Coal mine robot geological scene sensing modeling experiment system and method

Abstract

The invention discloses a coal mine robot geological scene sensing modeling experiment system and method, and relates to the field of geological exploration and experiments, and the system comprises a coal mine robot working environment simulation module, a track type mobile platform, a multi-source sensing module, a coal seam geological simulation module and an experiment analysis module. Different environments are created through the coal mine robot working environment simulation module and the rail type mobile platform to simulate the working process of the coal mine robot; a multi-source sensor module is used for obtaining scene in a complex environment and sensing information of a coal seam geology simulation module, and a sensing data set of each sensor is constructed; and comparing and analyzing the data by using an experimental analysis module, completing multi-source sensing information fusion, and constructing a geology and scene integrated three-dimensional model. The invention provides an effective experimental platform and an experimental method for researching a signal transmission mechanism sensed by each sensor under a complex working condition, echo characteristics of electromagnetic waves in different geologies and integrated modeling of a geological scene.

Description

Technical Field

[0001] The present invention relates to the field of geological exploration and experiment, and in particular to a coal mine robot geological scene perception modeling experimental system and method. Background Art

[0002] my country is rich in coal resources, but the geological conditions of coal mines are poor and the level of intelligent mining equipment is low, resulting in reduced mining efficiency. The environment of coal mine robot comprehensive mining face is complex and harsh. Traditional pre-mining detection and modeling are inefficient and have poor accuracy. It is impossible to dynamically correct the geological model in real time. Coal mine robots lack autonomous perception and real-time capabilities. Small faults, broken zones and other unfavorable geological structures during construction increase the difficulty of excavation technology and affect the excavation speed and safety. Therefore, it is very necessary to develop an experimental system and method that can realize real-time geological scene perception modeling.

[0003] Unmanned intelligent precision mining is the core of intelligent construction of coal mines. The precise mining of coal mine robots requires them to have autonomous environmental perception and real-time adjustment capabilities. Due to the harsh underground environment, especially the lack of light, high dust, high humidity and other factors that have different effects on sensors, there are certain defects in the integrated modeling of geological scenes in comprehensive mining faces. At present, geological modeling and scene modeling are carried out using different technologies, and both geological and scene modeling are pre-modeled manually, which has problems such as low modeling efficiency, high model coupling complexity, low model accuracy, and difficulty in updating. At present, ground penetrating radar geological modeling mainly focuses on the study of propagation characteristics in coal-rock media. The propagation mechanism of perception signals in complex geology of coal mining faces is still unclear. There is a lack of research on the propagation characteristics of electromagnetic waves in gas-coal-rock cross-medium echo signals during robot operation. The multi-sensor fusion perception and information joint interpretation mechanism in complex environments is still unclear.

[0004] Therefore, it is necessary to provide a coal mine robot geological scene perception modeling experimental system and method to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a coal mine robot geological scene perception modeling experimental system and method, which is used to simulate the coal mine robot operation process and its working environment, facilitate the laboratory simulation test of multi-source fusion perception modeling of the coal mine robot operation process, collect complex environment scene information, study the perception characteristics of multiple sensors in complex environments and multi-source information joint interpretation and perception enhancement methods, explore the further realization of real-time geological scene integrated detection and modeling, and use computers and other equipment to perform real-time modeling of geology and scenes. It is convenient to carry out laboratory simulation tests on ground penetrating radar coal-rock interface recognition, study the cross-medium propagation characteristics of electromagnetic waves, construct typical coal-rock data sets, and realize full waveform inversion and abnormal body recognition of detection.

[0006] To achieve the above-mentioned purpose, the present invention provides a coal mine robot geological scene perception modeling experimental system, including a closed box composed of wood board, iron sheet and insulation layer, a scene feature simulation device arranged on the left side of the closed box, a truth value measurement device arranged inside the closed box, and also including a coal mine robot working environment simulation module arranged inside the closed box, a coal seam geological simulation module arranged inside the closed box, a track-type mobile platform arranged in front of the coal seam geological simulation module, and a multi-source perception module carried on the track-type mobile platform; the multi-source perception module, the truth value measurement device and the coal mine robot working environment simulation module are all connected to the experimental analysis module, and the three-dimensional scene information and the position information of the track-type mobile platform and the sensor are obtained through the truth value measurement device, and the distance between the multi-source perception module and the coal seam geological simulation module is changed through the track-type mobile platform to change the perception information of the multi-source perception module.

[0007] Preferably, the material of the scene feature simulation device is set to a wooden structure, and the position and posture are adjusted according to the experimental content to simulate different scenes; the true value measurement device includes a laser tracker and a laser scanner, and the laser scanner is used to obtain scene information, and the laser tracker and the target are used to obtain the position and posture information of the track-type mobile platform and the sensor; the coal seam geological simulation module includes coal blocks and rock blocks of different shapes, and the coal blocks and rock blocks are cut from complete coal blocks and rock blocks according to requirements, and are quickly recombined into different simulated geological conditions by stacking coal blocks and rock blocks of different shapes.

