjust treatment measures for different types of risk points based on the risk types identified by the risk identification model;

[0102] Use AI language professional large models to integrate different information such as photos, texts, flow charts, etc. to generate risk assessment reports, risk disposal reports and expert consulting reports in text format (.doc / .docx).

[0103] Data upload module 2 is used for the rapid collection and upload of photos of foundation pit risk points.

[0104] In the embodiment, a large number of photos are taken regularly by construction inspectors at the foundation pit construction site. The photos taken should cover various key locations of the construction site, including but not limited to foundation pit slopes, support structures, groundwater treatment areas, and construction equipment and personnel activity areas.

[0105] The data preprocessing module 3 is used to preprocess the photos of the risk points of the foundation pit.

[0106] In the embodiment, the uploaded photos are pre-processed to generate risk identification input data, and the pre-processing method includes:

[0107] Image denoising: The AI ​​language professional large model uses common image processing algorithms such as median filtering and mean filtering to denoise the uploaded photos and eliminate noise in the image;

[0108] Image enhancement: AI language professional large model uses common image enhancement methods such as histogram equalization and contrast stretching to adjust contrast and brightness;

[0109] Geometric correction, the AI ​​language professional large model calls methods such as affine transformation and perspective transformation to perform geometric correction on tilted or deformed images.

[0110] The risk identification module 4 is used to identify the risks of the pre-processed foundation pit risk point photos through the foundation pit risk identification model and generate a preliminary risk assessment report.

[0111] like Figure 2 As shown, the risk identification module 4 includes: an input unit 41, a feedback unit 42 and a preliminary solution generating unit 43.

[0112] The input unit 41 automatically reads the risk identification input data through the trained foundation pit risk identification model, runs the model to identify the current risk point location and type, and marks it on the corresponding foundation pit risk point photo.

[0113] The feedback unit 42 is used to feed back the identified foundation pit risk point photos and risk types to the AI ​​language professional large model.

[0114] The preliminary solution generation unit 43 automatically generates a preliminary treatment measure plan for each risk point according to the type of risk point, and generates a preliminary risk assessment report, which is sent to the construction inspector. The preliminary treatment measure plan includes specific treatment steps, flow charts and recommended measures, and the preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measure plans.

[0115] The risk reporting module 5 is used to conduct the first review of the preliminary risk assessment report, and to send the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module 1 to generate the first risk assessment report.

[0116] like Figure 3 As shown, the risk reporting module 5 includes: a first review unit 51 and a first potential risk processing unit 52 .

[0117] The first review unit 51 is used for the construction inspector to conduct a first review of the risks identified in the preliminary risk assessment report, compare and analyze the risks with the actual situation on site, and determine whether the results of the risk identification model are accurate.

[0118] The first potential risk processing unit 52 is used by the construction inspector to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module 1 to the AI ​​language professional large model, and adjust the preliminary risk assessment report, generate the first risk assessment report, and automatically submit it to the project leader for review.

[0119] The initial risk processing module 6 is used to review whether the risk points in the first risk assessment report are accurate, and to send the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module 1 to generate a second risk assessment report.

[0120] like Figure 4 As shown, the initial risk processing module 6 includes: a second review unit 61 and a second potential risk processing unit 62 .

[0121] The second review unit 61 is used for the project leader to review the risk points identified in the first risk assessment report, compare and analyze them with the actual situation on site, and determine whether the results of the risk identification model are accurate.

[0122] The second potential risk processing unit 62 is used by the project leader to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module 1 to the AI ​​language professional large model, and adjust the first risk assessment report, generate a second risk assessment report, and automatically submit it to the project leader for review.

[0123] The risk reprocessing module 7 is used to review whether the treatment measures in the second risk assessment report are reasonable, generate a final risk assessment report based on the review results, and issue risk processing instructions through the human-computer interaction module 1.

[0124] In the embodiment, the review of whether the treatment measures in the second risk assessment report are reasonable includes three situations:

[0125] The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required;

[0126] Invalid solution: If the project leader believes that the treatment measure solution is invalid, he / she will issue a treatment measure adjustment instruction to the AI ​​language professional large model through the human-computer interaction module 1 to modify the invalid part of the solution;

[0127] Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader;

[0128] Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations:

[0129] The plan is effective: the second risk assessment report is used as the final risk assessment report;

[0130] Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report;

[0131] Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

[0132] The processing feedback module 8 is used to direct on-site construction personnel to handle risks according to the final risk assessment report. After all risk points have been handled, the construction inspector issues a processing completion instruction through the human-computer interaction module 1, and the AI ​​language professional large model automatically generates a risk disposal report.

