Target slope-based inspection report generation method, device, equipment and medium
By automatically generating slope inspection reports using lidar and intelligent models, the problem of cumbersome and inefficient generation processes in existing technologies has been solved, achieving efficient and reliable report generation.
Patent Information
- Application Number
- CN202610926450.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-25
- Publication Date
- 2026-11-03
AI Technical Summary
The existing process for generating target slope inspection reports is cumbersome, consumes a lot of human resources and time, and is easily affected by human intervention, resulting in low efficiency.
The system uses lidar to acquire slope scanning data, and combines retrieval enhancement generation models, text feature extraction models, and large language models to automatically generate inspection reports, including basic slope information, disease identification results, and treatment suggestions.
It reduces the time required to generate inspection reports, improves efficiency, enhances report reliability, and avoids the impact of human intervention.
Smart Images
Figure CN123030641A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of slope detection technology, and in particular to a method, apparatus, equipment and medium for generating inspection reports based on target slopes. Background Technology
[0002] Slope inspection reports are important reference materials in operation and maintenance management, and are also essential materials for project acceptance, qualification review, and project archiving. When managers review slope inspection reports, they can quickly grasp key issues and accurately understand the actual condition of the slope.
[0003] However, the current process for generating inspection reports for target slopes is cumbersome, hindering the improvement of report generation efficiency. This is because existing technologies primarily rely on manual compilation to generate these reports. Manual compilation requires on-site inspections, combining inspection experience with standardized formats to create the report. This process is not only cumbersome but also consumes significant human and time resources, increasing report generation time and making it susceptible to human intervention. Therefore, it is detrimental to improving report generation efficiency. Summary of the Invention
[0004] This application provides a method, apparatus, equipment, and medium for generating inspection reports based on target slopes, in order to solve the technical problem that the existing process for generating inspection reports for target slopes is cumbersome and not conducive to improving the efficiency of inspection report generation.
[0005] In a first aspect, embodiments of this application provide a method for generating inspection reports based on target slopes, applied to an electronic device connected to a lidar system. The method for generating inspection reports includes: Obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command; The three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point are obtained from the scan data. The slope monitoring points are the monitoring points set on the surface of the target slope. The three-dimensional spatial coordinates, echo intensity, and echo number of each slope monitoring point are used to form the data record corresponding to each slope monitoring point. The data records corresponding to each slope monitoring point are then spliced together to generate the splicing feature of the target slope. The splicing features of the target slope are input into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. From the document database, the basic inspection information, disease identification results, and treatment suggestions for the reference slope are obtained. These information, along with the treatment suggestions, are combined to form a prompt text. The prompt text is then input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
[0006] In one possible implementation of the first aspect, the steps of acquiring the scanning command of the target slope, sending the scanning command of the target slope to the lidar, and receiving the scanning data of the target slope returned by the lidar according to the scanning command include: Send a test request to the lidar and receive the response message returned by the lidar based on the test request; Obtain the status identifier of the lidar from the response message. When the status identifier is normal, obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command.
[0007] In one possible implementation of the first aspect, obtaining the three-dimensional spatial coordinates, echo intensity, and echo count corresponding to each slope monitoring point from the scan data, wherein the slope monitoring points are monitoring points deployed on the surface of the target slope, includes: The scan data is input into the parsing tool, which then processes the scan data to generate parsed scan data. From the parsed scan data, obtain the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point. The slope monitoring points are the monitoring points set up on the surface of the target slope.
[0008] In one possible implementation of the first aspect, the basic inspection information, disease identification results, and treatment suggestions for the reference slope are obtained from a document database. These information, along with the treatment suggestions, are combined to form a prompt text. The prompt text is then input into a text feature extraction model to generate features for the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope, including: Using JSON format, the slope identifier and query instructions are encapsulated into a query request. The query request is transmitted to the access interface of the document database. The response message returned by the access interface based on the query request is received. The basic inspection information of the reference slope, the disease identification results of the reference slope, and the treatment suggestions of the reference slope are obtained from the response message. The basic inspection information of the reference slope, the identification results of the reference slope's defects, and the treatment suggestions for the reference slope are combined into a prompt text. The prompt text is input into a text feature extraction model, which generates the features of the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope to generate the fused features of the target slope. The fused features of the target slope are then input into a large language model, which generates an inspection report for the target slope. The inspection report for the target slope includes the basic inspection information of the target slope, the identification results of the target slope's defects, and the treatment suggestions for the target slope.