[0008] Preferably, the coal mine robot working environment simulation module includes a dust generator, a water vapor generator, a lighting device, a dehumidification and dust removal equipment, a dust concentration detection sensor, an illumination detection sensor, and an air humidity detection sensor; the dust generator is arranged at the bottom of the coal powder box, the lighting device includes a controllable lighting lamp, and the dehumidification and dust removal equipment includes an intake fan and an exhaust fan.

[0009] Preferably, the multi-source perception module is composed of various sensors mounted on a track-type mobile platform and jointly deployed according to the experimental content. The sensors include ground penetrating radar, lidar, frequency modulated continuous wave radar, visible light camera, infrared camera and inertial navigation unit.

[0010] Preferably, guide wheels and a drive motor are arranged under the track-type mobile platform, and the track-type mobile platform includes a transverse guide rail and a longitudinal guide rail. The surfaces of the transverse guide rail and the longitudinal guide rail are both non-smooth surfaces. The transverse guide rail is arranged at the bottom of the closed box, and a guide rail groove is opened under the longitudinal guide rail. The guide rail groove is in contact with the transverse guide rail, and the guide wheel is in contact with the longitudinal guide rail. The drive motor is connected to the guide wheel, and the drive motor drives the guide wheel to move. An electric push rod is installed on the left side of the longitudinal guide rail, and the electric push rod is connected to the longitudinal guide rail. The longitudinal guide rail is pushed to make a transverse movement through the electric push rod.

[0011] Preferably, the experimental analysis module includes a computer and related software. The sensor transmits the collected data to the computer, and the data is analyzed and processed by the computer and related software. According to the experimental equipment and research content, a digital twin simulation engine software development platform is constructed to realize the perception modeling of the coal mine robot operation process and the visualization of the digital twin monitoring effect.

[0012] An experimental method for a coal mine robot geological scene perception modeling experimental method specifically includes the following steps:

[0013] S1: Through the coal mine robot working environment simulation module and the track-type mobile platform, the dust concentration, illumination and air humidity in the box environment are changed and the operation process of the coal mine robot is simulated;

[0014] S2: The coal mine robot multi-source perception module obtains scene information in complex environments and coal seam geological simulation module information to build a joint geological and scene data set;

[0015] S3: Using the scene data obtained in step S2, research is conducted on noise modeling and data enhancement methods for scene observation multi-mode heterogeneous sensors;

[0016] S4: Using the geological data obtained in step S2, construct a radar echo forward and inversion data set to study the propagation characteristics of ground penetrating radar echo signals in cross-medium processes and the anomaly recognition method;

[0017] S5: Process the data through the experimental analysis module, study the association of geological and scene multi-scale feature data and the joint interpretation and perception enhancement methods of multi-source information, and complete the fusion perception of multi-source information;

[0018] S6: Use geological and scene multi-source fusion perception information to build a geological and scene integrated three-dimensional model, and use the boundary area characteristics of the geological model and the scene model to modify the integrated three-dimensional model;

[0019] S7: Use the truth measurement device to obtain the true value of the scene three-dimensional model and the coal mine robot posture, and compare them with the multi-source information fusion perception and three-dimensional model results obtained in step S5 and step S6 to evaluate the experimental effect of the coal mine robot geological scene perception modeling method.

[0020] Preferably, in step S1, the coal mine robot working environment simulation module controls the shutdown and operation of the dust generator, water vapor generator, lighting device, dehumidification and dust removal equipment, dust concentration detection sensor, illumination detection sensor and air humidity detection sensor through the central controller, and the application of the coal mine robot working environment simulation module can be arbitrarily combined and adjusted according to the actual experimental content.

[0021] Preferably, in step S2, ground penetrating radar and infrared camera are used to detect the coal seam geological simulation module and collect relevant data; laser radar, frequency modulated continuous wave radar and visible light camera block are used to obtain original information of scenes in complex environments.

[0022] Preferably, in step S3, based on control experiment analysis and understanding of the form and influence of the environment and working conditions on the original information of the sensor, a noise observation model of each sensor is established.

[0023] Preferably, in step S4, a simulated coal seam geological simulation module is established and designed according to the parameters of the constructed coal seam geological physical model using gprMax simulation software, and a forward simulation is performed on the simulation model.

