Intelligent identification method and system of rock mass fractures by integrating three-dimensional scanning and image spectrum

Abstract

The present invention discloses a rock mass fracture intelligent identification system and method integrating three-dimensional scanning and image spectrum, including: obtaining three-dimensional point cloud data of target rock mass through three-dimensional scanning; processing the three-dimensional point cloud data to obtain fracture point cloud data set in two-dimensional coordinate system; obtaining image information and spectrum information of target rock mass through spectrum scanning; extracting mineral distribution characteristics based on spectrum information, and extracting fracture image characteristics based on image information; obtaining boundary coordinates of filled fractures and filling mineral information by using intelligent identification model of filled fracture boundary based on mineral distribution characteristics and fracture image characteristics; verifying and merging boundary coordinates of filled fractures with fracture point cloud data set in two-dimensional coordinate system, so as to realize comprehensive and accurate identification of rock mass fractures. The present invention can effectively make up for the defect that three-dimensional laser scanning cannot realize accurate identification of filled fractures and efficient extraction of parameters, and greatly improves the accuracy of fracture identification.

Description

Technical Field

[0001] The present invention relates to the technical field of rock mass crack identification, and in particular to a rock mass crack intelligent identification method and system integrating three-dimensional scanning and image spectrum. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] There are a large number of fracture structures in the rock mass, which have a great influence on the physical and mechanical performance indicators, deformation characteristics, seepage properties and long-term stability of the rock mass. Fractures can be divided into filled fractures and unfilled fractures according to whether they are filled or not. Common fracture fillings include breccia, sand and limestone formed by tectonic action, weathering action and groundwater action, which have a great difference from the surrounding rock composition and a low probability of consistent composition.

[0004] Looking at the field of rock mass crack identification, the common identification methods currently include naked eye observation, image recognition, and three-dimensional laser scanning. Naked eye observation often relies too much on manual experience, requires high professional skills and knowledge reserves of the observer, cannot achieve accurate quantitative analysis of cracks, is easily restricted by environmental factors such as weather and light, has a large workload, low efficiency, high labor costs, and low economic benefits; image recognition methods are often limited by image quality, have high requirements for algorithm and model integrity, and training data often need to be preprocessed. The original model parameters cannot be directly used, and the final recognition accuracy needs to be improved; three-dimensional laser scanning technology directly samples the rock surface in three dimensions through three-dimensional laser scanning equipment. Compared with traditional technologies, although it can quickly, in-situ, and high-precision identify rock mass cracks, since three-dimensional scanning can only sample data on the rock surface, and the internal space of the cracks filled with minerals, the crack boundaries obtained by scanning are prone to distortion or even misjudgment. In summary, using only one technology for crack identification has certain limitations in terms of identification accuracy. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a method and system for intelligent identification of rock fractures that integrates three-dimensional scanning and image spectrum. Three-dimensional scanning technology is used to realize accurate identification of unfilled fractures and fuzzy identification of filled fractures. Image spectrum technology is combined with neural network to realize accurate identification of filled fractures, and the three-dimensional scanning results are supplemented and verified, ultimately realizing accurate, comprehensive, rapid and intelligent identification of rock fractures.

[0006] In some embodiments, the following technical solutions are adopted:

[0007] A rock mass fracture intelligent identification method integrating three-dimensional scanning and image spectrum, comprising:

[0008] Acquire 3D point cloud data of the target rock mass through 3D scanning;

[0009] Processing the three-dimensional point cloud data, extracting the coordinates of the discontinuous point set, and performing dimension transformation on the extracted coordinates of the discontinuous point set to obtain a fracture point cloud data set in a two-dimensional coordinate system;

[0010] Acquire image information and spectral information of target rock mass through spectral scanning;

[0011] Extracting mineral distribution characteristics based on spectral information, and extracting fracture image characteristics based on image information; based on the mineral distribution characteristics and fracture image characteristics, using a filling fracture boundary intelligent recognition model, obtaining the boundary coordinates of the filling fracture and the filling mineral information;

[0012] The boundary coordinates of the filled fractures are checked and merged with the fracture point cloud data set in the two-dimensional coordinate system to achieve comprehensive and accurate identification of rock fractures.

