Automatic multifunctional three-dimensional scanning system and method based on language image model and intelligent path planning

Abstract

The invention discloses an automatic multifunctional three-dimensional scanning system and method based on a language image model and intelligent path planning, and the system comprises a voice / text input module which is used for receiving and processing the voice and text input of a user, and determining the description and report requirements of a workpiece to be scanned; the workpiece automatic tracking module is used for automatically searching, identifying and selecting a to-be-scanned workpiece according to information input by voice; the real-time scanning module is used for acquiring point cloud data and surface scanning data of the workpiece in real time; the automatic path planning and optimizing module is used for planning a scanning path of the mechanical arm according to the real-time point cloud data and the surface scanning data; the scanning area automatic switching module is used for scanning in the current visual field according to the scanning path and switching to a next non-scanning area after confirmation is completed; and the personalized report generation module is used for generating a detection report according to the specific requirements of the user and the workpiece characteristics.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial automation and intelligent manufacturing, and in particular relates to an automated multifunctional three-dimensional scanning system and method based on a language image model and intelligent path planning. Background Art

[0002] There are mainly the following types of workpiece scanning systems on the market: Traditional manual scanning systems rely on operators to scan workpieces with handheld scanning equipment. Such systems are usually equipped with portable scanners and computer terminals for real-time display and processing of scanned data. Disadvantages: Manual scanning systems require high skills from operators and require operators to have certain scanning experience and skills. Due to the influence of human factors, scanning accuracy and efficiency are often difficult to guarantee. In addition, long-term manual operation will cause operator fatigue and affect work quality. Fixed automatic scanning systems usually consist of fixed scanners and conveyor belts. The workpiece moves to the scanning area via the conveyor belt and is automatically scanned by the scanner. Disadvantages: Fixed systems have high requirements for environmental configuration and require specially designed and installed conveyor belts and scanning equipment. This system has poor flexibility and is difficult to adapt to workpieces of different sizes and shapes. For large workpieces, the coverage of fixed systems is limited and cannot achieve comprehensive scanning. Robotic arm assisted scanning systems use robotic arms to carry scanning equipment to scan workpieces. Such systems control the movement path of the robotic arm through programming to achieve automatic scanning. Disadvantages: Most existing robotic arm scanning systems rely on preset paths and lack intelligent path planning capabilities. When the position or shape of the workpiece changes, the path needs to be reprogrammed, which increases the complexity of the operation. In addition, such systems often lack an effective collision detection mechanism during the path planning process, which can easily cause the robot arm to collide with the workpiece or other equipment, affecting the safety and accuracy of the scan.

[0003] The existing technologies have the following defects: Most existing scanning systems rely on manual operation or fixed path control, with a low degree of automation, resulting in cumbersome operation and low efficiency. Existing systems have high requirements for environmental configuration and require specific lighting conditions and spatial layout, which limits their application in different production environments. Existing systems have high requirements for the professional skills of operators, which increases the complexity of operation and training costs. When scanning large workpieces, existing systems usually require multiple adjustments and scans to cover the entire surface, and rely entirely on manual operation, which is inefficient. Although the efficiency of small-scale scanning can be improved through automated equipment, these devices are usually expensive, making it difficult for small or medium-sized enterprises to afford them. The reports generated by existing systems are usually based on fixed templates, lacking personalized customization and targeted analysis, resulting in homogeneous report content and failing to meet the specific needs of different customers or tasks. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an automated multifunctional three-dimensional scanning system and method based on language image model and intelligent path planning, which can not only automatically complete the scanning of workpieces, but also generate detailed inspection reports. The reports contain specific text and data information, and provide detailed descriptions of the specific conditions of each workpiece.