[0024] Preferably, in step S5, the sensor data is processed by an experimental analysis module, a multi-scale visualization imaging method of the sensor is explored, and a multi-scale data characterization method is constructed, which specifically includes the following steps:

[0025] S51: By designing a temporal sliding window for temporal matching, a data association method for pixel-level alignment of sensor data on constructed artificial synthetic images is proposed;

[0026] S52: Artificially synthesized images constructed using multi-scale features and loss functions based on data association metrics;

[0027] S53: Use data-driven multi-task learning methods to infer the damage type and degree of multi-scale features and realize the joint interpretation of multimodal information;

[0028] S54: Perform feature prediction on artificially synthesized images under conditions where features are missing due to environmental interference, perform feature compensation on damaged information, and achieve enhanced sensor feature information perception under complex conditions.

[0029] Preferably, in step S6, the construction of the three-dimensional geological point cloud model is realized by using linear interpolation and point cloud discretization, which specifically includes the following steps:

[0030] S61: Acquire scene model data based on multi-sensor fusion simultaneous positioning and mapping technology to complete the construction of scene point cloud model;

[0031] S62: using the scene point cloud model and the geological point cloud model to fit the air-coal rock medium surface respectively;

[0032] S63: Cross-validate the accuracy of the scene point cloud model and the geological point cloud model based on the spatial continuity of the boundary area;

[0033] S64: Use the high-precision elevation model of the scene point cloud model to correct the geological point cloud model to achieve dynamic optimization;

[0034] S65: Obtaining coal-rock interface recognition results using the optimized scene point cloud model;

[0035] S66: Verify and update the medium physical properties of the geological point cloud model to achieve dynamic correction of the geological model through joint inversion of multi-source information;

[0036] S67: By constructing a new four-channel point cloud data storage data format, the intensity information is redesigned into data channels for reacting different media types such as gas-coal-rock on the basis of spatial three-dimensional coordinates and intensity information, so as to achieve consistent storage and integrated display of geological and scene point cloud models.

[0037] Therefore, the present invention adopts the above-mentioned coal mine robot geological scene perception modeling experimental system and method, which has the following beneficial effects:

[0038] (1) In the coal seam geological simulation module of the present invention, different stacking methods of coal blocks and rock blocks of different shapes can be used to change the coal seam geological simulation module, create different geological structures, and simulate typical structures such as faults, collapse columns, and cavities in real underground coal seams.

[0039] (2) The present invention can change the dust concentration, illumination, and air humidity in the box environment and simulate the operation process of the coal mine robot through the coal mine robot working condition simulation system and the track-type mobile platform equipped with sensors.

[0040] (3) The digital twin simulation engine software development platform constructed by the experimental analysis module of the present invention can realize the perception modeling of the coal mine robot operation process and the visualization of the digital twin monitoring effect.

[0041] (4) The coal mine robot geological scene perception modeling experimental system and method of the present invention are not only suitable for scientific research activities such as geological scene perception modeling of the coal mine robot operation process, but can also be used for scientific research on geological scene perception modeling of coal mining robots, rail-mounted inspection robots, tunneling robots and other operations.

[0042] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a structural schematic diagram of a coal mine robot geological scene perception modeling experimental system and method of the present invention;

[0044] Figure 2 This is a practical application diagram of a coal mine robot geological scene perception modeling experimental system and method of the present invention;

[0045] Figure 3It is a multi-sensor data fusion information weight experiment flow chart of a coal mine robot geological scene perception modeling experimental method of the present invention;

[0046] Figure 4 It is a sensor noise observation experiment flow chart of a coal mine robot geological scene perception modeling experiment method of the present invention;

[0047] Figure 5 This is a flow chart of the ground penetrating radar echo signal propagation characteristics and abnormal body recognition experiment of a coal mine robot geological scene perception modeling experimental method of the present invention;

[0048] Figure 6 This is a flow chart of an efficient modeling experiment of geological scene integration of a coal mine robot geological scene perception modeling experiment method of the present invention;

[0049] Figure Annotation

[0050] 1. Closed box; 2. Air intake fan; 3. Coal powder box; 4. Dust generator; 5. Water storage tank; 6. Water vapor generator; 7. Lighting device; 8. Dust concentration detection sensor; 9. Illumination detection sensor; 10. Air humidity detection sensor; 11. Visible light camera; 12. Infrared camera; 13. Coal block; 14. Rock block; 15. Collapse column; 16. Coal seam geological simulation module; 17. Coal seam cavity; 18. Fixing bolt; 19. Track-type mobile platform; 20. Ground penetrating radar; 21. LiDAR; 22. Frequency modulated continuous wave radar; 23. Sensor data collector; 24. Driving motor; 25. Horizontal guide rail; 26. Longitudinal guide rail; 27. Guide wheel; 28. Electric push rod; 29. ​​True value measurement device; 30. Exhaust fan; 31. Computer; 32. Data acquisition card; 33. Central controller; 34. Scene feature simulation device. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0052] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0053] The words "include" or "comprises" and the like used in the present invention mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of also including other elements. The orientation or position relationship indicated by the terms "inside", "outside", "upper", "lower", etc. is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. When the absolute position of the described object changes, the relative position relationship may also change accordingly. In the present invention, unless otherwise clearly specified and limited, the terms "attachment" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral body; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0054] like Figure 1 As shown, the present invention provides a coal mine robot geological scene perception modeling experimental system, including a coal mine robot working environment simulation module, a track-type mobile platform, a multi-source perception module, and an experimental analysis module.