[0013] As a further solution, the coordinates of the extracted discontinuous point set are transformed in dimension. The specific process is: for the three-dimensional (x, y, z) coordinates of the extracted discontinuous point set, the z-direction data is manually deleted and converted into a crack point cloud data set in a two-dimensional coordinate system.

[0014] As a further solution, the mineral distribution characteristics are extracted based on spectral information. The specific process is as follows:

[0015] The spectral curve of each pixel in the image is obtained, and whether the two adjacent spectral curves belong to the same type of rock mass is determined based on whether the spectral angle of the two adjacent spectral curves is less than the set threshold. The spectral curve is matched with the spectral database to determine the specific mineral type corresponding to the spectral curve; thereby determining the mineral type and mineral distribution corresponding to each pixel, and performing mineral mapping; wherein the spectral database pre-stores the spectral curve and the mineral molecular formula matching it.

[0016] As a further solution, based on the mineral distribution characteristics and fracture image characteristics, an intelligent recognition model of filled fracture boundaries is used to obtain the boundary coordinates of filled fractures and filling mineral information. The specific process is: input the fracture image characteristics and mineral distribution characteristics obtained by processing the imaging spectral image cube, match the fracture coordinates at the same position of the target rock mass with the mineral mapping area, extract the boundary of the overlapping area of ​​the two areas, regard it as the filled fracture area, and finally realize the output of the filled fracture boundary coordinates and mineral composition in combination with mineral mapping.

[0017] As a further solution, the filling crack boundary intelligent recognition model adopts a set neural network model, and the recognition process of the filling crack boundary intelligent recognition model specifically includes:

[0018] Image features and spectral features are extracted according to the imaging spectral image cube in the existing database, and the set ratio is divided into training set and test set. The training set is input into the neural network model for training, and the trained neural network model is optimized and verified using the test set to obtain the optimal neural network model; the image features and spectral features of the target rock mass that needs to be identified as filled cracks are input into the trained neural network model to obtain the filled crack identification results and their boundary coordinates set, and combined with the mineral composition, the boundary coordinates of the filled cracks and the filling mineral information are obtained.

[0019] As a further solution, the boundary coordinates of the filled cracks are checked and merged with the crack point cloud dataset in the two-dimensional coordinate system, specifically:

[0020] The coordinates of the fracture point cloud data in the two-dimensional coordinate system are aligned with the boundary coordinates of the filled fracture, the two coordinates are unified into the same reference coordinate system, and the boundary coordinates of the filled fracture replace the coordinates of the fracture point cloud data in the two-dimensional coordinate system of the corresponding position.

[0021] As a further solution, the identification results of rock mass fractures are obtained, including: the location of the fractures, the filling mineral information and the size of the fractures.

[0022] In other embodiments, the following technical solutions are adopted:

[0023] An intelligent rock fracture identification system integrating three-dimensional scanning and image spectrum, comprising:

[0024] A point cloud data acquisition module is used to acquire three-dimensional point cloud data of the target rock mass through three-dimensional scanning;

[0025] A fracture point cloud data preprocessing module is used to process the three-dimensional point cloud data, extract the coordinates of the discontinuous point set, and perform dimension transformation on the extracted discontinuous point set coordinates to obtain a fracture point cloud data set in a two-dimensional coordinate system;

[0026] An image spectrum data acquisition module is used to acquire image information and spectrum information of the target rock mass through spectrum scanning;

[0027] A filling crack data acquisition module is used to extract mineral distribution characteristics based on spectral information and fracture image characteristics based on image information; based on the mineral distribution characteristics and fracture image characteristics, the intelligent recognition model of the filling crack boundary is used to obtain the boundary coordinates of the filling crack and the filling mineral information;

[0028] The data fusion module is used to verify and merge the boundary coordinates of the filled fractures with the fracture point cloud data set in the two-dimensional coordinate system to obtain the identification results of the rock fractures.

[0029] As a further solution, the data fusion module aligns the coordinates of the crack point cloud data in the two-dimensional coordinate system with the boundary coordinates of the filled crack, unifies the two coordinates into the same reference coordinate system, and replaces the coordinates of the crack point cloud data in the two-dimensional coordinate system of the corresponding position with the boundary coordinates of the filled crack.

