TBM tunnel surrounding rock three-dimensional structure while drilling acquisition device and method
Patent Information
- Application Number
- CN202610643899.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-11
- Publication Date
- 2026-08-18
AI Technical Summary
通过对现有隧道探测技术的综合评价，现有的围岩感知实现方案在TBM掘进场景下均存在显著的局限性：
[0022] By using the above-mentioned technical means, the data acquisition step length is dynamically matched according to the distance between the TBM tunneling advance and the rock wall, ensuring that the axial overlap rate of the two consecutive acquisitions is stable and not lower than the preset value. This achieves synchronous matching between the acquisition rhythm and the TBM tunneling rhythm, forming a continuous spiral full-circumferential data chain, and solving the problems of disconnection between traditional acquisition and construction procedures and discontinuous data.
Smart Images
Figure CN122589408A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a device and method for obtaining the three-dimensional structure of surrounding rock in a TBM tunnel during drilling. It is applicable to the field of water conservancy and hydropower engineering, specifically rock tunnel engineering and intelligent sensing technology. Background Technology
[0002] In deep-buried, long tunnel projects, the construction efficiency and safety of TBMs (tunnel boring machines) are highly dependent on the real-time sensing of the surrounding rock geological characteristics. During the tunneling process, the dynamic matching of surrounding rock measurement data and support parameters directly determines the construction quality and safety. Through a comprehensive evaluation of existing tunnel detection technologies, existing surrounding rock sensing solutions all have significant limitations in TBM tunneling scenarios:
[0003] First, there is a conflict between limited space and installation requirements. Existing rock sensing solutions employ photogrammetry and 3D laser scanning equipment, both of which are "non-integrated, static sampling" methods. Photogrammetry typically requires a tripod to mount a DSLR camera, while 3D laser scanners require a spacious, open area to ensure circumferential coverage of the laser pulses. However, the internal mechanical structure of a TBM is extremely compact, with the main beam, propulsion system, and segment installer occupying the majority of the space. Furthermore, equipment can usually only be deployed on specific structural components such as the main beam or the base of the anchor drilling rig. This makes it impossible to deploy traditional non-contact detection equipment in real-time during tunneling intervals, making it difficult to adapt to the extremely limited physical space of a TBM.
[0004] Second, the data suffers from limited modality and lack of dimension. Existing technologies often rely on acquiring data from a single modality, leading to incomplete geological assessments. While single image data offers rich texture, it primarily provides two-dimensional planar information, making it difficult to accurately reflect the three-dimensional morphology, undulation, and spatial geological structure of rock masses. Conversely, while single three-dimensional point cloud data boasts high geometric spatial accuracy, the lack of color semantics and texture information makes it difficult for geological engineers to effectively identify lithological boundaries, seepage points, and detailed fracture features. This limitation in modal dimension directly restricts the ability to quantitatively analyze complex surrounding rock structures.
[0005] Third, the data acquisition process is discontinuous and disconnected from the operational procedures. Most existing non-contact measurement methods operate independently of construction equipment and lack standardized operating procedures. In the environment of continuous TBM tunneling, current technologies cannot reliably acquire standardized data with high overlap rates from multiple directions within the extremely short intervals between tunneling steps. The processing is often offline, with complex and time-consuming algorithms, making it difficult to achieve real-time perception and dynamic feedback during tunneling and failing to support the real-time requirements of intelligent decision-making in tunnel construction. Summary of the Invention
[0006] The technical problem to be solved by the present invention is to provide a device and method for obtaining the three-dimensional structure of the surrounding rock of a TBM tunnel during drilling, in view of the above-mentioned problems.