[0055] A coal mine robot geological scene perception modeling experimental system includes a closed box 1 composed of a wooden board, an iron sheet, and an insulation layer. The closed box 1 is composed of wooden boards, the iron sheet is fixed to the inner surface of the closed box 1 by iron nails, and a layer of insulation material is attached to the inner surface of the iron sheet. The wooden board, the iron sheet, and the insulation layer together form the closed box 1 to maintain a stable temperature in the closed box 1.

[0056] It also includes a scene feature simulation device 34 arranged inside the closed box 1, a coal seam geological simulation module 16 arranged inside the closed box 1, and a true value measurement device 29 arranged inside the closed box 1, and a coal mine robot working environment simulation module arranged above the closed box 1; the material of the scene feature simulation device 34 is set to a wooden structure, and the coal seam geological simulation module 16 is stacked by coal blocks 13 and rock blocks 14, and the coal blocks and rock blocks are both cut from complete coal blocks and rock blocks according to requirements, and then the coal blocks 13 and rock blocks 14 are stacked together and fixed in front of the track-type mobile platform 19 with fixing bolts 18, and preset-sized gaps are stacked by coal and rock blocks of different shapes to simulate coal seam cavities 17, collapse columns 15 and other geological structures.

[0057] A track-type mobile platform 19 is arranged in front of the coal seam geological simulation module 16. Sensors are deployed in combination according to the experimental content to form a multi-source perception module. The sensors include a ground penetrating radar 20, a lidar 21, a frequency-modulated continuous wave radar 22, a visible light camera 11 and an infrared camera 12. The sensor data collector 23, the true value measurement device 29 and the coal mine robot working environment simulation module are all connected to the experimental analysis module. The experimental analysis module includes a computer 31 and related software. The scene information and the position information of the track-type mobile platform 19 and the sensor are obtained through the true value measurement device 29. The distance between the sensor and the coal seam geological simulation module 16 is changed through the track-type mobile platform 19 to change the perception information of the multi-source perception module.

[0058] The coal mine robot working environment simulation module controls the dust generator 4, the water vapor generator 6, the lighting device 7, the dehumidification and dust removal equipment, the dust concentration detection sensor 8, the illumination detection sensor 9, and the air humidity detection sensor 10 through the central controller 33; the dust generator 4 is arranged at the bottom of the coal powder box 3, the lighting device 7 includes a controllable lighting lamp, and the dehumidification and dust removal equipment includes an intake fan 2 and an exhaust fan 30. The data detected by the dust concentration detection sensor 8, the air humidity detection sensor 10, and the illumination detection sensor 9 will be transmitted to the computer 31 by the data acquisition card 32.

[0059] The coal powder in the coal powder box 3 is released into the closed box 1 through the dust generator 4 to change the dust concentration in the closed box 1, and the dust concentration detection sensor 8 is used to detect the dust concentration in the environment; water vapor is released into the environment through the water vapor generator 6 to change the air humidity in the environment, and the air humidity detection sensor 10 is used to detect the air humidity in the working space; the ambient illumination in the working environment is changed by changing the brightness of the lighting device 7, and the ambient illumination in the box is detected by the illumination detection sensor 9; when the dust concentration or air humidity in the working channel is too high, the dust concentration or air humidity in the environment is reduced by the air intake fan 2 and the exhaust fan 30 and the air pressure balance in the box 1 is maintained.

[0060] A guide wheel 27 and a drive motor 24 are arranged below the track-type mobile platform 19. The track-type mobile platform 19 includes a transverse guide rail 25 and a longitudinal guide rail 26. The surface of the transverse guide rail and the surface of the longitudinal guide rail are non-smooth surfaces. The transverse guide rail 15 is arranged at the bottom of the closed box. A guide rail groove is opened under the longitudinal guide rail 16. The guide wheel 27 contacts the longitudinal guide rail 26 through the guide rail groove. The drive motor 24 is connected to the guide wheel 27. The drive motor 24 drives the guide wheel 27 to move so that the platform moves longitudinally. An electric push rod 28 is installed on the left side of the longitudinal guide rail. The electric push rod 28 is connected to the longitudinal guide rail 26. The longitudinal guide rail 26 is pushed by the electric push rod 28 to make a transverse movement, so that the track-type mobile platform 19 can move in the guide rail plane, simulating the walking operation process of the coal mine robot. Figure 2As shown, the rail-type mobile platform can be realized based on the coal mining machine robot, and can also be replaced by a rail-mounted inspection robot, tunneling robot, etc.