[0009] In one possible implementation of the first aspect, after obtaining the basic inspection information of the reference slope, the disease identification results of the reference slope, and the treatment suggestions for the reference slope from the document database, and composing the basic inspection information, disease identification results, and treatment suggestions of the reference slope into a prompt text, inputting the prompt text into a text feature extraction model, generating features of the prompt text through the text feature extraction model, and using a feature fusion model to fuse the features of the prompt text and the splicing features of the target slope to generate fused features of the target slope, inputting the fused features of the target slope into a large language model, and generating an inspection report of the target slope through the large language model, the inspection report of the target slope including the basic inspection information, disease identification results, and treatment suggestions of the target slope, the inspection report generation method further includes: Create a display window to show the inspection report of the target slope.
[0010] In one possible implementation of the first aspect, the basic information for the inspection of the reference slope includes the location of the reference slope, the name of the reference slope, the height of the reference slope, the slope of the reference slope, the inspection date of the reference slope, and the weather information of the reference slope. The disease identification results of the reference slope include at least one of the following: the reference slope has subsidence, the reference slope has cracks, and the reference slope has landslides. The treatment recommendations for the reference slope include at least one of the following: treatment recommendations for the reference slope with subsidence, treatment recommendations for the reference slope with cracks, and treatment recommendations for the reference slope with landslides.
[0011] In one possible implementation of the first aspect, the basic information for the inspection of the target slope includes the location of the target slope, the name of the target slope, the height of the target slope, the slope of the target slope, the inspection date of the target slope, and the weather information of the target slope. The results of the target slope disease identification include at least one of the following: the target slope has subsidence, the target slope has cracks, and the target slope has landslides. The treatment recommendations for the target slope include at least one of the following: treatment recommendations for the target slope with subsidence, treatment recommendations for the target slope with cracks, and treatment recommendations for the target slope with landslides.
[0012] Secondly, embodiments of this application provide a target slope-based inspection report generation device, applied to an electronic device connected to a lidar system, comprising: The first acquisition module is used to acquire the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command. The second acquisition module is used to acquire the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point from the scan data. The slope monitoring points are monitoring points set on the surface of the target slope. The first generation module is used to assemble the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point into a data record corresponding to each slope monitoring point, and to splice the data records corresponding to each slope monitoring point to generate the splicing feature of the target slope. The third acquisition module is used to input the splicing features of the target slope into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. The second generation module is used to obtain the basic inspection information, disease identification results, and treatment suggestions of the reference slope from the document database. It then combines these information into a prompt text, inputs it into a text feature extraction model, generates features for the prompt text, and uses a feature fusion model to fuse the features of the prompt text with the spliced features of the target slope to generate fused features of the target slope. Finally, it inputs these fused features into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
[0013] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the inspection report generation method described in the first aspect above.
[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the inspection report generation method described in the first aspect above.
[0015] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to execute the inspection report generation method described in the first aspect.
[0016] The beneficial effects of the embodiments of this application are as follows: Firstly, the system retrieves basic inspection information, disease identification results, and treatment suggestions for reference slopes from a document database. This information is then combined to form a prompt text. The prompt text is input into a text feature extraction model, which generates features for the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model, which generates an inspection report for the target slope. The inspection report includes basic inspection information, disease identification results, and treatment suggestions for the target slope. Since manual compilation is not required, the generation time of the target slope inspection report is reduced, thus improving the efficiency of target slope inspection report generation. Secondly, since large language models are not affected by human intervention, they are conducive to improving the reliability of inspection reports for target slopes. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This application scenario diagram illustrates the method for generating inspection reports provided in the embodiments of this application. Figure 2 This is a flowchart illustrating the inspection report generation method provided in this application embodiment; Figure 3 A flowchart illustrating the implementation of S202 provided in this application embodiment; Figure 4 A schematic block diagram of an inspection report generation device provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.
[0020] The inspection report generation method provided in this application embodiment can be applied to electronic devices. The electronic devices are connected to LiDAR, and the electronic devices include, but are not limited to, servers, mobile phones, tablets, wearable devices, vehicle-mounted devices, and laptops. This application embodiment does not impose any restrictions on the specific type of electronic device.
[0021] Please see Figure 1 , Figure 1 The application scenario diagram of the inspection report generation method provided in the embodiments of this application is described in detail below: The electronic device connects to the lidar; the electronic device sends a test request to the lidar and receives a response message from the lidar based on the test request; Obtain the status identifier of the lidar from the response message. When the status identifier is normal, obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command.