[0133] In the embodiment, the risk treatment report includes each risk point type, comparison photos before and after treatment, treatment measures and treatment completion status.

[0134] Information release module 9 is used for the circulation of risk assessment reports, risk disposal reports and expert consultation reports in the process of identifying, reporting, processing and feedback of foundation pit risk points.

[0135] In the embodiment, when the AI ​​language professional large model receives the report sending instruction issued by the human-computer interaction module 1, it will transfer the report to the next operator. According to the responsibilities of different personnel, it can be divided into:

[0136] When a risk is identified, the construction inspector issues a report circulation instruction through the human-computer interaction module 1, and the first risk assessment report will be sent to the project leader; when the risk is handled, the construction inspector issues a report circulation instruction through the human-computer interaction module 1, and the risk handling report will be sent to the project leader;

[0137] When expert processing is required, the project leader issues a report circulation instruction through the human-computer interaction module 1, and the expert consultation report will be sent to the relevant experts; when risk processing is required, the project leader issues a report circulation instruction through the human-computer interaction module 1, and the final risk assessment report will be sent to the construction inspector;

[0138] When the expert proposes a treatment measure, the expert issues a report circulation instruction through the human-computer interaction module 1, and the expert consultation report containing feedback treatment measure suggestions will be sent to the project leader;

[0139] When each person receives a new report, it will be displayed in the system and a reminder message will be sent via SMS.

[0140] In the embodiment, the formats of the generated preliminary risk assessment report, the first risk assessment report, the second risk assessment report, the final risk assessment report, the risk disposal report and the expert consultation report are all .doc.

[0141] The personnel management module 10 is used to clarify the responsibilities of personnel in each link of the foundation pit risk reporting and handling process.

[0142] In the embodiment, the content mainly entered by the personnel management module 10 includes:

[0143] The project leader enters the information of each participant, including name, position, unit, and responsibilities;

[0144] Automatically create the corresponding login account and initial password for the participants. The project leader logs in to the system for the first time, enters the information of each participant, and generates the corresponding login account and initial password. Among them, the corresponding permission functions are automatically opened according to the division of responsibilities of different personnel, including the following three types:

[0145] Construction inspector: responsible for collecting and reporting risks, reviewing preliminary risk assessment reports, and directing construction personnel to implement treatment measures for each risk point;

[0146] Project leader: responsible for reviewing the first risk assessment report, reviewing the treatment measures plan of the second risk assessment report, and issuing treatment measures instructions;

[0147] Expert: Responsible for proposing more professional treatment measures for risk points that cannot be resolved on site.

[0148] like Figure 5 As shown in the figure, the foundation pit risk reporting and processing method based on the AI ​​language professional large model includes:

[0149] Step S1, a construction inspector shall regularly take a large number of photos at the foundation pit construction site. The photos taken shall cover all key locations of the construction site, including but not limited to foundation pit slopes, support structures, groundwater treatment areas, and construction equipment and personnel activity areas.

[0150] Step S2, performing image denoising, image enhancement and geometric correction preprocessing operations on the photos of the risk points of the foundation pit to generate risk identification input data.

[0151] Step S3, using the foundation pit risk identification model to identify the risks of the pre-processed foundation pit risk point photos and generate a preliminary risk assessment report.

[0152] like Figure 6 As shown, step S3 includes:

[0153] Step S31, automatically reading risk identification input data through the trained foundation pit risk identification model, running the model to identify the current risk point location and type, and marking them on the corresponding foundation pit risk point photo;

[0154] Step S32, feeding back the identified foundation pit risk point photos and risk types to the AI ​​language professional large model;

[0155] Step S33, according to the type of risk point, the AI ​​language professional large model automatically generates a preliminary treatment measure plan for each risk point, and generates a preliminary risk assessment report, which is sent to the construction inspectors. The preliminary treatment measure plan includes specific processing steps, flow charts and recommended measures, and the preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measures.

[0156] In step S4, the construction inspector conducts the first review of the preliminary risk assessment report and sends the unidentified risk points or potential hazards to the AI ​​language professional large model to generate the first risk assessment report.

[0157] like Figure 7 As shown, step S4 includes:

[0158] Step S41, the construction inspector conducts a first review of the risks identified in the preliminary risk assessment report, compares and analyzes the risks with the actual situation on site, and determines whether the results of the risk identification model are accurate;

[0159] In step S42, the construction inspector sends the unidentified risk points or potential hidden dangers to the AI ​​language professional big model. The AI ​​language professional big model adjusts the preliminary risk assessment report, generates the first risk assessment report, and automatically submits it to the project leader for review.