[0005] On the one hand, to achieve the above-mentioned purpose, the present invention provides an automated multifunctional three-dimensional scanning system based on a language image model and intelligent path planning, comprising:

[0006] Voice / text input module, workpiece automatic tracking module, real-time scanning module, automatic path planning and optimization module, scanning area automatic switching module and personalized report generation module;

[0007] The voice / text input module is used to receive and process the user's voice and text input to determine the workpiece to be scanned;

[0008] The workpiece automatic tracking module is used to automatically find, identify and select the workpiece to be scanned according to the information input by voice;

[0009] The real-time scanning module is used to obtain point cloud data and surface scanning data of the workpiece in real time;

[0010] The automated path planning and optimization module is used to plan the scanning path of the robot arm based on the real-time point cloud data and surface scanning data;

[0011] The scanning area automatic switching module is used to scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation;

[0012] The personalized report generation module is used to generate a test report according to the user's specific needs and workpiece characteristics.

[0013] According to the automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the speech / text input module includes a speech / text receiving unit and a speech / text recognition unit;

[0014] The voice / text receiving unit is used to receive the user's voice and text input information;

[0015] The speech / text recognition unit is used to input a large language model according to the user's speech and text input information, parse the speech input information, extract the workpiece information and the scanning task requirements.

[0016] According to the automatic multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the workpiece automatic tracking module includes an automatic workpiece scanning unit and an automatic tracking unit;

[0017] The automatic workpiece scanning unit is used to automatically move the robot arm to find, identify and select the workpiece to be scanned according to the voice and text input information;

[0018] The automatic tracking unit is used to continuously track the position of the workpiece when the robot arm moves.

[0019] According to the automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the real-time scanning module includes a data acquisition unit and a data processing unit;

[0020] The data acquisition unit is used to acquire the point cloud data and the surface scanning data of the workpiece in real time;

[0021] The data processing unit is used to process and update the point cloud data of the current scanning area.

[0022] According to the automatic multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the automatic path planning and optimization module includes a path automatic planning unit and a path optimization unit;

[0023] The automatic path planning unit is used to plan the scanning path of the robot arm according to the real-time point cloud data and the surface scanning data;

[0024] The path optimization unit is used to optimize the moving path of the robot arm according to the real-time point cloud data and the surface scanning data.

[0025] According to the automatic multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the scanning area automatic switching module includes an area detection unit and an automatic switching unit;

[0026] The area detection unit is used to detect the scanning status within the current field of view;

[0027] The automatic switching unit is used to automatically move the robot arm to the next un-scanned area after the current area is scanned.

[0028] According to the automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning provided by the present invention, the personalized report generation module includes a report generation unit and a data analysis unit;

[0029] The report generation unit is used to process specific requirements and artifact characteristics of user input according to the large language model;

[0030] The data analysis unit is used to generate a test report based on the specific requirements and workpiece characteristics input by the user and in combination with the scanning data.

[0031] On the other hand, to achieve the above purpose, the present invention also provides an automated multifunctional three-dimensional scanning method based on a language image model and intelligent path planning, comprising the following steps:

[0032] Receive and process user voice and text input to determine the description of the workpiece to be scanned and the reporting requirements;

[0033] Identify and select the workpiece to be scanned based on the voice and text input information;

[0034] Acquire point cloud data and surface scanning data of workpieces in real time;

[0035] Plan the scanning path of the robot arm based on real-time point cloud data;

[0036] Scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation;

[0037] Generate inspection reports based on user-specific requirements and workpiece characteristics.

[0038] Technical effect of the invention: The present invention discloses an automated multifunctional three-dimensional scanning system and method based on a language image model and intelligent path planning. By introducing intelligent path planning and deep learning algorithms, the system can make autonomous decisions and plan scanning paths without human intervention, thus achieving highly automated workpiece scanning. The system is designed to be suitable for a variety of industrial environments, without the need for special lighting conditions or spatial layout, and can operate stably in complex and changeable production environments. The operator only needs to input the workpiece information through voice or text, and the system can automatically complete the scanning task, greatly reducing the skill requirements for the operator. By adopting intelligent path planning and collision detection algorithms, the system can efficiently cover the surface of large workpieces, reducing the need for multiple adjustments and scanning, and improving scanning efficiency. By using general-purpose robotic arms, depth cameras and other equipment, and optimizing the scanning process using intelligent algorithms, the system reduces equipment costs while maintaining efficient automation, making it more economical. Through a large language model, the system can generate personalized inspection reports based on the specific needs and workpiece characteristics input by the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0040] Figure 1 A schematic diagram of the structure of an automated multifunctional three-dimensional scanning system based on a language image model and intelligent path planning according to an embodiment of the present invention;

