CN-122016810-A - System and method for detecting wear of tool on face of underground shield tunneling machine

CN122016810ACN 122016810 ACN122016810 ACN 122016810ACN-122016810-A

Abstract

The embodiment of the application provides a system and a method for detecting wear of a tool on a face of an underground shield tunneling machine, and relates to the technical field of engineering machinery detection. The system comprises an unmanned plane, a detection component, a transmission component and a data processing device. The unmanned aerial vehicle can carry the detection assembly to perform image acquisition on the tunnel face cutter of the shield tunneling machine in the underground construction area so as to obtain detection image data. The transmission component transmits the detection image data and the control instruction through the optical fiber transmission line, so that the data processing equipment can identify the cutter head abrasion area and the defect information in the detection image data by using a cutter abrasion identification algorithm, and a structural detection result is generated. The system can carry out real-time image transmission by combining an unmanned aerial vehicle with an optical fiber, adopts MAML element learning and improved Swin-transform dual-stage architecture, adapts to small sample scenes and 1080p cutter abrasion real-time detection, can adapt to high-dust construction area environments in short distance and limited space, and improves the safety, adaptability and reliability of the detection process.

Inventors

WEI YINGJIE
Shao Zhenxin
ZHANG HANLIU
CHEN PENG
SHU JICHENG
DU GUIXIN

Assignees

中国地质大学（北京）
中铁十四局集团有限公司

Dates

Publication Date: 20260512
Application Date: 20260212

Claims (10)