[0022] LiDAR has high data acquisition efficiency and wide coverage. It can fully capture the deformation differences of various parts of the target slope, effectively avoiding the risk of missing key deformation areas due to sparse sensor deployment.
[0023] In this embodiment of the application, the electronic device receives the scanning data of the target slope returned by the lidar, which can overcome the spatial limitations of on-site operations and realize the remote processing of the scanning data.
[0024] like Figure 2 As shown in the embodiments of this application, the method for generating inspection reports includes the following steps, which are detailed below: S201: Obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command. The target slope is the slope to be identified, which includes, but is not limited to, slopes in roads, mines, and hydropower stations that have potential risks of collapse, landslides, and cracking.
[0025] S202, obtain the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point from the scan data. The slope monitoring points are the monitoring points set on the surface of the target slope. The three-dimensional spatial coordinates corresponding to a monitoring point are numerical parameters representing the spatial location of the monitoring point under a unified geographic coordinate system. They are generally represented by three sets of values: X, Y, and Z coordinates, in a geodetic coordinate system or a local engineering coordinate system. The X and Y coordinates determine the planar position of the monitoring point, while the Z coordinate corresponds to the point's elevation. This parameter accurately defines the actual geographical location of each monitoring point, providing a spatial benchmark for slope morphology measurement, deformation comparison across multiple data periods, and hazard location identification.
[0026] The echo intensity corresponding to a monitoring point is an indicator of the energy amplitude of the reflected signal after the probe beam emitted by the lidar comes into contact with targets such as the slope surface, rock mass, and attached objects. Different materials, roughness, and moisture content of slope objects will produce different echo intensity values, which can be used to distinguish different types of ground features such as soil and rock, vegetation, water accumulation, and loose rocks, and help identify abnormal areas on the slope. It is an important attribute information in point cloud data.
[0027] The echo count corresponding to a monitoring point is the total number of effective reflection signals generated sequentially along the propagation path of a single laser beam emitted by the lidar. When a single laser beam passes through multiple layers of media such as vegetation branches and leaves, uneven slopes, multiple echoes will be generated in sequence. Based on the echo count, surface points and vegetation points can be separated, vegetation interference can be eliminated, the true surface outline of the slope can be restored, and the effectiveness of the 3D model and monitoring data can be improved.
[0028] S203: The three-dimensional spatial coordinates, echo intensity, and echo number corresponding to each slope monitoring point are used to form the data record corresponding to each slope monitoring point. The data records corresponding to each slope monitoring point are spliced together to generate the splicing feature of the target slope. S204. Input the splicing features of the target slope into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, perform a similarity search on the splicing features of the target slope in the vector library to obtain the search results. In the search results, obtain the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope, and select the preset slope with the highest similarity as the reference slope. Among them, the Retrieval-Augmented Generation (RAG) model is a hybrid artificial intelligence model architecture that combines information retrieval and text generation.
[0029] Among them, the preset slope is the slope that has already been identified.
[0030] The vector library is a database that stores the preset features of each preset slope.
[0031] S205: Obtain the basic inspection information, disease identification results, and treatment suggestions for the reference slope from the document database. Combine these information into a prompt text. Input the prompt text into a text feature extraction model to generate features for the prompt text. Use a feature fusion model to fuse the features of the prompt text with the spliced features of the target slope to generate fused features for the target slope. Input the fused features of the target slope into a large language model to generate an inspection report for the target slope. The inspection report for the target slope includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
[0032] The process involves retrieving basic inspection information, disease identification results, and treatment suggestions for a reference slope from a document database. This information is then combined to form a prompt text. The prompt text is input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope. Using JSON format, the slope identifier and query instructions are encapsulated into a query request. The query request is transmitted to the access interface of the document database. The response message returned by the access interface based on the query request is received. The basic inspection information of the reference slope, the disease identification results of the reference slope, and the treatment suggestions of the reference slope are obtained from the response message. The basic inspection information of the reference slope, the identification results of the reference slope's defects, and the treatment suggestions for the reference slope are combined into a prompt text. The prompt text is input into a text feature extraction model, which generates the features of the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope to generate the fused features of the target slope. The fused features of the target slope are then input into a large language model, which generates an inspection report for the target slope. The inspection report for the target slope includes the basic inspection information of the target slope, the identification results of the target slope's defects, and the treatment suggestions for the target slope.