[0041] Figure 2A schematic diagram of a flow chart of an automated multifunctional three-dimensional scanning method based on a language image model and intelligent path planning according to an embodiment of the present invention;

[0042] Figure 3 This is a structural schematic diagram of an automated multifunctional three-dimensional scanning device based on a language image model and intelligent path planning according to an embodiment of the present invention, wherein 1 is a robotic arm, 2 is a depth camera, 3 is a three-dimensional scanning device, 4 is an operation panel, and 5 is a computing unit. DETAILED DESCRIPTION

[0043] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0044] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0045] like Figure 1 As shown, in this embodiment, an automated multifunctional three-dimensional scanning system based on a language image model and intelligent path planning is provided, including: a voice / text input module, an automatic workpiece tracking module, a real-time scanning module, an automated path planning and optimization module, an automatic scanning area switching module, and a personalized report generation module;

[0046] The voice / text input module is used to receive and process the user's voice and text input to determine the description of the workpiece to be scanned and the report requirements;

[0047] The workpiece automatic tracking module is used to automatically find, identify and select the workpiece to be scanned according to the information input by voice;

[0048] The real-time scanning module is used to obtain point cloud data and surface scanning data of the workpiece in real time;

[0049] The automated path planning and optimization module is used to plan the scanning path of the robot arm based on the real-time point cloud data and surface scanning data;

[0050] The scanning area automatic switching module is used to scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation;

[0051] The personalized report generation module is used to generate a test report according to the user's specific needs and workpiece characteristics.

[0052] Furthermore, the voice / text input module includes a voice / text receiving unit and a voice / text recognition unit; the voice / text receiving unit is used to receive the user's voice and text input information; the voice / text recognition unit is used to input the large language model according to the user's voice and text input information, parse the voice input information, extract the workpiece information and the scanning task requirements. The voice / text recognition unit is connected to the microphone and the large language model to process the voice and text data in real time. The large language model is integrated in the computing unit to parse the voice input and generate corresponding interactive text and control instructions.

[0053] Furthermore, the workpiece automatic tracking module includes an automatic workpiece scanning unit and an automatic tracking unit; the automatic workpiece scanning unit is used to automatically move the robot arm to find, identify and select the workpiece to be scanned according to voice and text input information; the automatic tracking unit is used to continuously track the position of the workpiece when the robot arm moves. The image language model processes the workpiece image in real time by connecting to the depth camera and generates a bounding box of the workpiece in the image. The long video tracking model is connected to the robot arm controller to ensure continuous tracking and updating of the workpiece position during the movement of the robot arm.

[0054] Furthermore, the real-time scanning module includes a data acquisition unit and a data processing unit; the data acquisition unit is used to acquire the point cloud data and surface scanning data of the workpiece in real time; the data processing unit is used to process and update the point cloud data of the current scanning area. The depth camera and scanner are installed at the end of the robot arm to collect the workpiece data in real time. The data processing unit is connected to the depth camera and scanner to process and update the point cloud data in real time.

[0055] Furthermore, the automated path planning and optimization module includes an automatic path planning unit and a path optimization unit; the automatic path planning unit is used to plan the scanning path of the robot arm according to the real-time point cloud data and surface scanning data; the path optimization unit is used to optimize the movement path of the robot arm according to the real-time point cloud data and surface scanning data. The path planning algorithm is integrated into the computing unit to generate and adjust the scanning path of the robot arm in real time. The reinforcement learning algorithm works in conjunction with the path planning algorithm to continuously optimize the scanning efficiency and accuracy of the robot arm.