1. An underground shield constructs machine face cutter wearing and tearing detecting system, characterized in that, the system includes: the unmanned aerial vehicle comprises a fusion positioning unit, wherein the fusion positioning unit is configured to calculate positioning data of the unmanned aerial vehicle through an extended Kalman filtering fusion algorithm; the detection assembly is arranged on the unmanned aerial vehicle and comprises a detection camera, a camera holder and a narrow-spectrum infrared light supplementing module, wherein the detection camera is arranged on the camera holder and is provided with an anti-fog and anti-fouling coated lens and a dust cover; the system comprises a detection module, a transmission assembly, a control module and a control module, wherein the detection module is configured to transmit the detection image data and the control command, the transmission assembly comprises an optical fiber transmission line and an optical transceiver, the optical fiber transmission line comprises a detection end and a control end, and the detection end of the optical fiber transmission line is connected with the unmanned aerial vehicle and the detection module through the optical transceiver; The data processing device is connected with the control end of the optical fiber transmission line through the optical terminal machine and is configured to: acquiring detection image data, wherein the detection image data comprises a cutter image; Invoking a cutter wear recognition algorithm, wherein the cutter wear recognition algorithm comprises a defect detection model, and the defect detection model is a neural network model obtained by training sample defect data; the defect detection model adopts MAML (maximum intensity markup language) element learning and improved Swin-converter double-stage training architecture, comprises a dark channel prior defogging module and an improved median filtering module, is used for removing image blurring and noise caused by dust, and comprises sample images containing typical defects of a cutter and defect labels of the sample images, wherein the defect labels comprise defect type labels and defect grade labels; identifying a cutter wear area and defect information in the detected image data using the cutter wear identification algorithm, the defect information including a defect type and a defect level output by the defect detection model based on the cutter image; And generating a structural detection result according to the cutter head abrasion area and the defect information, wherein the structural detection result comprises a defect position image, a defect coordinate, a defect size and the defect grade.
2. The system of claim 1, wherein the fused positioning unit comprises an inertial navigation module configured to detect a pose of the drone, an optical flow sensor configured to detect a position of the drone by analyzing a pattern of motion of pixels between successive image frames, and a laser ranging module configured to detect the position of the drone by laser ranging, the data processing device further configured to: acquiring positioning data acquired by the fusion positioning unit, wherein the positioning data comprises original detection data of the fusion positioning unit, the position of the unmanned aerial vehicle and the posture of the unmanned aerial vehicle; constructing a construction area model according to the positioning data and the detection image data; Determining an undetected area according to the construction area model and the detected image data; Generating detection control instructions based on the undetected areas, wherein the detection control instructions comprise a planned path for controlling the unmanned aerial vehicle to fly; and sending the detection control instruction to the unmanned aerial vehicle so that the unmanned aerial vehicle flies according to the planned path.
3. The system of claim 2, further comprising a display terminal coupled to the data processing device, wherein the data processing device is further configured to: generating an interface rendering background based on the construction area model and the detection image data; generating an algorithm marking result based on the construction area model and the structural detection result, wherein the algorithm marking result comprises a visual identification adopting color coding; the algorithm marking result is overlapped to the interface rendering background in real time to generate a result display interface; And displaying the result display interface through the display terminal.
4. The system of claim 3, wherein the display terminal is a virtual reality device, and wherein the data processing device is further configured to: acquiring an interaction instruction input by a user through the virtual reality equipment, wherein the interaction instruction comprises at least one of a review instruction, a correction instruction and a confirmation instruction; recording detection result information corresponding to the structured detection result according to the interaction instruction, wherein the detection result information comprises a confirmed defect position image, a defect coordinate and a defect grade; And generating a structured detection report according to the detection result information.
5. The system of claim 4, wherein the virtual reality device comprises a pose sensor configured to detect real-time pose data of a user, the data processing device further configured to: acquiring the real-time pose data and the historical pose data, wherein the historical pose data is obtained by recording the real-time pose data in the wearing process of a user; Calculating action amplitude according to the real-time pose data and the historical pose data, wherein the action amplitude is obtained by combining the construction area model calculation based on a weighted summation result of the dimension differences of the real-time pose data and the historical pose data; if the action amplitude is larger than a preset amplitude threshold, generating a visual field following instruction according to the real-time pose data, wherein the visual field following instruction is used for controlling the unmanned aerial vehicle to fly according to the real-time pose data; And sending the visual field following instruction to the unmanned aerial vehicle.
6. The system of claim 1, wherein the transmission component further comprises a graphics encoder coupled to the detection camera, the data processing device further configured to: Defining a region of interest based on the structural detection result, wherein the region of interest is a defect region with the number of samples of the cutter image smaller than a number threshold, or a defect region detected for the first time, or a defect region with the defect grade higher than a preset grade threshold in the cutter image; Determining a non-region of interest according to the region of interest, wherein the non-region of interest is a region outside the region of interest in the tool image; Generating a coding optimization instruction, wherein the coding optimization instruction comprises a first coding mode aiming at the region of interest and a second coding mode aiming at the non-region of interest; And sending the coding optimization instruction to the image transmission encoder so that the image transmission encoder can execute image coding on the original image data acquired by the detection camera according to the coding optimization instruction to generate detection image data.
7. The system of claim 6, wherein the data processing device is further configured to: acquiring a transmission data stream corresponding to the detection image data; Determining a first decoding mode according to the first coding mode, and determining a second decoding mode according to the second coding mode; decoding a key region image in the transmission data stream by using the first decoding mode according to the region of interest; Decoding non-critical area images in the transmission data stream by using the second decoding mode according to the non-interested area; and stitching the critical area image and the non-critical area image to generate the tool image.
8. The system of claim 1, wherein the data processing device is further configured to: Responding to a remote control instruction input by a user, and reading flight control parameters from the remote control instruction; Acquiring a detection strategy, wherein the detection strategy comprises a cutter detection position and a flight limiting condition; Correcting the flight control parameters based on the detection strategy, and generating flight control instructions based on the corrected flight control parameters; And sending the flight control instruction to the unmanned aerial vehicle so as to control the unmanned aerial vehicle to fly to the cutter detection position according to the flight limiting condition.
9. The system of claim 1, wherein the data processing device is further configured to: extracting a plurality of tool images from the detection image data according to detection time; Acquiring a plurality of positioning data acquired at the same detection time as the cutter image; stitching a plurality of the tool images based on the positioning data to generate an overview image; and taking the overview image as a rendering background, and generating the defect position image by superposing the cutter head abrasion area and the defect information in real time.
10. A method for detecting wear of a tool on a face of an underground shield tunneling machine, which is applied to the system of any one of claims 1 to 9, and comprises the following steps: acquiring detection image data, wherein the detection image data comprises a cutter image; Invoking a cutter wear recognition algorithm, wherein the cutter wear recognition algorithm comprises a defect detection model, and the defect detection model is a neural network model obtained by training sample defect data; the defect detection model adopts MAML (maximum intensity markup language) element learning and improved Swin-converter double-stage training architecture, comprises a dark channel prior defogging module and an improved median filtering module, is used for removing image blurring and noise caused by dust, and comprises sample images containing typical defects of a cutter and defect labels of the sample images, wherein the defect labels comprise defect type labels and defect grade labels; identifying a cutter wear area and defect information in the detected image data using the cutter wear identification algorithm, the defect information including a defect type and a defect level output by the defect detection model based on the cutter image; And generating a structural detection result according to the cutter head abrasion area and the defect information, wherein the structural detection result comprises a defect position image, a defect coordinate, a defect size and the defect grade.