[0160] In step S5, the project leader reviews whether the risk points in the first risk assessment report are accurate, and issues unidentified risk points or potential hidden dangers to the AI ​​language professional large model to generate a second risk assessment report.

[0161] like Figure 8 As shown, step S5 includes:

[0162] Step S51, the project leader reviews the risks identified in the first risk assessment report, compares and analyzes them with the actual situation on site, and determines whether the results of the risk identification model are accurate;

[0163] In step S52, the project leader sends the unidentified risk points or potential hidden dangers to the AI ​​language professional big model. The AI ​​language professional big model adjusts the first risk assessment report, generates a second risk assessment report, and automatically submits it to the project leader for review.

[0164] Step S6: The project leader reviews whether the treatment measures in the second risk assessment report are reasonable, generates a final risk assessment report based on the review results, and issues risk treatment instructions.

[0165] In the embodiment, the review of whether the treatment measures in the second risk assessment report are reasonable includes three situations:

[0166] The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required;

[0167] Invalid solution: If the project leader believes that the treatment measure solution is invalid, he will issue treatment measure adjustment instructions to the AI ​​language professional large model through the human-computer interaction module to modify the invalid part of the solution;

[0168] Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader;

[0169] Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations:

[0170] The plan is effective: the second risk assessment report is used as the final risk assessment report;

[0171] Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report;

[0172] Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

[0173] In step S7, the construction inspector directs the on-site construction personnel to handle the risks according to the final risk assessment report. After all risk points have been handled, the construction inspector issues a completion instruction, and the AI ​​language professional large model automatically generates a risk handling report, in which the risk handling report includes the type of each risk point, comparison photos before and after the treatment, the treatment measures plan and the completion status of the treatment.

[0174] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. The foundation pit risk reporting and processing system based on the AI ​​language professional large model is characterized by: include: The human-computer interaction module is used to report foundation pit risks and issue instructions during the processing process, and automatically generate risk treatment measures and related texts; Data upload module, used for rapid collection and upload of photos of foundation pit risk points; Data preprocessing module, used to preprocess photos of foundation pit risk points; The risk identification module is used to identify the risks of the pre-processed photos of the risk points of the foundation pit through the foundation pit risk identification model and generate a preliminary risk assessment report; The risk reporting module is used to conduct the first review of the preliminary risk assessment report, and to issue the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module to generate the first risk assessment report; The initial risk processing module is used to review whether the risk points in the first risk assessment report are accurate, and to issue the unidentified risk points or potential hidden dangers to the AI ​​language professional large model through the human-computer interaction module to generate a second risk assessment report; The risk reprocessing module is used to review whether the treatment measures in the second risk assessment report are reasonable, generate a final risk assessment report based on the review results, and issue risk processing instructions through the human-computer interaction module; The processing feedback module is used to direct on-site construction personnel to handle risks according to the final risk assessment report. After all risk points are handled, the construction inspector issues a processing completion instruction through the human-computer interaction module, and the AI ​​language professional large model automatically generates a risk disposal report; Information release module, used for the circulation of risk assessment reports, risk disposal reports and expert consultation reports in the process of identifying, reporting, processing and feedback of foundation pit risk points; The personnel management module is used to clarify the responsibilities of personnel in each link of the foundation pit risk reporting and handling process.

2. According to claim 1, the foundation pit risk reporting and processing system based on the AI ​​language professional large model is characterized in that: In the human-computer interaction module, the human-computer dialogue interaction capability of the AI ​​language professional large model is used to identify the instructions issued by personnel at all levels and automatically execute them, including data reception before and after risk point processing, risk identification, risk adjustment, adjustment of treatment measures, expert processing, risk processing and completion of processing; Utilize the integration capability of the professional knowledge base of large-scale foundation pit engineering in AI language to generate and adjust treatment measures for different types of risk points based on the risk types identified by the risk identification model; Use AI language professional big models to integrate different information such as photos, texts, flow charts, etc. to generate risk assessment reports, risk disposal reports and expert consultation reports in text format; In the data uploading module, the construction inspectors shall take a large number of photos at the foundation pit construction site on a regular basis. The photos taken shall cover all key locations of the construction site, including but not limited to the foundation pit slope, support structure, groundwater treatment area, and construction equipment and personnel activity area; In the data preprocessing module, the uploaded photos are preprocessed to generate risk identification input data, and the preprocessing method includes: Image denoising: The AI ​​language professional large model uses common image processing algorithms such as median filtering and mean filtering to denoise the uploaded photos and eliminate noise in the image; Image enhancement: AI language professional large model uses common image enhancement methods such as histogram equalization and contrast stretching to adjust contrast and brightness; Geometric correction, the AI ​​language professional large model calls methods such as affine transformation and perspective transformation to perform geometric correction on tilted or deformed images.