[0061] The experimental analysis module includes computers and related software. The sensors transmit the collected data to the computer, and the data is analyzed and processed through the computer and related software. According to the experimental equipment and research content, a digital twin simulation engine software development platform is constructed to realize the perception modeling of the coal mine robot operation process and the visualization of the digital twin monitoring effect.

[0062] A method for a coal mine robot geological scene perception modeling experimental system, specifically comprising the following steps:

[0063] S1: Through the coal mine robot working environment simulation module and the track-type mobile platform, the dust concentration, illumination and air humidity in the box environment are changed and the operation process of the coal mine robot is simulated;

[0064] In step S1, the coal mine robot working environment simulation module controls the shutdown and operation of the dust generator, water vapor generator, lighting device, dehumidification and dust removal equipment, dust concentration detection sensor, illumination detection sensor and air humidity detection sensor through the central controller. The application of the coal mine robot working environment simulation module can be arbitrarily combined and adjusted according to the actual experimental content.

[0065] S2: The information of scenes in complex environments and information of coal seam geological simulation modules are obtained through the multi-source perception module of the coal mine robot, and a joint geological and scene data set is constructed; in step S2, ground penetrating radar and infrared camera are used to detect the coal seam geological simulation module and collect relevant data; laser radar, frequency modulated continuous wave radar and visible light camera blocks are used to obtain the original information of scenes in complex environments.

[0066] S3: Using the scene data obtained in step S2, the noise modeling and data enhancement method of multi-mode heterogeneous sensors for scene observation is studied; in step S3, based on the control experiment analysis and the form and influence of the environment and working conditions on the original information of the sensor, the noise observation model of each sensor is established.

[0067] S4: Using the geological data obtained in step S2, construct a radar echo forward and inversion data set to study the propagation characteristics of ground penetrating radar echo signals in cross-medium processes and the method of identifying anomalies; in step S4, use the gprMax simulation software to design and establish a simulated coal seam geological simulation module based on the constructed coal seam geological physical model parameters, and perform forward simulation on the simulation model.

[0068] S5: Process the data through the experimental analysis module, study the association of geological and scene multi-scale feature data and the multi-source information joint interpretation and perception enhancement method, and complete the multi-source information fusion perception; In step S5, the sensor data is processed through the experimental analysis module, the sensor multi-scale visualization imaging method is explored, and a multi-scale data characterization method is constructed, which specifically includes the following steps:

[0069] S51: By designing a temporal sliding window for temporal matching, a data association method for pixel-level alignment of sensor data on constructed artificial synthetic images is proposed;

[0070] S52: Artificially synthesized images constructed using multi-scale features and loss functions based on data association metrics;

[0071] S53: Use data-driven multi-task learning methods to infer the damage type and degree of multi-scale features and realize the joint interpretation of multimodal information;

[0072] S54: Perform feature prediction on artificially synthesized images under conditions where features are missing due to environmental interference, perform feature compensation on damaged information, and achieve enhanced sensor feature information perception under complex conditions.

[0073] S6: Use geological and scene multi-source fusion perception information to build a geological and scene integrated three-dimensional model, and use the boundary area characteristics of the geological model and the scene model to correct the integrated three-dimensional model. In step S6, linear interpolation and point cloud discretization are used to realize the construction of the three-dimensional geological point cloud model, which specifically includes the following steps:

[0074] S61: Acquire scene model data based on multi-sensor fusion simultaneous positioning and mapping technology to complete the construction of scene point cloud model;

[0075] S62: using the scene point cloud model and the geological point cloud model to fit the air-coal rock medium surface respectively;

[0076] S63: Cross-validate the accuracy of the scene point cloud model and the geological point cloud model based on the spatial continuity of the boundary area;

[0077] S64: Use the high-precision elevation model of the scene point cloud model to correct the geological point cloud model to achieve dynamic optimization;

[0078] S65: Obtaining coal-rock interface recognition results using the optimized scene point cloud model;

[0079] S66: Verify and update the medium physical properties of the geological point cloud model to achieve dynamic correction of the geological model through joint inversion of multi-source information;

[0080] S67: By constructing a new four-channel point cloud data storage data format, the intensity information is redesigned into data channels for reacting different media types such as gas-coal-rock on the basis of spatial three-dimensional coordinates and intensity information, so as to achieve consistent storage and integrated display of geological and scene point cloud models.