[0007] The technical solution adopted in this invention is: a device for obtaining three-dimensional structure of surrounding rock in a TBM tunnel during drilling, comprising: The rotating detection unit has at least one RGB-D binocular camera module, which is installed on the base of the TBM anchor drilling rig. It is used to rotate with the base to complete the circumferential scanning of the surrounding rock of the tunnel and collect multimodal raw data of the surrounding rock. The auxiliary observation unit is fixedly installed on the TBM main beam or side support structure to provide a global spatial registration benchmark for the rotating detection unit and assist in the spatial positioning of the acquired data. The edge computing terminal communicates with the rotating detection unit, auxiliary observation unit, TBM control system, and TBM navigation system. It is used to receive tunneling condition signals from the TBM control system. During TBM tunneling transitions or downtime, it collects surrounding rock data through the rotating detection unit and combines the reference data from the auxiliary observation unit and the positioning information from the TBM navigation system to complete the spatial calibration and coordinate transformation of the collected data.
[0008] By employing the aforementioned technical means, the traditional method of external tripod mounting for detection equipment is abandoned. The core detection unit is directly integrated into the TBM's built-in anchor drilling rig base, and the auxiliary observation unit is integrated into the TBM's main beam structure. This eliminates the need for additional construction space and fundamentally resolves the conflict between traditional equipment and the TBM's limited and compact space. Through the linkage between the edge computing terminal and the TBM control and navigation systems, deep coupling between surrounding rock data acquisition and TBM construction procedures is achieved. Data acquisition is only performed during tunneling transitions, without interfering with the TBM's main tunneling process, thus adapting to the rhythm of TBM's rapid and continuous tunneling. Simultaneously, the RGB-D binocular camera module synchronously acquires color images and depth point cloud data of the surrounding rock, forming a complementary multimodal data source of two-dimensional texture semantics and three-dimensional geometric space, solving the problems of single data modality and lack of dimension in traditional technologies.
[0009] As a preferred option, it also includes: An adaptive feedback unit is installed on the base of the anchor drilling rig and is used to collect real-time radial distance data from the center of the base to the surface of the surrounding rock. The edge computing terminal is communicatively connected to the adaptive feedback unit and is used to dynamically adjust the acquisition parameters of the rotating detection unit based on the radial distance data from the adaptive feedback unit.
[0010] By employing the aforementioned technical means, the distance data between the sensor and the rock wall is acquired in real time through the adaptive feedback unit, providing real-time feedback for the dynamic adjustment of the collected parameters. This solves the problems of insufficient overlap rate of fixed parameter acquisition and point cloud splicing discontinuity caused by the undulation of the rock wall, ensuring the continuity and standardization of the collected data, and laying the data foundation for subsequent high-precision 3D reconstruction.
[0011] As a preferred embodiment, the RGB-D binocular camera module integrates an infrared laser supplementary lighting device to provide stable lighting conditions in low-light or no-light conditions inside the tunnel.
[0012] The above-mentioned technical means have solved the problems of image color distortion and insufficient contrast caused by incandescent lamp supplementary lighting in the low-light environment of tunnels, improved the quality of the original image from the source of data acquisition, and reduced the impact of environmental interference on perception accuracy.
[0013] As a preferred embodiment, the rotating detection unit has three sets of RGB-D binocular camera modules, which are evenly arranged around the base of the TBM anchor drilling rig at an angle of 120° to each other.
[0014] Through the above-mentioned technical means, the three camera modules are evenly distributed at 120°, which can achieve 360° full circumference coverage of the tunnel surrounding rock without blind spots during the rotation of the anchor drilling rig base. At the same time, the complementary field of view of multiple cameras improves the efficiency and data integrity of single-loop acquisition, and adapts to the short acquisition window requirements of TBM step-change intervals.
[0015] As a preferred embodiment, the auxiliary observation unit has at least one set of fixed depth cameras.
[0016] By employing the aforementioned technical means, a stable global spatial reference is provided by a depth camera fixed to the main beam of the TBM. This allows for precise calibration of the relative pose of the rotating detection unit with respect to the TBM head, eliminating the cumulative pose error of the rotating detection unit during the drilling rig's movement. This ensures the accuracy of the spatial positioning of the acquired data and provides a unified spatial reference for multi-frame data stitching and full-tunnel data tracing.