[0030] In other embodiments, the following technical solutions are adopted:

[0031] A rock mass fracture intelligent identification device integrating three-dimensional scanning and image spectrum, used to implement the above-mentioned rock mass fracture intelligent identification method integrating three-dimensional scanning and image spectrum, characterized in that the system comprises:

[0032] Work platform;

[0033] An electromagnetic wave transceiver is arranged on the working platform, and the electromagnetic wave transceiver can transmit electromagnetic waves to the side rock wall. The distance between the electromagnetic wave transceiver and the side rock wall can be obtained by using the round-trip propagation time of the electromagnetic wave, so as to control the working platform to automatically move according to the set route;

[0034] The beams are arranged relative to the working platform, and the beams are connected to the working platform through lifting devices; the three-dimensional scanning device and the spectrum scanning device are respectively mounted on different beams, and can rotate and translate along the beams; the working platform can drive the three-dimensional scanning device and the spectrum scanning device to automatically move to the set target position for data collection;

[0035] The control unit is used to receive data collected by the three-dimensional scanning device and the spectrum scanning device, obtain a fracture point cloud data set and a boundary coordinate data set of the filled fractures based on the data, merge the two data sets, and obtain all fractures of the target rock mass.

[0036] As a further solution, it also includes: a micro positioning camera arranged on the working platform, which accurately locates the identification range of the rock mass to be measured by locating the characteristic points of the target rock mass through multi-angle shooting, and transmits the positioning information to the three-dimensional scanning device to achieve regional locking of the target area to be measured.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] (1) The present invention proposes to use image spectroscopy technology to accurately identify filled fractures. Based on the existing lithology spectral information database (the database consists of an imaging spectral image cube and spectral images of various minerals with corresponding spectral numbers and molecular formulas, combined with fracture image information characteristics), a neural network model is combined to train an intelligent recognition model for filled fractures, thereby realizing the intelligent extraction and output of the filled fracture boundary coordinate set and the filling material parameters, and finally realizing the accurate recognition of filled fractures and the efficient extraction of parameters (parameters include fracture shape, size, position, filling material mineral composition, etc.).

[0039] (2) The present invention combines three-dimensional laser scanning technology to identify the cracks in the target rock mass. The two crack identification results complement and verify each other to achieve comprehensive and accurate identification of the cracks in the target rock mass. The method of the present invention can effectively make up for the defects of three-dimensional laser scanning that cannot accurately identify filled cracks and cannot extract filling materials, improve data quality, and greatly improve the accuracy of crack identification.

[0040] (3) The present invention designs an intelligent rock fracture identification system that integrates three-dimensional scanning and image spectrum. The system can autonomously move to the target location based on a predetermined route to perform three-dimensional scanning and spectral scanning. At the same time, it can autonomously adjust the specific position of the scanning device and remotely control the fracture identification process, thereby achieving fully intelligent, accurate identification and efficient processing of fracture identification, greatly improving the work efficiency of rock fracture identification.

[0041] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the intelligent identification method of rock mass fractures integrating three-dimensional scanning and image spectrum in an embodiment of the present invention;

[0043] Figure 2 It is a flow chart of the intelligent identification system of rock mass fractures integrating three-dimensional scanning and image spectrum in an embodiment of the present invention;

[0044] Figure 3 It is a schematic diagram of the structure of a rock mass fracture intelligent identification device integrating three-dimensional scanning and image spectrum in an embodiment of the present invention;

[0045] Figure 4 Schematic diagram of height adjustment of the rock mass fracture intelligent identification device in an embodiment of the present invention;

[0046] Among them, 1. 3D laser scanner; 2. spectral imager; 3. crossbeam; 4. robotic arm; 5. hydraulic column; 6. working platform; 7. intelligent operation display screen; 8. light source; 9. micro positioning camera; 10. distribution box; 11. handle; 12. fill light; 13 electromagnetic wave transceiver. DETAILED DESCRIPTION

[0047] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0048] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0049] Embodiment 1

[0050] In one or more embodiments, a method for intelligently identifying rock fractures that integrates three-dimensional scanning and image spectrum is disclosed. The fractures are divided into two categories: filled fractures and unfilled fractures. Three-dimensional scanning technology and data processing models are used to accurately extract unfilled fractures and fuzzy extract filled fractures. Image spectrum technology is used to combine image features and mineral distribution features, and neural networks are used to accurately identify filled fractures. The two methods each perform their respective functions, and the results complement and verify each other, ultimately achieving complete, intelligent, and efficient identification of fractures with full coverage.