[0056] Furthermore, the scanning area automatic switching module includes an area detection unit and an automatic switching unit; the area detection unit is used to detect the scanning status in the current field of view; the automatic switching unit is used to automatically move the robot arm to the next unscanned area after the current area is scanned. The area detection unit is connected to the data processing unit to monitor the scanning progress in real time. The automatic switching controller controls the robot arm to automatically move to the unscanned area according to the detection results to ensure that the entire workpiece is covered and scanned.

[0057] Furthermore, the personalized report generation module includes a report generation unit and a data analysis unit; the report generation unit is used to process the specific requirements and workpiece characteristics input by the user according to the large language model; the data analysis unit is used to generate a detection report in combination with the scan data according to the specific requirements and workpiece characteristics input by the user. The report generation unit is connected to the large language model and the data analysis unit to generate a personalized detection report including specific scan data, analysis results, answers to specific questions and suggestions.

[0058] like Figure 2 As shown, this embodiment provides an automated multifunctional three-dimensional scanning method based on a language image model and intelligent path planning, comprising the following steps:

[0059] Receive and process user voice and text input to determine the name of the workpiece to be scanned;

[0060] Identify and select the workpiece to be scanned based on the voice and text input information;

[0061] Acquire point cloud data and surface scanning data of workpieces in real time;

[0062] Plan the scanning path of the robot arm based on real-time point cloud data;

[0063] Scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation;

[0064] Generate inspection reports based on user-specific requirements and workpiece characteristics.

[0065] The specific implementation process includes the following:

[0066] The operator inputs the name of the workpiece for which information needs to be collected through voice or text, and follows the system prompts to complete a simple description of the workpiece and subsequent report requirements.

[0067] The Large Language Model processes voice or text information and separates the object information requested to be scanned, including appearance descriptions such as name, color, shape, and report requirement information.

[0068] While the robot arm rotates automatically, it cooperates with the visual language model to detect the position of the workpiece in real time. Based on the object information and the photos displayed by the camera, the visual language model first adjusts the image taken by the camera to the required size and divides it into image blocks, and then processes these image blocks through the Vision Transformer (ViT) encoder to generate an image feature vector. At the same time, the text information describing the workpiece is converted into a text feature representation through the text encoder. Then, using these image and text features, the detection head calculates the category probability and bounding box coordinates of each object in the image. When the confidence score reaches the preset standard, the robot arm stops moving, selects the object that meets the standard in the image, and marks its boundaries and numbers.

[0069] The operator determines whether there is a target object in the image. If so, the operator enters the target object number for the next step of generating a mask. If not, the operator continues the previous scanning operation.

[0070] The image captured by the camera is input into the image encoder of the vision model to extract high-dimensional image features. Then, the detected object bounding box is input into the prompt encoder as the prompt information. Next, the image features are combined with the prompt information through the mask decoder to generate high-precision object masks and mark these masks in the image.

[0071] The operator judges whether the generated mask fits perfectly. If not, the operator can manually add or reduce mask points in the operation panel and regenerate until the operator is satisfied, and then start scanning.

[0072] The long video tracking model receives the initial image and the corresponding object mask, and initializes multi-level memory storage including sensory memory, working memory, and long-term memory. For each frame, the model extracts features through memory read operations and generates a mask of the target object. The mask is used to continuously track the position of the target object during the movement of the robotic arm. At the same time, the depth information obtained by the depth camera is used to generate the corresponding point cloud map, marking the position and shape of the object in three-dimensional space.

[0073] Use the convex hull algorithm (Quickhull algorithm) to calculate the outer boundary of the object point cloud. Construct a hypertetrahedron based on the boundary points to form the 3D outline of the point cloud. Calculate the normal vector of the enveloping point cloud. For each point, select a neighborhood point set within a radius from the point cloud. Use PCA to analyze the neighborhood point set and calculate the normal vector of each point.