[0017] A method for obtaining the three-dimensional structure of the surrounding rock of a TBM tunnel based on the aforementioned drilling acquisition device includes: S1. Receive tunneling condition signals from the TBM control system and trigger the drilling data acquisition program during TBM tunneling step changes or shutdown intervals. S2. The radial distance data from the center of the sensing base to the surrounding rock wall is collected in real time through the adaptive feedback unit; S3. Based on real-time radial distance data, dynamically calculate the rotation step angle of the anchor drilling rig base and control the overlap rate of adjacent acquisition frames in the circumferential direction to be no less than the first preset value. S4. Real-time acquisition of TBM tunneling advance data, combined with the effective axial field of view of the rotating detection unit sensing hardware, adaptively calculates the axial scanning step length, and controls the tunnel axial overlap rate of the two acquisitions to be no less than the second preset value. S5. Drive the anchor drilling rig base to complete circumferential rotation according to the adaptively calculated rotation step angle and axial scanning step length, and simultaneously collect color images and depth point cloud data of the surrounding rock. S6. The collected multimodal data, combined with the reference data of the auxiliary observation unit and the positioning information of the TBM navigation system, are correlated to the tunnel global geographic coordinate system, and output circumferential color point cloud data of the surrounding rock with spatial index.
[0018] Through the aforementioned technical means, adaptive acquisition of TBM tunnel surrounding rock data was achieved, deeply coupled with the TBM construction process. Data acquisition was only performed during the intervals between tunneling steps, without affecting the main construction period. By dynamically adjusting the rotation step angle through real-time radial distance feedback and adaptively matching the axial scanning step length through tunneling footage, the radial overlap rate and axial coincidence rate of the acquired data were forcibly guaranteed. This solved the problems of data discontinuity and insufficient splicing accuracy caused by traditional fixed parameter acquisition, and achieved continuous and standardized acquisition of surrounding rock data. At the same time, by associating global coordinates, each set of data was assigned a unique spatial index, enabling the traceability and splicing of the acquired data, and providing a high-quality standardized data source for subsequent three-dimensional reconstruction of the surrounding rock structure and geological identification.
[0019] As a preferred embodiment, step S3 includes: ; in, To rotate the step angle, The first preset value, For the image sensor width, Focal length For radial distance data, Let be the radius of rotation of the base.
[0020] By using the above-mentioned technical means, the stepping angle is dynamically adjusted according to the real-time radial distance to ensure that the radial overlap rate of adjacent acquisition frames at different distances on the rock wall is stable and not lower than the preset value. This solves the problem of acquisition overlap rate fluctuation caused by rock wall undulation and provides mathematical guarantee for seamless stitching of full-circumferential point clouds.
[0021] As a preferred embodiment, step S4 includes: ; in, This is the axial scan step size. For radial distance data, The effective field of view along the axis, This is the second preset value.
[0022] By using the above-mentioned technical means, the data acquisition step length is dynamically matched according to the distance between the TBM tunneling advance and the rock wall, ensuring that the axial overlap rate of the two consecutive acquisitions is stable and not lower than the preset value. This achieves synchronous matching between the acquisition rhythm and the TBM tunneling rhythm, forming a continuous spiral full-circumferential data chain, and solving the problems of disconnection between traditional acquisition and construction procedures and discontinuous data.
[0023] The beneficial effects of this invention are: This invention abandons the traditional mode of externally mounted detection equipment, and directly integrates the rotating detection unit and adaptive feedback unit into the anchor drilling rig base of the TBM, and integrates the auxiliary observation unit into the main beam structure of the TBM. It does not require additional construction site, and is fully adapted to the extremely compact physical space inside the TBM. At the same time, it reuses the rotating drive mechanism of the anchor drilling rig as the scanning execution unit, and does not require the addition of independent rotating equipment.