[0081] S7: Use the truth measurement device to obtain the true value of the scene three-dimensional model and the coal mine robot posture, and compare them with the multi-source information fusion perception and three-dimensional model results obtained in step S5 and step S6 to evaluate the experimental effect of the coal mine robot geological scene perception modeling method.

[0082] Example

[0083] A coal mine robot geological scene perception modeling experimental system and method, in the process of scene modeling experiment, such as Figure 3 and Figure 4 As shown, the coal mine robot working environment simulation module is used to create clean environment, single dust environment, single water vapor environment, single lighting environment, dust and water vapor environment, dust lighting environment, water vapor lighting environment, dust and water vapor lighting environment and complex environmental conditions such as different dust concentrations in the same environment, air humidity, and light intensity. The program controls the drive motor and electric push rod to make the track-type mobile platform move on the non-smooth guide rail to simulate the vibration conditions of the coal mine robot during operation. The platform is equipped with laser radar, visible light camera, infrared camera, and frequency-modulated continuous wave radar to collect scene information in various environments and construct original data sets. The data collected in different environments or with different parameter information in the same environment are compared, analyzed, and processed by computers and related software to explore the interference of the environment and working conditions on the original information of the sensor. The form and influence law of the noise observation model of each sensor are established, and methods to improve the information quality are proposed based on the experimental results; the scene modeling and multi-sensor modeling information fusion are carried out on the data collected by the sensor through the experimental analysis module, and compared with the scene model true value data obtained by the true value measurement device. The accuracy and error of each sensor scene modeling under different working conditions are explored. According to the accuracy and advantages and disadvantages of each sensor modeling, the allocation of information weights during multi-sensor data fusion is studied, and a method for constructing cross-modal features is proposed. By evaluating the strength and direction of constraints provided by different modal features, the optimization decision and learning methods are used to realize the automatic adjustment of weights for different environmental conditions during information fusion. Based on the factor graph optimization framework, real-time fusion of multi-sensor data is realized by building an efficient optimization solver. In the process of geological modeling experiments, such as Figure 5, through program-controlled drive motors and electric push rods, the track-type mobile platform moves on uneven guide rails to simulate the vibration conditions of coal mine robots during operation. The platform is equipped with ground-penetrating radar and infrared cameras to detect the constructed coal seam geological simulation module and collect original data. At the same time, the gprMax simulation software is used to design and establish a simulated coal seam geological simulation module according to the physical properties of the constructed coal seam geological simulation module, and the simulated coal seam geological simulation module is forward simulated. The forward data and the data actually collected by the ground-penetrating radar are compared and analyzed and tested through the experimental analysis module, and a data set corresponding to the wave field response characteristics of typical geological structures and the actual structural parameters is constructed. Combined with software simulation forward simulation and ground-penetrating radar simulation experiments, the echo signal propagation characteristics of the ground-penetrating radar in the gas-coal-rock cross-medium process are studied, and by changing the polarization direction of the ground-penetrating radar antenna, the coal seam geology Different physical parameters and other parameters are used to explore the influence mechanism of different conditions on the ground penetrating radar echo signal, and to clarify the relationship between dielectric constants, electromagnetic characteristics, radar frequency, antenna deployment and other factors in typical coal and rock geological structures; the software simulation and actual simulation experimental data are analyzed and processed through the experimental analysis module, and radar echo data sets of different geological structures are established to realize the identification of different geological anomalies, providing a basis for ground penetrating radar inversion imaging and modeling; the B-scan data collected in the experiment are processed using methods such as multi-profile and linear interpolation, and the spatial position and size of the geological structural anomaly are determined by the radar echo time difference and the propagation speed of electromagnetic waves in the medium, so as to realize the three-dimensional modeling of coal seam geology, and compare and analyze the three-dimensional geological model established by the simulation experiment with the coal seam geological simulation module with known structural dimensions, so as to explore the accuracy and error of geological three-dimensional modeling under different conditions. In the process of geological scene integrated modeling experiment, such as Figure 6 As shown, the position information of sensors such as laser radar, visible light camera, infrared camera, ground penetrating radar, etc. is obtained through a true value measurement device. The data information of the scene by laser radar, visible light camera, frequency modulated continuous wave radar, and the data information of the coal seam geological simulation module by ground penetrating radar and infrared camera are simultaneously collected through the relevant steps in the scene modeling experiment and the relevant steps in the geological modeling experiment. The collected scene and geological information are three-dimensionally modeled through the experimental analysis system. Combined with geological observation information, scene multi-source observation information and other constraints, the sensor position information is used to unify the coordinate systems of the two models to realize the integrated modeling of geological scenes. The accuracy and error of the integrated modeling of geological scenes are analyzed according to the scene truth obtained by the true value measurement device and the truth value of the geological model, the description method of the integrated model of geology and scenes is studied, and a multi-layer probability grid map construction strategy is proposed to realize the seamless connection and homogeneous expression of the geological model and the scene model. By designing a dynamic correction mechanism for mutual verification and joint inversion of the scene model and the geological model, efficient and high-precision modeling of the geology and scenes of the comprehensive mining working face is realized.

