Search

CN-121994135-A - Tunnel blasting block degree in-situ information instant detection method based on video measurement

CN121994135ACN 121994135 ACN121994135 ACN 121994135ACN-121994135-A

Abstract

The invention discloses a tunnel blasting block degree in-situ instant detection method based on video measurement, which is realized by installing cameras and high-intensity light sources on a cab, a movable arm and a bucket rod of a slag extractor and matching the cameras and the high-intensity light sources; the method comprises the steps of completing system calibration by using azimuth sensors, calibration plates and tunnel profile information, automatically collecting a detonation pile change video in a slag discharging process, screening typical frame segments through a change detection algorithm, marking change types and time-space information, restoring disturbed rock blocks based on marking information, combining excavation exposure sequences, constructing a detonation pile multilayer three-dimensional form model from the surface to the inside, and finally extracting a block size grading curve, rock block space distribution and detonation pile shape information. The system correspondingly comprises a camera, a light source, an analysis processing module, a calibration module and the like. The invention solves the problems of difficult acquisition of the internal block degree of the explosion stack, poor detection aging, insufficient environmental adaptability and the like in the prior art, and realizes automatic, instant and comprehensive detection of the tunnel blasting in-situ block degree.

Inventors

  • SUN PENGCHANG
  • MENG HAILI
  • XUE LI
  • KANG YONGQUAN
  • LIU YIWEI
  • SUN CUIYUAN
  • CAI YANG

Assignees

  • 中国铁道科学研究院集团有限公司铁道建筑研究所
  • 中国铁道科学研究院集团有限公司
  • 北京铁科特种工程技术有限公司

Dates

Publication Date
20260508
Application Date
20260407

Claims (10)