[0051] Combination Figure 1 The method of this embodiment specifically includes the following process:

[0052] Step (1): obtaining three-dimensional point cloud data of the target rock mass through three-dimensional scanning;

[0053] Step (2): Perform noise filtering on the three-dimensional point cloud data, extract the coordinates of the discontinuous point set, transform the dimension of the extracted discontinuous point set coordinates, manually delete the three-dimensional (x, y, z) coordinates of the extracted discontinuous point set, and obtain a fracture point cloud dataset in a two-dimensional coordinate system. The dataset includes an unfilled fracture point cloud dataset and a filled fracture point cloud dataset. The recognition result of the filled fracture point cloud dataset may be biased. Therefore, by merging the two-dimensional filled fracture dataset coordinates obtained by the filled fracture boundary intelligent recognition model, an accurate and comprehensive unfilled and filled fracture point cloud dataset can be obtained.

[0054] It should be noted that in the process of collecting 3D point cloud data, the surface of the object is measured by a 3D scanning instrument, and the data obtained generally have many geometric discontinuities, which are usually in the form of discrete points, called discontinuities.

[0055] In this embodiment, processing the three-dimensional point cloud data specifically includes: point cloud intensity correction, point cloud data registration under different viewing angles, and non-target point cloud filtering;

[0056] Among them, point cloud intensity correction specifically involves correcting the point cloud intensity through a preset point cloud intensity calibration model, selecting the most appropriate laser incident angle, optimal laser scanner position, etc.;

[0057] The point cloud data registration under different viewing angles specifically uses an automatic point cloud data stitching method based on reflectance value images to register the point cloud data obtained by multiple scans.

[0058] Step (3): obtaining image information and spectral information of the target rock mass through spectral scanning;

[0059] In this embodiment, the obtained spectral information is preprocessed by variance normalization to eliminate the spectral error between samples caused by different solid particle sizes, scattering or measurement optical path. The variance normalization method is as follows:

[0060]

[0061] Among them, I snv is the spectral data obtained by spectral scanning, the spectral data obtained by the first scan is set as the sample, and its variance E is calculated α with the spectral variance σ.

[0062] Step (4): extracting mineral distribution characteristics based on spectral information, and extracting fracture image characteristics based on image information; based on the mineral distribution characteristics and fracture image characteristics, using a filling fracture boundary intelligent recognition model, obtaining the boundary coordinates of the filling fracture and the filling mineral information;

[0063] In this embodiment, the mineral distribution characteristics are extracted based on the spectral information, and the specific process is as follows:

[0064] The spectral curve of each pixel in the image is obtained, and whether the spectral angles of two adjacent spectral curves are less than a set threshold is used to determine whether the two pixels belong to the same type of rock mass. The spectral curve is matched with the spectral database to determine the specific mineral type corresponding to the spectral curve; thereby determining the mineral type and mineral distribution corresponding to each pixel, and performing mineral mapping on this basis; wherein the spectral database pre-stores the spectral curve and the mineral molecular formula that matches it.

[0065] The image information of the rock mass containing fractures is subjected to fracture detection, filtering, point distribution, boundary line fitting, connection and correction to obtain fracture image features.

[0066] Then, the extracted image features and spectral features are used, and the trained intelligent recognition model of the filling fracture boundary is used to extract the boundary coordinates of the filling fracture and output the information of the filling minerals.

[0067] In this embodiment, the intelligent recognition model for filling crack boundaries can be implemented by using an existing neural network model, such as a deep neural network model. The specific processing process of the intelligent recognition model for filling crack boundaries on input data is as follows:

[0068] The fracture image features and mineral distribution features obtained by processing the imaging spectral image cube are input, the fracture coordinates at the same position of the target rock mass are matched with the mineral mapping area, the overlapping area boundary of the two areas is extracted and regarded as the filled fracture area, and finally the filled fracture boundary coordinates and mineral composition are output in combination with mineral mapping.