3. The foundation pit risk reporting and processing system based on the AI ​​language professional large model according to claim 2 is characterized in that: The risk identification module includes: The input unit automatically reads the risk identification input data through the trained foundation pit risk identification model, runs the model to identify the current risk point location and type, and marks it on the corresponding foundation pit risk point photo; Feedback unit, used to feed back the photos of identified foundation pit risk points and risk types to the AI ​​language professional large model; The preliminary plan generation unit automatically generates preliminary treatment measures for each risk point based on the type of risk point, and generates a preliminary risk assessment report, which is sent to the construction inspector. The preliminary treatment measures include specific treatment steps, flow charts and recommended measures. The preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measures. The risk reporting module includes: The first review unit is used by construction inspectors to conduct the first review of the risks identified in the preliminary risk assessment report, compare and analyze them with the actual situation on site, and determine whether the results of the risk identification model are accurate; The first potential risk processing unit is used by construction inspectors to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module to the AI ​​language professional large model, adjust the preliminary risk assessment report, generate the first risk assessment report, and automatically submit it to the project leader for review.

4. The foundation pit risk reporting and processing system based on the AI ​​language professional large model according to claim 1 is characterized in that: The risk initial processing module includes: The second review unit is used by the project leader to review the risks identified in the first risk assessment report, compare and analyze them with the actual situation on site, and determine whether the results of the risk identification model are accurate; The second potential risk processing unit is used by the project leader to send the unidentified risk points or potential hidden dangers issued by the human-computer interaction module to the AI ​​language professional large model, and adjust the first risk assessment report, generate a second risk assessment report, and automatically submit it to the project leader for review.

5. The foundation pit risk reporting and processing system based on the AI ​​language professional large model according to claim 4 is characterized in that: In the risk reprocessing module, the rationality of the treatment measures in the second risk assessment report is reviewed, including three situations: The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required; Invalid solution: If the project leader believes that the treatment measure solution is invalid, the human-computer interaction module will issue a treatment measure adjustment instruction to the AI ​​language professional large model to modify the invalid part of the solution; Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader; Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations: The plan is effective: the second risk assessment report is used as the final risk assessment report; Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report; Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

6. The foundation pit risk reporting and processing system based on the AI ​​language professional large model according to claim 1 is characterized in that: In the processing feedback module, the risk treatment report includes the type of each risk point, comparison photos before and after the treatment, the treatment measure plan and the completion status of the treatment; In the information release module, when the AI ​​language professional model receives the report sending instruction issued by the human-computer interaction module, it will transfer the report to the next operator. According to the responsibilities of different personnel, it can be divided into: When a risk is identified, the construction inspector issues a report circulation instruction through the human-computer interaction module, and the first risk assessment report will be sent to the project leader; when the risk is handled, the construction inspector issues a report circulation instruction through the human-computer interaction module, and the risk handling report will be sent to the project leader; When expert processing is required, the project leader issues a report circulation instruction through the human-computer interaction module, and the expert consultation report will be sent to the relevant experts; when risk processing is required, the project leader issues a report circulation instruction through the human-computer interaction module, and the final risk assessment report will be sent to the construction inspector; When the expert proposes a treatment measure, the expert issues a report circulation instruction through the human-computer interaction module, and the expert consultation report containing feedback treatment measure suggestions will be sent to the project leader; After each person receives a new report, it will be displayed in the system and a reminder message will be sent via SMS; The personnel management module includes: The project leader enters the information of each participant, including name, position, unit, and responsibilities; Automatically create the corresponding login account and initial password for the participants. According to the division of responsibilities of different personnel, the corresponding permission functions are automatically opened, including the following three types: Construction inspector: responsible for collecting and reporting risks, reviewing preliminary risk assessment reports, and directing construction personnel to implement treatment measures for each risk point; Project leader: responsible for reviewing the first risk assessment report, reviewing the treatment measures plan of the second risk assessment report, and issuing treatment measures instructions; Expert: Responsible for proposing more professional treatment measures for risk points that cannot be resolved on site.