  1. 1. The tunnel blasting block degree in-situ instant detection method based on video measurement is characterized by comprising the following steps of: (1) The method comprises the steps that a high-intensity light source and a camera are arranged on a cab front window frame, a movable arm and a bucket rod of an excavator for tunnel slag tapping, and the light source and the camera are connected and configured, so that the irradiation angle and the brightness of the light source are matched with the acquisition requirement of the camera, and a video acquisition hardware system suitable for a tunnel low-illumination environment is constructed; (2) By recording attitude data of an azimuth sensor of the camera in a reference state and a working state of the excavator, combining geometric information of a calibration plate and a tunnel contour, establishing a single camera calibration model and solving a space transformation relation among a plurality of cameras, and completing three-dimensional calibration of a system; (3) When carrying an excavator for deslagging operation, adjusting parameters of a camera and a light source according to the requirements of environmental brightness and block analysis precision, automatically and continuously acquiring an overall process video from formation of a blasting pile to emptying, and recording position change and form change of broken rock blocks of the blasting pile in the deslagging process; (4) Transmitting and storing the acquired video, automatically screening typical video segments and image frames with the detonation pile changed in the deslagging process through a change detection algorithm, and marking the frame segments with corresponding camera calibration information, deslagging time sequence information and detonation pile change type information, wherein the detonation pile change type information comprises the change of the detonation pile in-situ block degree and the change of the undisturbed detonation pile in-situ block degree; (5) The method comprises the steps of sequencing marked frame segments according to slag discharge time sequence, carrying out image correction, restoring the positions and the morphological changes of rock blocks marked as disturbance of in-situ block degree in the frame segments due to slag discharge by utilizing adjacent frame information to obtain clean explosion stack frame segments, reconstructing apparent three-dimensional forms based on the clean explosion stack frame segments of a plurality of time sequences, combining excavation pit change information marked as the interior of an undisturbed explosion stack, recursively superposing apparent three-dimensional form models of different slag discharge time from the surface to the inside through a time sequence difference set and reverse motion compensation to generate a complete in-situ explosion stack multilayer three-dimensional form model, measuring characteristic sizes of each broken rock block according to the model, and calculating and outputting explosion block degree grading curves, spatial distribution characteristics of the rock blocks of different sizes and explosion stack shape curves.
  2. 2. The method according to claim 1, wherein the step (1) specifically comprises: ① A high-intensity light source is arranged on a cab front window frame, a movable arm and a bucket rod of the excavator for tunnel slag discharge; ② Installing a camera on a cab front window frame, a movable arm and a bucket rod of the excavator for tunnel slag discharge; ③ The high-intensity light source assembly and the camera assembly are connected, and the irradiation angle and the brightness of the high-intensity light source are set to be matched with the video acquisition requirement of the camera.
  3. 3. The method according to claim 1, wherein the step (2) specifically comprises: ① The method comprises the steps of adjusting the deslagging excavator to a cab homing state and a movable arm fully-retracted state, taking the cab homing state and the movable arm fully-retracted state as a reference state, and recording the azimuth and the angle in an azimuth sensor of the camera in the reference state; ② Driving the excavator into the tunnel, adjusting the positions of the cab and the movable arm to a working state, and recording the azimuth and the angle in an azimuth sensor of the camera in the working state; ③ Placing a camera calibration plate in a video acquisition range of a camera, and starting the camera to acquire video images comprising the calibration plate and tunnel contours; ④ Establishing a calibration model of the tunnel blasting block video detection system by using the tunnel profile direction and geometric information, calibration plate geometric information, azimuth and angle information in azimuth sensors under different states of the camera and a built-in coordinate system of the camera; ⑤ According to the calibration result of the single camera, a rotation matrix Rm and a translation vector Tm between any two cameras are calculated, a transformation matrix Tm between the cameras is solved, and the three-dimensional calibration of the system is completed.
  4. 4. The method according to claim 1, wherein the step (3) specifically comprises: ① The excavator runs to the front of the tunnel face blasting pile and is adjusted to a slag discharging working state; ② Setting shooting frame rate and image definition of a camera according to actual precision requirements of tunnel blasting block analysis; ③ Adjusting the irradiation intensity and angle of a light source according to the brightness environment of the tunnel slag discharging operation section so as to match the shooting requirement of a camera; ④ And in the process of normally carrying out tunnel slag discharge operation, automatically collecting the whole process video from post-explosion accumulation to gradual emptying of the explosion stack.
  5. 5. The method according to claim 1, wherein the step (4) specifically comprises: ① Transmitting and storing tunnel blasting slag discharging videos shot and collected by each camera in real time; ② Analyzing tunnel explosion pile change conditions of slag video display by adopting an unsupervised change detection method based on inter-frame difference and background modeling, namely carrying out gray level and filtering pretreatment on continuous video frames, establishing and dynamically updating a background model, calculating a difference image of a current frame and the background model, extracting a change area mask after thresholding and morphological operation; ③ Reading and identifying the segment video frequency section frame by frame, isolating the frame section without any change of the explosion stack, and selecting the typical frame section; ④ Marking each section frame section with calibration information of the camera and slag discharging time sequence information by combining shooting azimuth and angle information of the section frame section camera; ⑤ Marking broken rock blocks of the explosion pile and positions thereof with position change and form change according to the change condition of the explosion pile in the deslagging process, and marking the change information of the explosion pile, namely a pit hole formed by normal deslagging and excavation; ⑥ For the frame section with the position and the form of the broken rock of the explosion pile, the information of the immediately previous frame section is marked in an associated way; ⑦ Integrating all the marking information and outputting all the frame segments carrying the marking information.
  6. 6. The method according to claim 1, wherein the step (5) specifically comprises: ① Sequencing shot and collected explosion stack frame segments according to the slag discharging time sequence; ② Correcting each frame segment according to the calibration information filled in the frame segment, and replacing the original frame segment with the corrected frame segment; ③ According to the position and form change information of the broken rock of the explosion pile of disturbance in-situ block degree of each frame section image mark filling, utilizing the marked adjacent frame section information, adopting a motion compensation and image restoration method based on feature matching to restore the rock related to the rock position and form change caused by slag discharge process in each corrected frame image, tracking the motion track of the rock in the adjacent frame through scale-invariant feature transformation or optical flow method, restoring the image content to the position before disturbance by reverse affine transformation, filling the gap generated after removal by utilizing an image restoration technology, restoring to the original position and form, and forming a clean explosion pile frame section; ④ Image segmentation and edge detection processing are carried out on each cleaned detonation reactor frame segment, and an apparent three-dimensional morphological model of each detonation reactor frame segment is obtained; ⑤ Combining slag discharging time sequence, marking slag discharging excavation formed by the original position block degree of the undisturbed detonation pile filled with each frame section image, and utilizing marked adjacent frame section information to stack apparent three-dimensional morphological models with different slag discharging time from the outside to the inside to generate a complete multilayer three-dimensional morphological model of the detonation pile; Is provided with And (3) with The slag is discharged from the slag discharge device at the adjacent slag discharge time sequence moment; the apparent three-dimensional morphological models at corresponding moments are respectively represented in a point cloud form; for a pair of Registering and aligning point clouds; Is that The apparent three-dimensional morphological model of the detonation reactor after cleaning and restoring at the moment i is represented by a point cloud set; By difference set operation Extracted at A time of day newly exposed internal rock mass surface point cloud, wherein To at the same time The point cloud of the surface of the newly exposed internal rock mass at the moment; representing a set difference set operation; rock displacement vector field determined by combining excavator bucket motion trail and image optical flow field Reconstructing an original surface point cloud of a removed rock mass using an inverse motion compensation method Wherein To remove the space transformation matrix of rock from the original pose to the exposed pose, the vector field is displaced by the rock It is determined that the number of the cells, Is the inverse matrix thereof; taking the reconstructed original surface point cloud as a first Layer rock block model With an initial surface rock mass model Recursion superposition to generate finally constructed in-situ detonation stack multilayer three-dimensional morphological model Wherein n is the total number of key frames acquired in the deslagging process, ⋃ represents union operation of the set; ⑥ Comparing the tunnel contour geometric information in the frame section, and correcting to form an in-situ detonation stack multilayer three-dimensional morphological model by correcting deviation or taking average value of the tunnel contour geometric information and the actual geometric information; ⑦ And measuring the maximum characteristic size of each broken rock in the model according to the modified in-situ explosion stack multilayer three-dimensional morphological model, and calculating and outputting an explosion block size grading curve, different-size rock spatial distribution characteristics and explosion stack shape curves.
  7. 7. The method of claim 1, wherein the steps (1) and (2) are performed before the excavator is put into use, and the steps (3) to (5) are performed in situ and automatically during the deslagging process of the excavator.
  8. 8. A tunnel blasting blocking in situ instant detection system for implementing the method of any of claims 1-7, comprising: The camera assembly comprises a plurality of cameras arranged on a front window frame, a movable arm and a bucket rod of the cab of the excavator and is used for collecting a video of the change of the explosion pile in the deslagging process; The light source assembly comprises a plurality of high-intensity light sources arranged on a front window frame, a movable arm and a bucket rod of the cab of the excavator and is used for providing illumination for the camera; The analysis processing device is in communication connection with the camera assembly, and is used for receiving and storing video data and executing the step (4) and the step (5); And the calibration module is used for executing the step (2) based on the camera azimuth sensor data, the calibration plate and the tunnel contour information.
  9. 9. The system of claim 8, wherein the camera in the camera assembly has vibration-proof, adjustable shooting angle, and timed acquisition functions and is configured with an orientation sensor.
  10. 10. The system of claim 8 or 9, further comprising a control and matching unit, connected to the light source assembly and the camera assembly, respectively, for dynamically adjusting the illumination angle and brightness of the light source assembly based on ambient brightness and camera acquisition parameters.