[0033] The method for generating a reference slope includes retrieving basic inspection information, disease identification results, and treatment suggestions from a document database. This information is then combined with other relevant data to form a prompt text. The prompt text is input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope. The method for generating the inspection report also includes: Create a display window to show the inspection report of the target slope.
[0034] The basic information for the inspection of the reference slope includes the location of the reference slope, the name of the reference slope, the height of the reference slope, the slope of the reference slope, the inspection date of the reference slope, and the weather information of the reference slope. The disease identification results of the reference slope include at least one of the following: the reference slope has subsidence, the reference slope has cracks, and the reference slope has landslides. The treatment recommendations for the reference slope include at least one of the following: treatment recommendations for the reference slope with subsidence, treatment recommendations for the reference slope with cracks, and treatment recommendations for the reference slope with landslides.
[0035] The basic information for the inspection of the target slope includes the location of the target slope, the name of the target slope, the height of the target slope, the slope of the target slope, the inspection date of the target slope, and the weather information of the target slope. The results of the target slope disease identification include at least one of the following: the target slope has subsidence, the target slope has cracks, and the target slope has landslides. The treatment recommendations for the target slope include at least one of the following: treatment recommendations for the target slope with subsidence, treatment recommendations for the target slope with cracks, and treatment recommendations for the target slope with landslides.
[0036] For ease of explanation, the following example is provided: For example, the recommended treatment for subsidence on the target slope is to delineate a warning area and manage the risks, block water infiltration through seepage prevention and drainage, and restore the bearing capacity of the soil and rock mass by layered backfilling and compaction and grouting reinforcement, depending on the depth of subsidence. The recommended treatment for cracks in the target slope is to seal and fill the cracks according to their specifications, and to add anchoring structures to constrain the displacement of expanding cracks. The recommended measures for dealing with landslides on the target slope are to prioritize early warning evacuation and full closure of the slope in response to the landslide hazard, combine unloading and weight reduction, drainage and diversion to reduce sliding force, and deploy anti-sliding support structures to suppress overall sliding.
[0037] The beneficial effects of the embodiments of this application are as follows: Firstly, the system retrieves basic inspection information, disease identification results, and treatment suggestions for reference slopes from a document database. This information is then combined to form a prompt text. The prompt text is input into a text feature extraction model, which generates features for the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model, which generates an inspection report for the target slope. The inspection report includes basic inspection information, disease identification results, and treatment suggestions for the target slope. Since manual compilation is not required, the generation time of the target slope inspection report is reduced, thus improving the efficiency of target slope inspection report generation. Secondly, since large language models are not affected by human intervention, they are conducive to improving the reliability of inspection reports for target slopes.
[0038] Please see Figure 3 , Figure 3 The implementation flowchart of S202 provided in the embodiments of this application is described in detail below: S301, input the scan data into the parsing tool, and use the parsing tool to parse and process the scan data to generate parsed scan data; S302. From the parsed scan data, obtain the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point. The slope monitoring points are monitoring points set on the surface of the target slope.
[0039] In this embodiment, the scan data is input into a parsing tool, which then processes the scan data to generate parsed scan data. The parsed scan data can be stripped of redundant frames, invalid packets, and interference information, thereby improving the usability and readability of the parsed scan data.
[0040] For the inspection report generation method described in the above embodiments, please refer to [link / reference]. Figure 4 , Figure 4 This is a schematic block diagram of the inspection report generation device provided in the embodiments of this application. Figure 4 The inspection report generation device 400 shown can be applied to, for example... Figure 1 The application scenario diagram shows electronic devices. The following section uses electronic devices as an example to illustrate this. Figure 4 The inspection report generation device 400 shown will be described in detail. The inspection report generation device 400 may include a first acquisition module 401, a second acquisition module 402, a first generation module 403, a third acquisition module 404, and a second generation module 405.