Description

System and method for detecting wear of tool on face of underground shield tunneling machine Technical Field The application relates to the technical field of engineering machinery detection, in particular to a system and a method for detecting wear of a tool on a face of an underground shield tunneling machine. Background The shield cutter is used as a core component of a shield tunneling process, and the abrasion state of the shield cutter directly influences the construction efficiency and the engineering cost. In order to ensure the construction efficiency, the shield cutter needs to be detected, maintained and replaced regularly. In special engineering practice, the time for checking, maintaining and replacing the cutter is about 1/4-1/3 of the total tunneling time. The shield cutter detection can find out the defects of cutter abrasion, mud cake formation and the like in time, and is used for avoiding the occurrence of chain damage, reducing the maintenance cost and delaying the construction period. The shield cutter detection can rely on a manual detection method, namely a detection person enters a construction area to carry out close-range visual or simple tool measurement, and the cutter abrasion condition is judged according to visual or tool measurement results. Because the shield machine construction area is an area with high risk and limited space, potential safety hazards exist, and the efficiency of manual detection is low and the subjectivity is strong. Therefore, instead of manual tool detection, a robot may be used, that is, a wheeled or crawler robot is used to move in a construction area, and visual image acquisition is performed on the tool based on an image acquisition device such as a camera, so that the tool wear condition is determined through image analysis processing. However, because the construction area environment of the shield machine is muddy, complex and limited in space, the detection capability of the robot is limited, the large-diameter cutterhead is difficult to cover on the whole surface, and particularly, the problem of difficult detection on the top and edge areas exists, so that the safety, adaptability and reliability of cutter detection are reduced. Disclosure of Invention In view of the above, the embodiment of the application provides a system and a method for detecting the wear of a tool on the face of an underground shield tunneling machine, which are used for solving the problems of low safety, adaptability and reliability of the detection of the tool of the shield tunneling machine. According to a first aspect of the present application, there is provided an underground shield tunneling machine face cutter wear detection system, the system comprising: the unmanned aerial vehicle comprises a fusion positioning unit, wherein the fusion positioning unit is configured to calculate positioning data of the unmanned aerial vehicle through an extended Kalman filtering fusion algorithm; the detection assembly is arranged on the unmanned aerial vehicle and comprises a detection camera, a camera holder and a narrow-spectrum infrared light supplementing module, wherein the detection camera is arranged on the camera holder and is provided with an anti-fog and anti-fouling coated lens and a dust cover; the system comprises a detection module, a transmission assembly, a control module and a control module, wherein the detection module is configured to transmit the detection image data and the control command, the transmission assembly comprises an optical fiber transmission line and an optical transceiver, the optical fiber transmission line comprises a detection end and a control end, and the detection end of the optical fiber transmission line is connected with the unmanned aerial vehicle and the detection module through the optical transceiver; The data processing device is connected with the control end of the optical fiber transmission line through the optical terminal machine and is configured to: acquiring detection image data, wherein the detection image data comprises a cutter image; Invoking a cutter wear recognition algorithm, wherein the cutter wear recognition algorithm comprises a defect detection model, and the defect detection model is a neural network model obtained by training sample defect data; the defect detection model adopts MAML (maximum intensity markup language) element learning and improved Swin-converter double-stage training architecture, comprises a dark channel prior defogging module and an improved median filtering module, is used for removing image blurring and noise caused by dust, and comprises sample images containing typical defects of a cutter and defect labels of the sample images, wherein the defect labels comprise defect type labels and defect grade labels; identifying a cutter wear area and defect information in the detected image data using the cutter wear identification algorithm, the defect information including a defect type an