[0024] This invention uses an RGB-D binocular camera module to simultaneously acquire color images and depth point cloud data of the surrounding rock. It also obtains two-dimensional semantic information on the lithology, texture, and seepage traces of the surrounding rock, as well as three-dimensional geometric information on the dip angle, dip direction, and undulation of the surrounding rock. This forms a complementary multimodal data source, which makes up for the core defects of single images lacking three-dimensional spatial information and single point clouds lacking texture and semantic information. This provides a data foundation for the comprehensive and accurate identification of the three-dimensional structure of the surrounding rock.
[0025] This invention links the edge computing terminal with the TBM control system to strictly lock the acquisition window during the TBM tunneling step-changing interval, without interfering with the main construction process, and adapts to the rhythm of rapid and continuous TBM tunneling. Through the adaptive acquisition logic of real-time feedback from laser ranging, the rotation step angle and axial scanning step length are dynamically adjusted to ensure that the radial overlap rate and axial coincidence rate are not lower than the preset values, thereby realizing continuous and standardized acquisition of surrounding rock data and solving the problems of discontinuous data and insufficient real-time performance of traditional offline acquisition.
[0026] This invention enables non-contact surrounding rock sensing throughout the entire process. Technicians do not need to enter unsupported dangerous areas to conduct manual geological sketching, which greatly reduces construction safety risks. The collected surrounding rock data can be spatially calibrated and coordinate transformed in real time, and can be directly used for the reconstruction of the three-dimensional structure of the surrounding rock and quality classification. This significantly shortens the geological decision-making cycle and provides real-time data support for TBM tunneling parameter adjustment and support scheme optimization. Attached Figure Description
[0027] Figure 1 This is a schematic diagram of the drilling data acquisition device in the embodiment.
[0028] Figure 2 This is a schematic diagram of the arrangement structure of the rotating detection unit in the embodiment.
[0029] Figure 3 This is a schematic diagram of the hardware structure of the edge computing terminal in the embodiment.
[0030] 1. RGB-D binocular camera module; 2. Laser rangefinder sensor; 3. TBM anchor drill base; 4. Edge computing terminal; 401. Processor; 402. Memory; 403. Input / output interface; 404. Communication interface; 405. Sensor interface; 406. Power interface. Detailed Implementation
[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of this invention are only for explaining this invention and are not intended to limit this invention.
[0032] This embodiment is a drilling-while-drilling device for obtaining the three-dimensional structure of the surrounding rock in a TBM tunnel, including a rotating detection unit, an adaptive feedback unit, an auxiliary observation unit, and an edge computing terminal.
[0033] In this example, the rotating detection unit is the core data acquisition module of the device. It is integrated and installed on the TBM anchor drilling rig base. The anchor drilling rig base is connected to the rotating drive mechanism of the TBM, which can drive the rotating detection unit to complete a 360° circumferential rotation scan.
[0034] In this embodiment, the rotating detection unit is equipped with three sets of high-protection-level RGB-D binocular camera modules. The three modules are evenly distributed around the base of the anchor drilling rig at a 120° angle to each other, achieving 360° all-around, blind-spot-free acquisition of surrounding rock data. Each RGB-D binocular camera module integrates an infrared laser illumination device, which is equipped with a PWM dimming unit and can communicate with the edge computing terminal. It dynamically adjusts the illumination intensity according to the brightness of the tunnel environment, providing stable and distortion-free lighting conditions in low-light or no-light conditions inside the tunnel, and acquiring color images and depth point cloud data of the surrounding rock surface.
[0035] The RGB-D binocular camera module uses industrial-grade equipment with high protection level, which can withstand the harsh construction environment of high dust, high humidity and strong vibration in the tunnel. The raw data it collects includes color image data and depth point cloud data of the surrounding rock surface, forming a complementary multimodal data source of two-dimensional texture semantics and three-dimensional geometric space.
[0036] In this embodiment, the adaptive feedback unit is fixedly installed on the anchor drilling rig base. The core of the unit is an industrial-grade laser rangefinder. The laser rangefinder is connected to the edge computing terminal via RS485 / RS232 serial port protocol. It is used to measure the radial distance data from the center of the anchor drilling rig base to the surrounding rock wall in real time. This distance data serves as the core feedback quantity for the adaptive acquisition and control while drilling. It is transmitted to the edge computing terminal to dynamically adjust the acquisition parameters such as the rotation step angle of the rotating detection unit.