[0084] Therefore, the present invention adopts the above-mentioned coal mine robot geological scene perception modeling experimental system and method. Through the coal mine robot working environment simulation module and the track-type mobile platform, it can change the dust concentration, illumination, air humidity in the box environment and simulate the coal mine robot operation process. The multi-source perception module is used to obtain complex scene perception information, construct a data set, and conduct noise modeling and data enhancement method research of multi-mode heterogeneous sensors for scene observation; obtain geological data under complex environments, construct radar echo forward and inversion data sets, and explore the cross-medium process ground penetrating radar echo signal propagation characteristics and abnormal body recognition methods; process the data through computers and related software, explore multi-scale feature data association and multi-source information joint interpretation and perception enhancement methods, complete multi-source information fusion, and realize integrated real-time modeling of geological scenes. It can not only carry out scientific research activities such as integrated modeling of geological scenes in the operation process of coal mine robots, but also can be used for geological detection and inspection of coal mine tunnels through different combinations of sensors, and provide relevant information for tunnel stability and maintenance.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A coal mine robot geological scene perception modeling experimental system, comprising a closed box composed of a wood board, an iron sheet, and an insulation layer, a scene feature simulation device arranged on the left side of the closed box, and a true value measurement device arranged inside the closed box, characterized in that: It also includes a coal mine robot working environment simulation module arranged inside a closed box, a coal seam geology simulation module arranged inside a closed box, a track-type mobile platform arranged in front of the coal seam geology simulation module, and a multi-source perception module carried on the track-type mobile platform; the multi-source perception module, the truth measurement device and the coal mine robot working environment simulation module are all connected to the experimental analysis module, and the three-dimensional scene information and the position information of the track-type mobile platform and the sensor are obtained through the truth measurement device, and the distance between the multi-source perception module and the coal seam geology simulation module is changed through the track-type mobile platform, thereby changing the perception information of the multi-source perception module.

2. A coal mine robot geological scene perception modeling experimental system according to claim 1, characterized in that: The material of the scene feature simulation device is set to a wooden structure, and the position and posture are adjusted according to the experimental content to simulate different scenes; the true value measurement device includes a laser tracker and a laser scanner, which uses the laser scanner to obtain scene information, and uses the laser tracker and target to obtain the position and posture information of the track-type mobile platform and the sensor; the coal seam geological simulation module includes coal blocks and rock blocks of different shapes, which are cut from complete coal blocks and rock blocks according to requirements, and are quickly recombined into different simulated geological conditions by stacking coal blocks and rock blocks of different shapes.

3. A coal mine robot geological scene perception modeling experimental system according to claim 1, characterized in that: The coal mine robot working environment simulation module includes a dust generator, a water vapor generator, a lighting device, a dehumidification and dust removal equipment, a dust concentration detection sensor, an illumination detection sensor, and an air humidity detection sensor; the dust generator is arranged at the bottom of the coal powder box, the lighting device includes a controllable lighting lamp, and the dehumidification and dust removal equipment includes an intake fan and an exhaust fan.

4. A coal mine robot geological scene perception modeling experimental system according to claim 1, characterized in that: The multi-source perception module is composed of various sensors mounted on a tracked mobile platform and deployed jointly according to the experimental content. The sensors include ground penetrating radar, lidar, frequency-modulated continuous wave radar, visible light camera, infrared camera and inertial navigation unit.

5. The coal mine robot geological scene perception modeling experimental system according to claim 1 is characterized by: Guide wheels and a drive motor are arranged under the track-type mobile platform. The track-type mobile platform includes a transverse guide rail and a longitudinal guide rail. The surfaces of the transverse guide rail and the longitudinal guide rail are both non-smooth surfaces. The transverse guide rail is arranged at the bottom of the closed box. A guide rail groove is opened under the longitudinal guide rail. The guide rail groove is in contact with the transverse guide rail, and the guide wheel is in contact with the longitudinal guide rail. The drive motor is connected to the guide wheel, and the drive motor drives the guide wheel to move. An electric push rod is installed on the left side of the longitudinal guide rail, and the electric push rod is connected to the longitudinal guide rail. The longitudinal guide rail is pushed to make a transverse movement through the electric push rod.

6. A coal mine robot geological scene perception modeling experimental system according to claim 1, characterized in that: The experimental analysis module includes computers and related software. The sensors transmit the collected data to the computer, and the data is analyzed and processed through the computer and related software. According to the experimental equipment and research content, a digital twin simulation engine software development platform is constructed to realize the perception modeling of the coal mine robot operation process and the visualization of the digital twin monitoring effect.