[0074] Define the target point and posture of the end of the robot arm according to the normal vector and point cloud data to ensure full coverage of the scan. Use the shortest path algorithm such as A* or Dijkstra to plan the movement path of the robot arm. During the path planning process, perform collision detection between the robot arm and the workpiece to avoid collision between the robot arm and the workpiece. Use the GJK (Gilbert-Johnson-Keerthi) algorithm for real-time collision detection.

[0075] After the current part of the inspection is completed, the end of the robot moves to the boundary area of ​​the point cloud, reduces the point cloud repetition rate through the reinforcement learning algorithm, and scans the current view again. Use the deep Q learning (DQN) or Proximal Policy Optimization (PPO) algorithm to optimize the movement path of the robot arm and reduce redundant scanning.

[0076] Repeat the above steps until the point cloud information completely covers the workpiece and the scanning is considered complete.

[0077] The report requirements are analyzed through a large language model, and the scan data and specific requirements entered by the user are combined to generate a personalized test report. The report content includes specific scan data, analysis results, answers to specific questions and suggestions, ensuring the personalization and pertinence of the report.

[0078] like Figure 3As shown, in this embodiment, an automated multifunctional three-dimensional scanning device based on a language image model and intelligent path planning is provided, including: a mechanical arm 1, a depth camera 2, a three-dimensional scanning device 3, an operation panel 4 and a computing unit 5; all of these components are installed on a platform vehicle to ensure the mobility and flexibility of the system. The mechanical arm 1 is installed on the platform vehicle, has multiple degrees of freedom, and can flexibly move and adjust its posture to adapt to workpieces of various shapes and sizes. A depth camera 2 and a three-dimensional scanning device 3 are installed at the end of the mechanical arm 1. The depth camera 2 is used to obtain the depth image of the workpiece in real time, capture the three-dimensional structural information of the workpiece, and provide basic data for subsequent point cloud processing and path planning. The three-dimensional scanning device 3 is responsible for high-precision scanning of the workpiece surface and generating detailed three-dimensional point cloud data. Both devices are connected to the computing unit on the platform vehicle via a cable to transmit the scanning data to the central processing system for analysis and processing. The operation panel 4 is also installed on the platform vehicle, providing a user interface, and the operator can move the platform vehicle, control the mechanical arm 1, and enter task information, such as the name of the workpiece to be scanned and specific report requirements through the operation panel 4. The operation panel 4 is connected to the computing unit 5, and the operator's input information will be transmitted to the computing unit 5 in real time for processing. The computing unit 5 is integrated on the platform vehicle and contains a processor of a large language model, an image language model and a deep learning algorithm for processing data and planning paths. The control system in the computing unit 5 consists of multiple modules, including a speech recognition module, a large language model module, an image processing module, a path planning module and a data analysis module. The speech recognition module is responsible for parsing the operator's voice input and converting it into processable text information; the large language model module is used to understand and process the operator's text input and extract task requirements and report requirements; the image processing module uses the data of the depth camera 2 and the three-dimensional scanning device 3 to generate a three-dimensional point cloud map of the workpiece, and performs corresponding image processing and feature extraction; the path planning module uses the intelligent path planning algorithm and the collision detection algorithm based on the point cloud data and normal vector information to plan the movement path of the robot 1 to ensure the comprehensiveness and safety of the scan; the data analysis module analyzes the scan data and generates a personalized inspection report according to the operator's report requirements.