[0037] In this embodiment, the auxiliary observation unit consists of at least one fixed depth camera, which is fixedly installed on the TBM main beam or side support structure, without occupying additional tunneling work space, and is communicatively connected to the edge computing terminal. The fixed depth camera's field of view covers the entire range of motion of the rotating detection unit, and can stably acquire the appearance features of the rotating detection unit and the surrounding TBM structural features, providing a global spatial registration benchmark for the rotating detection unit, assisting in completing the pose calibration of the rotating detection unit relative to the TBM head, eliminating cumulative pose errors during the acquisition process, and ensuring the accuracy of spatial positioning of the acquired data.
[0038] In this example, the edge computing terminal is connected to the rotating detection unit, the auxiliary observation unit, the TBM control system, and the TBM navigation system. It is used to receive tunneling condition signals from the TBM control system. During TBM tunneling transitions or downtime, the rotating detection unit collects surrounding rock data and combines the reference data from the auxiliary observation unit with the positioning information from the TBM navigation system to complete the spatial calibration and coordinate transformation of the collected data.
[0039] The edge computing terminal is the core control and computing unit of this device. It adopts an industrial-grade high-performance edge computing terminal, equipped with a multi-core CPU and GPU / AI neural network accelerator, which can meet the computing power requirements for real-time data processing and algorithm inference in the tunnel. The edge computing terminal achieves communication connections between various functional modules through a system bus, specifically including: Processor: Used to execute program instructions stored in memory to complete the entire process of acquisition control, data processing, spatial calibration, algorithm inference, etc. Memory: Employs non-transitory computer-readable storage media, including high-speed RAM and SSD non-volatile memory, to store acquired raw data, program instructions, deep learning model weights, digital twin assets, and other content; Input / output interface: Physically connected to the TBM control system via CAN / Modbus industrial bus to receive tunneling condition signals and output early warning signals; Communication interface: It communicates with the TBM navigation system via industrial Ethernet to obtain navigation data such as the absolute geographic coordinates, attitude, and tunneling progress of the TBM head in real time; Sensor interface: It communicates with the RGB-D binocular camera module of the rotating detection unit, the laser rangefinder of the adaptive feedback unit, and the fixed depth camera of the auxiliary observation unit via the GigEVision / USB3.0 protocol to realize the synchronous acquisition and transmission of multimodal data; Power interface: Connects to the industrial power supply system (DC24V / 48V) of the TBM, and integrates a voltage regulator and filter module and surge protection circuit, which can operate stably in the strong electromagnetic interference environment caused by the start and stop of the high-power motor of the TBM.
[0040] Phase 1: System Integration Initialization and Multimodal Calibration
[0041] This stage aims to establish the spatial topology between the sensing system hardware and the TBM host, and to complete the standardized mapping of sensor parameters. This embodiment adopts the following steps:
[0042] S0-1. Hardware Coordinated Deployment and Coordinate System Anchoring: During the equipment arrival phase, three sets of high-protection-level RGB-D binocular camera modules are installed at a 120° angle to each other on the TBM anchor drilling rig base, and fixed auxiliary observation units are deployed on the TBM main beam as global reference nodes. High-precision measuring instruments are used to determine the three-dimensional offset vector of the measurement module center relative to the shield machine head center. and installation offset angle The absolute geographic coordinates and attitude of the TBM nose are obtained in real time through the communication interface with the TBM navigation system. Based on this, a transformation matrix between the local coordinate system of the measurement module and the absolute geographic coordinate system of the tunnel is established. To ensure that the sensed data can be automatically converted into absolute geographic coordinates in real time during the tunneling process, the conversion formula is as follows: ; Where, transformation matrix It is the real-time absolute coordinate translation matrix of the TBM head. With attitude rotation matrix Sequence (rotation angle around axis) The composite matrix is formed by the sensor's preset relative offset and the sensor's own offset. ;
[0043] S0-2, RGB-D Camera Intrinsic and Extrinsic Parameter Calibration and Modal Registration: For the RGB-D module, intrinsic parameter calibration of the color and depth sensors is performed separately to obtain the focal length ( ), principal point coordinates ( ) and distortion coefficient ( Establish the intrinsic parameter matrix K of the RGB-D camera. Perform stereo calibration to calculate the transformation matrix between the two sensors. This achieves precise registration of different modal data at the pixel level. Subsequently, it is combined with the tunnel radius... Radius of rotation of the base Establish from pixel coordinates and pixel depth value To a point in three-dimensional space The inverse projection mapping model lays the geometric foundation for subsequent feature fusion.