[0069] In this embodiment, the training process for the neural network model is as follows:

[0070] The above-mentioned image features and spectral features are extracted according to the imaging spectral image cube in the existing database, and divided into a training set and a test set in a ratio of 3:1. The training set is input into the neural network model for training, and the trained neural network model is optimized and verified using the test set to obtain the optimal neural network model; the image features and spectral features of the target rock mass that needs to be identified for filled cracks are input into the trained neural network model to obtain the filled crack identification results and their boundary coordinate sets, and combined with the mineral composition, the output of the filled cracks is completed.

[0071] Step (5): The fracture point cloud data set in the two-dimensional coordinate system is merged with the boundary coordinates of the filled fractures to obtain all fracture parameter data of the target rock mass, including data on fracture shape, size, location and mineral composition of the filling material.

[0072] Among them, the merging process is specifically as follows: align the two-dimensional discontinuous point set obtained by dimensional transformation with the coordinates of the filled fracture boundary, so that the two coordinate systems are in the same reference coordinate system, combine the image information obtained by the two technologies, determine the fracture coordinate data set with two fracture identification results at the same position, replace the original fracture coordinate set with the filled fracture boundary data, and finally determine all the fracture data of the target rock mass.

[0073] The merged fracture boundary data are processed for the second time, and operations such as boundary line fitting, connection and correction are performed on them in turn.

[0074] Combined with the distribution of filling materials in the cracks, the image spectrum data and the crack data are input into the sketching system to draw a sketch of the rock cracks. During the drawing process, a one-to-one correspondence between the sketch and the crack parameters is achieved to facilitate data retrieval.

[0075] Embodiment 2

[0076] In one or more embodiments, a rock mass fracture intelligent identification system integrating three-dimensional scanning and image spectrum is disclosed; Figure 2 , the system of this embodiment specifically includes:

[0077] The area delineation module is used to delineate the target area to be measured, reduce redundant data acquisition, and reduce the difficulty of subsequent data processing.

[0078] A point cloud data acquisition module is used to acquire three-dimensional point cloud data of the target rock mass through three-dimensional scanning;

[0079] A crack point cloud data preprocessing module is used to perform noise filtering on the three-dimensional point cloud data, extract the coordinates of the discontinuous point set, and perform dimensional transformation on the extracted coordinates of the discontinuous point set to obtain a crack point cloud data set in a two-dimensional coordinate system;

[0080] An image spectrum data acquisition module is used to acquire image information and spectrum information of the target rock mass through spectrum scanning;

[0081] The filling crack intelligent extraction module is used to extract mineral distribution characteristics based on spectral information and fracture image characteristics based on image information; based on the mineral distribution characteristics and fracture image characteristics, the filling crack boundary intelligent recognition model is used to obtain the boundary coordinates of the filling crack and the filling mineral information;

[0082] The data fusion module is used to verify and merge the fracture point cloud data set with the boundary coordinates of the filled fractures to obtain the identification results of the rock fractures.

[0083] It should be noted that the specific implementation methods of the above modules have been described in Example 1 and will not be described in detail here.

[0084] Embodiment 3

[0085] In one or more embodiments, a rock mass fracture intelligent identification system integrating three-dimensional scanning and image spectrum is disclosed, and the system can implement the rock mass fracture intelligent identification method integrating three-dimensional scanning and image spectrum described in Example 1;

[0086] Combination Figure 3 , the system of this embodiment specifically includes:

[0087] Cross beams 3 are respectively arranged on both sides of the platform, and the two cross beams 3 are respectively connected to the working platform 6 through a lifting device; the three-dimensional scanning device and the spectrum scanning device are respectively mounted on different cross beams, and a transmission device and a rotating shaft are arranged on the cross beams. The horizontal displacement and rotation of the three-dimensional scanning device or the spectrum scanning device can be realized through the transmission device and the rotating shaft on the cross beams.