7. The foundation pit risk reporting and processing method based on the AI ​​language professional large model is characterized by: include: Step S1, a construction inspector regularly takes a large number of photos at the foundation pit construction site, and the photos taken should cover various key locations of the construction site, where the photos taken include but are not limited to foundation pit slopes, support structures, groundwater treatment areas, and construction equipment and personnel activity areas; Step S2, performing image denoising, image enhancement and geometric correction preprocessing operations on the photos of the risk points of the foundation pit to generate risk identification input data; Step S3, using the foundation pit risk identification model to identify the risks of the pre-processed foundation pit risk point photos and generate a preliminary risk assessment report; Step S4: The construction inspector conducts a first review of the preliminary risk assessment report and issues unidentified risk points or potential hazards to the AI ​​language professional large model to generate the first risk assessment report; Step S5: The project leader reviews whether the risk points in the first risk assessment report are accurate, and issues unidentified risk points or potential risks to the AI ​​language professional large model to generate a second risk assessment report; Step S6: The project leader reviews whether the treatment measures in the second risk assessment report are reasonable, generates a final risk assessment report based on the review results, and issues risk treatment instructions; In step S7, the construction inspector directs the on-site construction personnel to handle the risks according to the final risk assessment report. After all risk points have been handled, the construction inspector issues a completion instruction, and the AI ​​language professional large model automatically generates a risk handling report, in which the risk handling report includes the type of each risk point, comparison photos before and after the treatment, the treatment measures plan and the completion status of the treatment.

8. The method for processing foundation pit risk reporting based on the AI ​​language professional large model according to claim 7 is characterized in that: The step S3 comprises: Step S31, automatically reading risk identification input data through the trained foundation pit risk identification model, running the model to identify the current risk point location and type, and marking them on the corresponding foundation pit risk point photo; Step S32, feeding back the identified foundation pit risk point photos and risk types to the AI ​​language professional large model; Step S33, according to the type of risk point, the AI ​​language professional large model automatically generates a preliminary treatment measure plan for each risk point, and generates a preliminary risk assessment report, which is sent to the construction inspectors. The preliminary treatment measure plan includes specific processing steps, flow charts and recommended measures, and the preliminary risk assessment report contains all original photos, risk point annotated photos, risk types and preliminary treatment measures.

9. The method for processing foundation pit risk reporting based on the AI ​​language professional large model according to claim 7 is characterized in that: The step S4 comprises: Step S41, the construction inspector conducts a first review of the risks identified in the preliminary risk assessment report, compares and analyzes the risks with the actual situation on site, and determines whether the results of the risk identification model are accurate; Step S42: The construction inspector sends the unidentified risk points or potential hazards to the AI ​​language professional big model, and the AI ​​language professional big model adjusts the preliminary risk assessment report, generates the first risk assessment report, and automatically submits it to the project leader for review; The step S5 comprises: Step S51, the project leader reviews the risks identified in the first risk assessment report, compares and analyzes them with the actual situation on site, and determines whether the results of the risk identification model are accurate; In step S52, the project leader sends the unidentified risk points or potential hidden dangers to the AI ​​language professional big model. The AI ​​language professional big model adjusts the first risk assessment report, generates a second risk assessment report, and automatically submits it to the project leader for review.

10. The method for processing foundation pit risk reporting based on the AI ​​language professional large model according to claim 7 is characterized in that: In step S6, whether the treatment measures in the second risk assessment report are reasonable is reviewed, including three situations: The plan is effective: If the project leader believes that the treatment measure plan is effective, no modification is required; Invalid solution: If the project leader believes that the treatment measure solution is invalid, the human-computer interaction module will issue a treatment measure adjustment instruction to the AI ​​language professional large model to modify the invalid part of the solution; Expert processing: If the project leader believes that the treatment measures are ineffective and cannot come up with a reliable solution, he / she will issue an expert processing instruction. The AI ​​language professional large model will automatically generate a detailed expert consultation report, including the basic profile of the project, photos of risk points, risk types, and treatment measures, and send this report to one or more experts. After receiving the risk report in the information release module, the relevant experts will propose more professional treatment measures based on their professional knowledge and experience, and send the expert consultation report containing feedback on treatment measures to the project leader; Based on the review results of the treatment measures, the AI ​​language professional model will generate a final risk assessment report in three situations: The plan is effective: the second risk assessment report is used as the final risk assessment report; Invalid solution: The AI ​​language professional model automatically adjusts the second risk assessment report based on the treatment measures provided by the project leader and generates a final risk assessment report; Expert processing: The AI ​​language professional model automatically adjusts the second risk assessment report and generates a final risk assessment report based on the processing measures provided by the project leader and relevant experts.

Images