7. An experimental method for a coal mine robot geological scene perception modeling experimental system according to any one of claims 1 to 6, characterized in that The specific steps include: S1: Through the coal mine robot working environment simulation module and the track-type mobile platform, the dust concentration, illumination and air humidity in the box environment are changed and the operation process of the coal mine robot is simulated; S2: The coal mine robot multi-source perception module obtains scene information in complex environments and coal seam geological simulation module information to build a joint geological and scene data set; S3: Using the scene data obtained in step S2, research is conducted on noise modeling and data enhancement methods for scene observation multi-mode heterogeneous sensors; S4: Using the geological data obtained in step S2, construct a radar echo forward and inversion data set to study the propagation characteristics of ground penetrating radar echo signals in cross-medium processes and the anomaly recognition method; S5: Process the data through the experimental analysis module, study the association of geological and scene multi-scale feature data and the joint interpretation and perception enhancement methods of multi-source information, and complete the fusion perception of multi-source information; S6: Use geological and scene multi-source fusion perception information to build a geological and scene integrated three-dimensional model, and use the boundary area characteristics of the geological model and the scene model to modify the integrated three-dimensional model; S7: Use the truth measurement device to obtain the true value of the scene three-dimensional model and the coal mine robot posture, and compare them with the multi-source information fusion perception and three-dimensional model results obtained in step S5 and step S6 to evaluate the experimental effect of the coal mine robot geological scene perception modeling method.

8. The method for modeling geological scene perception of a coal mine robot according to claim 7, characterized in that: In step S1, the coal mine robot working environment simulation module controls the shutdown and operation of the dust generator, water vapor generator, lighting device, dehumidification and dust removal equipment, dust concentration detection sensor, illumination detection sensor and air humidity detection sensor through the central controller. The application of the coal mine robot working environment simulation module can be arbitrarily combined and adjusted according to the actual experimental content.

9. The method for modeling geological scene perception of a coal mine robot according to claim 7, characterized in that: In step S2, ground penetrating radar and infrared camera are used to detect the coal seam geological simulation module and collect relevant data; laser radar, frequency modulated continuous wave radar and visible light camera blocks are used to obtain original information of scenes in complex environments.

10. The method for modeling geological scene perception of a coal mine robot according to claim 7, characterized in that: In step S3, based on the control experiment analysis and the form and influence of the environment and working conditions on the original information of the sensor, a noise observation model of each sensor is established.

11. The method for modeling geological scene perception of a coal mine robot according to claim 7, characterized in that: In step S4, the gprMax simulation software is used to design and establish a simulated coal seam geological simulation module according to the constructed coal seam geological physical model parameters, and a forward simulation is performed on the simulation model.

12. The method for modeling geological scene perception of a coal mine robot according to claim 7, characterized in that: In step S5, the sensor data is processed by the experimental analysis module, the multi-scale visualization imaging method of the sensor is explored, and a multi-scale data characterization method is constructed, which specifically includes the following steps: S51: By designing a temporal sliding window for temporal matching, a data association method for pixel-level alignment of sensor data on constructed artificial synthetic images is proposed; S52: Artificially synthesized images constructed using multi-scale features and loss functions based on data association metrics; S53: Use data-driven multi-task learning methods to infer the damage type and degree of multi-scale features and realize the joint interpretation of multimodal information; S54: Perform feature prediction on artificially synthesized images under conditions where features are missing due to environmental interference, perform feature compensation on damaged information, and achieve enhanced sensor feature information perception under complex conditions.

13. A coal mine robot geological scene perception modeling experimental method according to claim 7, characterized in that: In step S6, linear interpolation and point cloud discretization are used to construct a three-dimensional geological point cloud model, which specifically includes the following steps: S61: Acquire scene model data based on multi-sensor fusion simultaneous positioning and mapping technology to complete the construction of scene point cloud model; S62: using the scene point cloud model and the geological point cloud model to fit the air-coal rock medium surface respectively; S63: Cross-validate the accuracy of the scene point cloud model and the geological point cloud model based on the spatial continuity of the boundary area; S64: Use the high-precision elevation model of the scene point cloud model to correct the geological point cloud model to achieve dynamic optimization; S65: Obtaining coal-rock interface recognition results using the optimized scene point cloud model; S66: Verify and update the medium physical properties of the geological point cloud model to achieve dynamic correction of the geological model through joint inversion of multi-source information; S67: By constructing a new four-channel point cloud data storage data format, the intensity information is redesigned into data channels for reacting different media types such as gas-coal-rock on the basis of spatial three-dimensional coordinates and intensity information, so as to achieve consistent storage and integrated display of geological and scene point cloud models.

Images