[0088] In this embodiment, the lifting device adopts a hydraulic column 5, which can freely adjust the height. Figure 4 As shown, it is ensured that the devices on the two beams do not block each other, and a mechanical arm 4 is also installed on the hydraulic column to realize remote control of the equipment.

[0089] The three-dimensional scanning device is mainly composed of a three-dimensional laser scanner 1, and the spectrum scanning device is mainly composed of a light source and a spectrum imager 2. The light source 8 adopts a halogen lamp, and the brightness and angle of the light source are adjustable.

[0090] The working platform is also equipped with a micro positioning camera 9, a fill light 12, an intelligent operation display screen 7 and other test equipment and objects on the platform. The micro positioning camera 9 accurately locates the identification range of the rock mass to be tested by shooting the characteristic points of the target rock mass at multiple angles, and synchronizes the positioning information to the three-dimensional laser scanner through the sensor to achieve regional locking of the target area to be tested, reducing the difficulty of later data processing. The fill light 12 uses a high-brightness LED light source to fill in the light in insufficiently lit environments such as tunnels, ensuring the smooth progress of image data acquisition and micro camera positioning.

[0091] The working platform is equipped with shock-absorbing and wear-resistant high-performance tires to ensure the smooth movement of the device; the working platform can drive the three-dimensional scanning device and the spectrum scanning device to move to the set target position for data collection.

[0092] The working platform is also equipped with a plurality of electromagnetic wave transceivers. The electromagnetic wave transceiver 13 transmits electromagnetic waves through a built-in pulse generator, and uses the number of pulses in the time interval between the round trip propagation of the pulses on the measuring line to obtain the distance from the surrounding rock wall, so as to control the working platform to automatically travel on a predetermined route. The device also includes a distribution box 10, the purpose of which is to provide energy for the entire device and reasonably configure the electrical energy.

[0093] The device also comprises a handle 11 for transporting the device over a short distance and for manual position adjustment.

[0094] The system of this embodiment also includes: a control unit, which is used to receive data collected by the three-dimensional scanning device and the spectral scanning device, and based on these data, obtain a point cloud data set of unfilled fractures and a boundary coordinate data set of filled fractures, respectively, and merge the two data sets to obtain all fractures of the target rock mass.

[0095] The specific working process of the control unit is consistent with the method disclosed in Example 1. The control unit mainly includes the following components:

[0096] (1) Data collection and processing system: It consists of a data collection module and a data processing module. The data collection module is responsible for receiving the image, spectrum, point cloud and other data acquired by the fracture scanning integrated system, and classifying and transferring these data; the data processing module consists of a spectral data processing unit and a point cloud data processing unit. The spectral data processing unit first pre-processes the spectral information and extracts the image and mineral features, and then inputs the extracted features into the filling fracture intelligent recognition model. The filling fracture intelligent recognition model uses a neural network to extract the filling fracture boundary coordinate set and output the filler parameters through the extracted image features and spectral features. The point cloud data processing unit inputs the point cloud data and image data into the point cloud data intelligent processing model to process the point cloud data. The point cloud data intelligent processing model can sequentially perform tasks including point cloud intensity correction, point cloud data registration under different viewing angles, non-target point cloud filtering and discontinuous point set coordinate extraction on the point cloud data, and on this basis, the depth dimension of the obtained coordinate set is eliminated to achieve the dimension conversion of the coordinates.

[0097] (2) Data merging and storage system: It consists of a data integration module and a storage module. The data integration module merges the received data-processed filled fracture boundary coordinate data set and the dimensionally converted fracture point cloud data, and performs fracture numbering, and realizes the pairing of filled fractures and fillings. The merging process realizes coordinate registration, coordinate verification, coordinate elimination and coordinate merging of two-dimensional coordinate sets from different sources, and the merged fracture boundary coordinate data is input into the rock fracture output system after secondary processing. The data storage module is used to store data information and program information for data migration, backup analysis, and program upgrade detection and repair.