[0041] The first acquisition module 401 is used to acquire the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command. The second acquisition module 402 is used to acquire the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point from the scan data. The slope monitoring points are monitoring points set on the surface of the target slope. The first generation module 403 is used to assemble the three-dimensional spatial coordinates, echo intensity, and echo number corresponding to each slope monitoring point into a data record corresponding to each slope monitoring point, and to splice the data records corresponding to each slope monitoring point to generate the splicing feature of the target slope. The third acquisition module 404 is used to input the splicing features of the target slope into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. The second generation module 405 is used to obtain the basic inspection information, disease identification results, and treatment suggestions of the reference slope from the document database. It then combines these information into a prompt text, inputs it into a text feature extraction model, generates features for the prompt text, and uses a feature fusion model to fuse the features of the prompt text with the spliced features of the target slope to generate fused features of the target slope. Finally, it inputs these fused features into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
[0042] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0043] The beneficial effects of the embodiments of this application are as follows: Firstly, the system retrieves basic inspection information, disease identification results, and treatment suggestions for reference slopes from a document database. This information is then combined to form a prompt text. The prompt text is input into a text feature extraction model, which generates features for the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model, which generates an inspection report for the target slope. The inspection report includes basic inspection information, disease identification results, and treatment suggestions for the target slope. Since manual compilation is not required, the generation time of the target slope inspection report is reduced, thus improving the efficiency of target slope inspection report generation. Secondly, since large language models are not affected by human intervention, they are conducive to improving the reliability of inspection reports for target slopes.
[0044] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.
[0045] like Figure 5 As shown, Figure 5The electronic device includes: at least one processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the at least one processor 20, wherein the processor 20 executes the computer program 22 to implement the steps in any of the above method embodiments.
[0046] The electronic device may include, but is not limited to, processor 20 and memory 21. Those skilled in the art will understand that... Figure 5 This is merely an example of an electronic device and does not constitute a limitation on electronic devices. It may include more or fewer components than shown in the illustration, or combinations of certain components, or different components. For example, it may also include input / output devices, network access devices, etc.
[0047] The processor 20 is used to run a computer program 22 stored in the memory 21, and performs the following steps when executing the computer program 22: Obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command; The three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point are obtained from the scan data. The slope monitoring points are the monitoring points set on the surface of the target slope. The three-dimensional spatial coordinates, echo intensity, and echo number of each slope monitoring point are used to form the data record corresponding to each slope monitoring point. The data records corresponding to each slope monitoring point are then spliced together to generate the splicing feature of the target slope. The splicing features of the target slope are input into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. From the document database, the basic inspection information, disease identification results, and treatment suggestions for the reference slope are obtained. These information, along with the treatment suggestions, are combined to form a prompt text. The prompt text is then input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
[0048] The processor 20 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors, field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0049] Furthermore, the memory 21 may include both internal storage units and external storage devices of the electronic device. The memory 21 is used to store the operating system, applications, boot loader, data, and other programs, such as the program code of the computer program. The memory 21 can also be used to temporarily store data that has been output or will be output.
[0050] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.
[0051] Since the computer program stored in the computer-readable storage medium can execute any of the target slope-based inspection report generation methods provided in the embodiments of this application, the computer-readable storage medium can achieve the beneficial effects that any of the target slope-based inspection report generation methods provided in the embodiments of this application can achieve, as detailed in the preceding embodiments, and will not be repeated here.
[0052] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A method for generating inspection reports based on target slopes, characterized in that, Applied to electronic devices connected to LiDAR, the inspection report generation method includes: Obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command; The three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point are obtained from the scan data. The slope monitoring points are the monitoring points set on the surface of the target slope. The three-dimensional spatial coordinates, echo intensity, and echo number of each slope monitoring point are used to form the data record corresponding to each slope monitoring point. The data records corresponding to each slope monitoring point are then spliced together to generate the splicing feature of the target slope. The splicing features of the target slope are input into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. From the document database, the basic inspection information, disease identification results, and treatment suggestions for the reference slope are obtained. These information, along with the treatment suggestions, are combined to form a prompt text. The prompt text is then input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
2. The inspection report generation method according to claim 1, characterized in that, The process of acquiring the scanning command for the target slope, sending the scanning command for the target slope to the lidar, and receiving the scanning data of the target slope returned by the lidar based on the scanning command includes: Send a test request to the lidar and receive the response message returned by the lidar based on the test request; Obtain the status identifier of the lidar from the response message. When the status identifier is normal, obtain the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command.
3. The inspection report generation method according to claim 1, characterized in that, The process involves obtaining the three-dimensional spatial coordinates, echo intensity, and echo count for each slope monitoring point from the scanned data. These slope monitoring points are those deployed on the surface of the target slope. The scan data is input into the parsing tool, which then processes the scan data to generate parsed scan data. From the parsed scan data, obtain the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point. The slope monitoring points are the monitoring points set up on the surface of the target slope.