[0044] ;
[0045] S0-3, Initial Pose Calculation and Attitude Initialization: Based on the original quaternion pose data recorded by the module. The conversion formula is used to solve it into Euler angles, i.e., heading angles. Pitch angle and roll angle : ; By accurately calibrating the initial azimuth and tilt of the sensing module in the mechanical structure of the tunneling machine using the calculated Euler angles, it is ensured that each step of the sensor during the rotational scanning process can achieve a high degree of adaptive matching with the axial tunneling advance parameters of the TBM.
[0046] Phase Two: Feedback-based Adaptive Rotary Acquisition While Drilling
[0047] The method for obtaining the three-dimensional structure of the surrounding rock during drilling based on the TBM tunnel surrounding rock three-dimensional structure acquisition device in this embodiment includes:
[0048] S1. The edge computing terminal receives tunneling status signals from the TBM control system in real time. When it detects that the TBM has entered a tunneling intermittent state, it automatically triggers the drilling data acquisition program. If a TBM tunneling restart signal is received during the acquisition process, the acquisition is immediately terminated and the anchor drilling rig base is reset to ensure that the acquisition operation does not interfere with the TBM's main tunneling process. S2. The radial distance data from the center of the anchor drilling rig base to the surrounding rock wall is acquired in real time through the laser rangefinder of the adaptive feedback unit. The data is transmitted to the edge computing terminal. S3, the edge computing terminal uses real-time radial distance data... Dynamically calculate the rotation step angle of the anchor drilling rig base. Control the circumferential overlap rate of adjacent acquisition frames to be no less than a first preset value. (In this embodiment) The calculation formula is: ; In the formula, For the image sensor width, For camera focal length, The radius of rotation of the base is given. This formula is based on the geometric relationship between the visual field of view and the object distance. The farther away from the rock wall, the smaller the coverage angle of a single frame image, and the rotation step angle decreases synchronously, ensuring that the radial overlap rate of all acquired frames in the circumference is stable and meets the standard, avoiding point cloud stitching discontinuities caused by rock wall undulations. S4, the edge computing terminal acquires the tunneling advance data of the TBM navigation system in real time, combined with the axial effective field of view of the RGB-D binocular camera. Adaptive calculation of axial scan step size The tunnel axial overlap rate between the two consecutive data acquisitions is controlled to be no less than the second preset value. (In this embodiment) The calculation formula is: ;
[0049] This formula achieves dynamic matching between the axial step length and the TBM tunneling advance, ensuring that the axial overlap rate of the two consecutive acquisitions meets the standard, forming a continuous spiral full-circumferential data chain. S5: The edge computing terminal drives the anchor drilling rig base to complete a 360° circumferential step rotation according to the adaptively calculated rotation step angle and axial scanning step length. Through hardware synchronous trigger signals, all RGB-D binocular camera modules are controlled to synchronously expose at the same trigger point, acquiring color images and depth point cloud data of the surrounding rock, ensuring the spatiotemporal consistency of multimodal data. S6: Using the composite transformation matrix established in Phase 1, combined with the global reference data of the auxiliary observation unit and the positioning information of the TBM navigation system, each set of acquired multimodal data is correlated in real time to the tunnel's global geographic coordinate system, generating full-circumferential true-color point cloud data of the surrounding rock with a unique spatial index.