[0098] (3) Rock fracture output system: It consists of a rock fracture sketch module and a fracture parameter output module; the rock fracture sketch module refers to inputting fracture data into the fracture sketch software to realize the graphical depiction and output of rock fractures, and draw each fracture in detail. The fracture output module consists of a fracture output unit and a liquid crystal display screen. The fracture output unit realizes the matching and output of fracture images and parameter data, which are finally displayed on the liquid crystal display screen. The system can be controlled, fracture data queried and extracted through the display screen. The output parameter information mainly includes: the orientation, shape, size of the fracture, and the mineral composition of the filling material of the fracture;

[0099] (4) Intelligent control system: The intelligent control system can remotely control the system, start and stop the spectrometer and three-dimensional laser scanner and adjust their posture through the wireless signal transceiver and sensor in the control unit, freely adjust the transmission device and shaft on the robotic arm and beam, and the light source, and control the electromagnetic wave transceiver to ensure the precise movement of the device. The optimal scheduling and reasonable configuration of each system and supporting facilities can be achieved. The intelligent control system can also be combined with the remote control platform to realize the mobile terminal remote operation of the above process, as well as the viewing and extraction of crack and parameter information.

[0100] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A rock mass fracture intelligent identification method integrating three-dimensional scanning and image spectrum, characterized in that: include: Acquire 3D point cloud data of the target rock mass through 3D scanning; Processing the three-dimensional point cloud data, extracting the coordinates of the discontinuous point set, and performing dimension transformation on the extracted coordinates of the discontinuous point set to obtain a fracture point cloud data set in a two-dimensional coordinate system; Acquire image information and spectral information of the target rock mass through spectral scanning; Extract mineral distribution characteristics based on spectral information, and extract crack image characteristics based on image information; Based on the mineral distribution characteristics and crack image characteristics, the boundary coordinates of the filled cracks and the filling mineral information are obtained by using the filled crack boundary intelligent recognition model; The recognition process of the intelligent recognition model for filling fracture boundaries specifically includes: Image features and spectral features are extracted from the imaging spectral image cube in the existing database, and the set ratio is divided into a training set and a test set. The training set is input into the neural network model for training, and the trained neural network model is optimized and verified using the test set to obtain the optimal neural network model; the image features and spectral features of the target rock mass that needs to be identified for filled fractures are input into the trained neural network model to obtain the filled fracture identification results and their boundary coordinates set, and combined with the mineral composition, the boundary coordinates of the filled fractures and the filling mineral information are obtained; The boundary coordinates of the filled fractures are checked and merged with the fracture point cloud data set in the two-dimensional coordinate system to obtain the identification result of the rock fracture. The process of checking and merging the boundary coordinates of the filled fractures with the fracture point cloud data set in the two-dimensional coordinate system specifically includes: The coordinates of the fracture point cloud data in the two-dimensional coordinate system are aligned with the boundary coordinates of the filled fracture, the two coordinates are unified into the same reference coordinate system, and the boundary coordinates of the filled fracture replace the coordinates of the fracture point cloud data in the two-dimensional coordinate system of the corresponding position.

2. The method for intelligently identifying rock fractures by integrating three-dimensional scanning and image spectrum according to claim 1, characterized in that: The three-dimensional point cloud data is processed, and after the coordinates of the discontinuous point set are extracted, the extracted discontinuous point set coordinates are transformed in dimension. The specific process is: for the three-dimensional (x, y, z) coordinates of the extracted discontinuous point set, the z-direction data is manually deleted and converted into a crack point cloud data set in a two-dimensional coordinate system.

3. The method for intelligently identifying rock fractures by integrating three-dimensional scanning and image spectrum according to claim 1, characterized in that: Mineral distribution characteristics are extracted based on spectral information. The specific process is as follows: The spectral curve of each pixel in the image is obtained, and whether the two adjacent spectral curves belong to the same type of rock mass is determined based on whether the spectral angle of the two adjacent spectral curves is less than the set threshold. The spectral curve is matched with the spectral database to determine the specific mineral type corresponding to the spectral curve; thereby determining the mineral type and mineral distribution corresponding to each pixel, and performing mineral mapping; wherein the spectral database pre-stores the spectral curve and the mineral molecular formula matching it.

4. The method for intelligently identifying rock fractures by integrating three-dimensional scanning and image spectrum according to claim 1, characterized in that: The identification results of rock mass fractures are obtained, including: the location of the fracture, filling mineral information and the size of the fracture.