4. The inspection report generation method according to claim 1, characterized in that, From the document database, the basic inspection information, disease identification results, and treatment suggestions for the reference slope are obtained. These information, along with the treatment suggestions, are combined to form a prompt text. This prompt text is then input into a text feature extraction model to generate its features. A feature fusion model is used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope. Using JSON format, the slope identifier and query instructions are encapsulated into a query request. The query request is transmitted to the access interface of the document database. The response message returned by the access interface based on the query request is received. The basic inspection information of the reference slope, the disease identification results of the reference slope, and the treatment suggestions of the reference slope are obtained from the response message. The basic inspection information of the reference slope, the identification results of the reference slope's defects, and the treatment suggestions for the reference slope are combined into a prompt text. The prompt text is input into a text feature extraction model, which generates the features of the prompt text. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope to generate the fused features of the target slope. The fused features of the target slope are then input into a large language model, which generates an inspection report for the target slope. The inspection report for the target slope includes the basic inspection information of the target slope, the identification results of the target slope's defects, and the treatment suggestions for the target slope.
5. The method for generating inspection reports according to claim 1, characterized in that, The method for generating a reference slope includes retrieving basic inspection information, disease identification results, and treatment suggestions from a document database. This information is then combined with other relevant data to form a prompt text. The prompt text is input into a text feature extraction model to generate its features. A feature fusion model is then used to fuse the features of the prompt text with the spliced features of the target slope, generating a fused feature for the target slope. This fused feature is then input into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope. The method for generating the inspection report also includes: Create a display window to show the inspection report of the target slope.
6. The inspection report generation method according to claim 1, characterized in that, The basic information for the inspection of the reference slope includes the location of the reference slope, the name of the reference slope, the height of the reference slope, the slope of the reference slope, the inspection date of the reference slope, and the weather information of the reference slope. The disease identification results of the reference slope include at least one of the following: the reference slope has subsidence, the reference slope has cracks, and the reference slope has landslides. The treatment recommendations for the reference slope include at least one of the following: treatment recommendations for the reference slope with subsidence, treatment recommendations for the reference slope with cracks, and treatment recommendations for the reference slope with landslides.
7. The method for generating inspection reports according to claim 1, characterized in that, The basic information for the inspection of the target slope includes the location of the target slope, the name of the target slope, the height of the target slope, the slope of the target slope, the inspection date of the target slope, and the weather information of the target slope. The results of the target slope disease identification include at least one of the following: the target slope has subsidence, the target slope has cracks, and the target slope has landslides. The treatment recommendations for the target slope include at least one of the following: treatment recommendations for the target slope with subsidence, treatment recommendations for the target slope with cracks, and treatment recommendations for the target slope with landslides.
8. A device for generating inspection reports based on target slopes, characterized in that, Applications in electronic devices, where the electronic device connects to the lidar, including: The first acquisition module is used to acquire the scanning command of the target slope, send the scanning command of the target slope to the lidar, and receive the scanning data of the target slope returned by the lidar according to the scanning command. The second acquisition module is used to acquire the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point from the scan data. The slope monitoring points are monitoring points set on the surface of the target slope. The first generation module is used to assemble the three-dimensional spatial coordinates, echo intensity, and echo count of each slope monitoring point into a data record corresponding to each slope monitoring point, and to splice the data records corresponding to each slope monitoring point to generate the splicing feature of the target slope. The third acquisition module is used to input the splicing features of the target slope into the preset retrieval enhancement generation model. Through the retrieval enhancement generation model, the splicing features of the target slope are searched for similarity in the vector library to obtain the search results. In the search results, the similarity between the preset features of each preset slope in the vector library and the splicing features of the target slope is obtained, and the preset slope with the highest similarity is selected as the reference slope. The second generation module is used to obtain the basic inspection information, disease identification results, and treatment suggestions of the reference slope from the document database. It then combines these information into a prompt text, inputs it into a text feature extraction model, generates features for the prompt text, and uses a feature fusion model to fuse the features of the prompt text with the spliced features of the target slope to generate fused features of the target slope. Finally, it inputs these fused features into a large language model to generate an inspection report for the target slope. The inspection report includes the basic inspection information, disease identification results, and treatment suggestions for the target slope.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the inspection report generation method as described in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the inspection report generation method as described in any one of claims 1 to 7.