Claims
1. A device for obtaining three-dimensional structure of surrounding rock in a TBM tunnel during drilling, characterized in that, include: The rotating detection unit has at least one RGB-D binocular camera module, which is installed on the base of the TBM anchor drilling rig. It is used to rotate with the base to complete the circumferential scanning of the surrounding rock of the tunnel and collect multimodal raw data of the surrounding rock. The auxiliary observation unit is fixedly installed on the TBM main beam or side support structure to provide a global spatial registration benchmark for the rotating detection unit and assist in the spatial positioning of the acquired data. The edge computing terminal communicates with the rotating detection unit, auxiliary observation unit, TBM control system, and TBM navigation system. It is used to receive tunneling condition signals from the TBM control system. During TBM tunneling transitions or downtime, it collects surrounding rock data through the rotating detection unit and combines the reference data from the auxiliary observation unit and the positioning information from the TBM navigation system to complete the spatial calibration and coordinate transformation of the collected data.
2. The TBM tunnel surrounding rock three-dimensional structure acquisition device according to claim 1, characterized in that, Also includes: An adaptive feedback unit is installed on the base of the anchor drilling rig and is used to collect real-time radial distance data from the center of the base to the surface of the surrounding rock. The edge computing terminal is communicatively connected to the adaptive feedback unit and is used to dynamically adjust the acquisition parameters of the rotating detection unit based on the radial distance data from the adaptive feedback unit.
3. The TBM tunnel surrounding rock three-dimensional structure acquisition device according to claim 1, characterized in that, The RGB-D binocular camera module integrates an infrared laser supplementary lighting device to provide stable lighting conditions in low-light or no-light conditions inside the tunnel.
4. The TBM tunnel surrounding rock three-dimensional structure acquisition device according to claim 1, characterized in that, The rotating detection unit has three RGB-D binocular camera modules, which are evenly arranged around the base of the TBM anchor drilling rig at an angle of 120° to each other.
5. The TBM tunnel surrounding rock three-dimensional structure acquisition device according to claim 1, characterized in that, The auxiliary observation unit has at least one set of fixed depth cameras.
6. A method for obtaining the three-dimensional structure of the surrounding rock in a TBM tunnel based on the drilling acquisition device according to any one of claims 1 to 5, characterized in that, include: S1. Receive tunneling condition signals from the TBM control system and trigger the drilling data acquisition program during TBM tunneling step changes or shutdown intervals. S2. The radial distance data from the center of the sensing base to the surrounding rock wall is collected in real time through the adaptive feedback unit; S3. Based on real-time radial distance data, dynamically calculate the rotation step angle of the anchor drilling rig base and control the overlap rate of adjacent acquisition frames in the circumferential direction to be no less than the first preset value. S4. Real-time acquisition of TBM tunneling advance data, combined with the effective axial field of view of the rotating detection unit sensing hardware, adaptively calculates the axial scanning step length, and controls the tunnel axial overlap rate of the two consecutive acquisitions to be no less than the second preset value. S5. Drive the anchor drilling rig base to complete circumferential rotation according to the adaptively calculated rotation step angle and axial scanning step length, and simultaneously collect color images and depth point cloud data of the surrounding rock. S6. The collected multimodal data, combined with the reference data of the auxiliary observation unit and the positioning information of the TBM navigation system, are correlated to the tunnel global geographic coordinate system, and output circumferential color point cloud data of the surrounding rock with spatial index.
7. The method for obtaining the three-dimensional structure of the surrounding rock in a TBM tunnel during drilling, as described in claim 6, is characterized in that... Step S3 includes: ； in, To rotate the step angle, The first preset value, For the image sensor width, Focal length For radial distance data, Let be the radius of rotation of the base.
8. The method for obtaining the three-dimensional structure of the surrounding rock in a TBM tunnel during drilling, as described in claim 6, is characterized in that... Step S4 includes: ； in, This is the axial scan step size. For radial distance data, The effective field of view along the axis, This is the second preset value.