5. An intelligent rock fracture identification system integrating three-dimensional scanning and image spectrum, characterized in that: include: A point cloud data acquisition module is used to acquire three-dimensional point cloud data of the target rock mass through three-dimensional scanning; A fracture point cloud data preprocessing module is used to process the three-dimensional point cloud data, extract the coordinates of the discontinuous point set, and perform dimension transformation on the extracted discontinuous point set coordinates to obtain a fracture point cloud data set in a two-dimensional coordinate system; An image spectrum data acquisition module is used to acquire image information and spectrum information of the target rock mass through spectrum scanning; The filling crack data acquisition module is used to extract mineral distribution characteristics based on spectral information and extract crack image characteristics based on image information; Based on the mineral distribution characteristics and crack image characteristics, the boundary coordinates of the filled cracks and the filling mineral information are obtained by using the filled crack boundary intelligent recognition model; The recognition process of the intelligent recognition model for filling fracture boundaries specifically includes: Image features and spectral features are extracted from the imaging spectral image cube in the existing database, and the set ratio is divided into a training set and a test set. The training set is input into the neural network model for training, and the trained neural network model is optimized and verified using the test set to obtain the optimal neural network model; the image features and spectral features of the target rock mass that needs to be identified for filled fractures are input into the trained neural network model to obtain the filled fracture identification results and their boundary coordinates set, and combined with the mineral composition, the boundary coordinates of the filled fractures and the filling mineral information are obtained; The data fusion module is used to check and merge the boundary coordinates of the filled fractures with the fracture point cloud data set in the two-dimensional coordinate system to obtain the identification result of the rock fracture. The process of checking and merging the boundary coordinates of the filled fractures with the fracture point cloud data set in the two-dimensional coordinate system specifically includes: The coordinates of the fracture point cloud data in the two-dimensional coordinate system are aligned with the boundary coordinates of the filled fracture, the two coordinates are unified into the same reference coordinate system, and the boundary coordinates of the filled fracture replace the coordinates of the fracture point cloud data in the two-dimensional coordinate system of the corresponding position.

6. The intelligent rock mass fracture identification system integrating three-dimensional scanning and image spectrum as claimed in claim 5, characterized in that: The data fusion module aligns the coordinates of the crack point cloud data in the two-dimensional coordinate system with the boundary coordinates of the filled crack, unifies the two coordinates into the same reference coordinate system, and replaces the coordinates of the crack point cloud data in the two-dimensional coordinate system of the corresponding position with the boundary coordinates of the filled crack.

7. A rock mass fracture intelligent identification system integrating three-dimensional scanning and image spectrum, used to implement the rock mass fracture intelligent identification method integrating three-dimensional scanning and image spectrum as described in any one of claims 1 to 4, characterized in that: The system comprises: Work platform; An electromagnetic wave transceiver is arranged on the working platform, and the electromagnetic wave transceiver can transmit electromagnetic waves to the side rock wall. The distance between the electromagnetic wave transceiver and the side rock wall can be obtained by using the round-trip propagation time of the electromagnetic wave, so as to control the working platform to automatically move according to the set route; The beams are arranged relative to the working platform, and the beams are connected to the working platform through lifting devices; the three-dimensional scanning device and the spectrum scanning device are respectively mounted on different beams, and can rotate and translate along the beams; the working platform can drive the three-dimensional scanning device and the spectrum scanning device to automatically move to the set target position for data collection; The control unit is used to receive data collected by the three-dimensional scanning device and the spectrum scanning device, obtain a fracture point cloud data set and a boundary coordinate data set of the filled fractures based on the data, merge the two data sets, and obtain all fractures of the target rock mass.

8. The intelligent rock mass fracture identification system integrating three-dimensional scanning and image spectrum as claimed in claim 7, characterized in that: Also includes: A micro positioning camera is arranged on the working platform, which accurately locates the identification range of the rock mass to be measured by shooting the characteristic points of the target rock mass at multiple angles, and transmits the positioning information to the three-dimensional scanning device to achieve regional locking of the target area to be measured.

Images