[0079] The present invention supports voice input, and the operator can complete the workpiece scanning task only through voice commands. The skill requirements for operators are greatly reduced, so that operators without professional technical backgrounds can also use it easily, which improves the usability and promotion of the system. The present invention uses an image language model and a long video tracking model, and the system can automatically identify and track the position of the workpiece without manual intervention. The present invention improves the automation of the scanning process, ensures the continuity and accuracy of the scanning, and reduces the complexity and error rate of manual operation. The present invention can obtain the point cloud data of the workpiece in real time and dynamically adjust the scanning path during the scanning process. The present invention improves the scanning efficiency and accuracy, ensuring that all workpiece surfaces can be fully covered and scanned. The present invention adopts an intelligent path planning algorithm and a reinforcement learning algorithm, and the system can optimize the scanning path and reduce repeated scanning. The present invention improves the scanning efficiency, shortens the scanning time, and reduces the energy and resource consumption of the system. The present invention can automatically detect the completion status of the current scanning area and automatically switch to the next unscanned area. The present invention ensures full coverage scanning of the workpiece surface without manual intervention, further improving the level of automation. The present invention can generate a personalized inspection report based on the specific needs and workpiece characteristics input by the user through a large language model. The report content includes specific scanning data, analysis results, answers to specific questions and suggestions, ensuring that the report is personalized and targeted to meet the specific needs of different customers or tasks.

[0080] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An automated multifunctional 3D scanning system based on language image model and intelligent path planning, characterized in that: include: Voice / text input module, workpiece automatic tracking module, real-time scanning module, automatic path planning and optimization module, scanning area automatic switching module and personalized report generation module; The voice / text input module is used to receive and process the user's voice and text input to determine the description of the workpiece to be scanned and the report requirements; The workpiece automatic tracking module is used to automatically find, identify and select the workpiece to be scanned according to the information input by voice; The real-time scanning module is used to obtain point cloud data and surface scanning data of the workpiece in real time; The automated path planning and optimization module is used to plan the scanning path of the robot arm based on the real-time point cloud data and surface scanning data; The scanning area automatic switching module is used to scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation; The personalized report generation module is used to generate a test report according to the user's specific needs and workpiece characteristics.

2. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The speech / text input module includes a speech / text receiving unit and a speech / text recognition unit; The voice / text receiving unit is used to receive the user's voice and text input information; The speech / text recognition unit is used to input a large language model according to the user's speech and text input information, parse the speech input information, extract the workpiece information and the scanning task requirements.

3. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The workpiece automatic tracking module includes an automatic workpiece scanning unit and an automatic tracking unit; The automatic workpiece scanning unit is used to automatically move the robot arm to find, identify and select the workpiece to be scanned according to the voice and text input information; The automatic tracking unit is used to continuously track the position of the workpiece when the robot arm moves.

4. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The real-time scanning module includes a data acquisition unit and a data processing unit; The data acquisition unit is used to acquire the point cloud data and the surface scanning data of the workpiece in real time; The data processing unit is used to process and update the point cloud data of the current scanning area.

5. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The automatic path planning and optimization module includes an automatic path planning unit and a path optimization unit; The automatic path planning unit is used to plan the scanning path of the robot arm according to the real-time point cloud data and the surface scanning data; The path optimization unit is used to optimize the moving path of the robot arm according to the real-time point cloud data and the surface scanning data.

6. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The scanning area automatic switching module includes an area detection unit and an automatic switching unit; The area detection unit is used to detect the scanning status within the current field of view; The automatic switching unit is used to automatically move the robot arm to the next un-scanned area after the current area is scanned.

7. The automated multifunctional three-dimensional scanning system based on language image model and intelligent path planning as claimed in claim 1, characterized in that: The personalized report generation module includes a report generation unit and a data analysis unit; The report generation unit is used to process specific requirements and artifact characteristics of user input according to the large language model; The data analysis unit is used to generate a test report based on the specific requirements and workpiece characteristics input by the user and in combination with the scanning data.

8. The method of an automated multifunctional three-dimensional scanning system based on a language image model and intelligent path planning according to any one of claims 1 to 7, characterized in that: The following steps are involved: Receive and process user voice and text input to determine the description of the workpiece to be scanned and the reporting requirements; Identify and select the workpiece to be scanned based on the voice and text input information; Acquire point cloud data and surface scanning data of workpieces in real time; Plan the scanning path of the robot arm based on real-time point cloud data; Scan the current field of view according to the scanning path, and switch to the next unscanned area after confirmation; Generate inspection reports based on user-specific requirements and workpiece characteristics.

Images