Description

Tunnel blasting block degree in-situ information instant detection method based on video measurement Technical Field The invention relates to the technical field of engineering blasting and tunnel engineering, in particular to a tunnel blasting block degree in-situ instant detection method based on video measurement. Background The tunnel blasting block size is used as a key index for evaluating blasting quality, and the distribution characteristics of the tunnel blasting block size directly influence the energy consumption and the efficiency of the shoveling transportation of the tunnel slag and the subsequent recycling utilization, so that the tunnel blasting block size has important guiding significance in tunnel engineering. In order to realize optimal adjustment of blasting parameters, effective control of construction cost and accurate assessment of blasting construction quality, accurate and timely blasting block detection method and data are required to be relied on. Currently, tunnel blasting block detection methods mainly comprise a standard screening method, an image analysis method, a laser scanning method and an emerging method based on physical signals. The standard screening method is used as a traditional detection means, the rock samples after explosion are required to be screened and classified, the mass percentages of all the particle fractions are calculated, the result is visual and reliable, the process is tedious and time-consuming, and quick and large-scale field detection is difficult to realize. The image analysis method collects the surface image of the detonation heap by means of a camera or an unmanned aerial vehicle, and the blocking distribution is counted by an image recognition technology, so that the method has the advantages of non-contact and high speed, but is limited by an image collection mode, and the method can only acquire the blocking information of the surface of the detonation heap, cannot recognize the blocking distribution condition in the detonation heap, and is easily interfered by environmental factors such as illumination, dust and the like. The laser scanning method acquires point cloud data on the surface of the explosion stack through three-dimensional laser scanning, so that a rock three-dimensional model is rebuilt, the method has higher measurement precision, and the spatial distribution characteristics of the block degree can be partially reflected, however, the method has high equipment cost, long data acquisition and processing period, is difficult to acquire block degree information in the explosion stack, and has poor adaptability in low-illumination and high-dust environments in tunnels. The method based on the physical signal (such as sound wave or vibration analysis) can realize indirect and rapid evaluation of the post-blasting block degree by collecting the physical signal in the blasting process and constructing a correlation model between the signal characteristics and the block degree, but the model depends on specific geology and blasting conditions, has low universality and needs on-site calibration to ensure the precision. In summary, the existing tunnel blasting block detection methods have the following defects, although each has a emphasis: 1. The identification and the acquisition of the internal block degree of the explosion stack are difficult to realize, and most methods can only acquire the surface layer information of the explosion stack and cannot comprehensively reflect the real distribution characteristics of the explosion block degree; 2. the detection result lacks of block morphology and space distribution information, and multi-dimensional data support cannot be provided for blasting parameter optimization and shovel loading process improvement; 3. The detection timeliness and engineering adaptability are insufficient, the traditional method is complex in operation and long in period, is difficult to adapt to the rapid continuous operation requirement of a slag discharging link after tunnel blasting, and is reduced in reliability in severe environments such as low illumination and high dust; 4. the automation and integration degree is low, the existing method mostly needs manual intervention or off-line processing, the in-situ, instant and continuous acquisition and processing of blasting block information can not be realized, and the requirement of intelligent construction on real-time data feedback is difficult to meet. Therefore, a novel in-situ instant detection method which can realize synchronous acquisition of internal and external block information of a blasting pile under a severe environment after tunnel blasting, has high timeliness, high automation degree and strong adaptability is needed, so as to fill the blank of the prior art in the aspects of comprehensive sensing and real-time feedback of blasting block. Disclosure of Invention The invention aims to provide a tunnel blasting block degree in-situ instant