An underground inspection robot with a radar system

Abstract

The invention discloses an underground inspection robot with a radar system, and the invention relates to the technical field of inspection robots, and solves the problem that there is no processing method for automatically determining an abnormal area to determine an abnormal model. The invention performs relevant detection in a channel based on a radar system, and then generates a relevant channel model belonging to the internal channel, and then decomposes the channel model into a plurality of model surfaces based on the model characteristics, and determines the center point between each model surface, and determines the corresponding abnormal model surface based on the specific change of the center point between each model surface, and performs overall determination of the corresponding abnormal model in the entire channel model based on the area parameters of the abnormal model surface and subsequent other model surfaces. This determination processing method can quickly and effectively complete the determination process of the corresponding abnormal model, and its determination rate is faster and has a wider range of use.

Description

Technical Field

[0001] The invention relates to the technical field of inspection robots, and in particular to an underground inspection robot with a radar system. Background Art

[0002] The underground inspection robot is an intelligent device specially designed to perform inspection tasks in complex and dangerous underground environments such as mines.

[0003] The application with publication number CN114879655A provides an intelligent robot inspection method for underground conveyor belts in coal mines, which utilizes an inspection robot to perform inspection movements along the transmission path of the conveyor belt in the coal mine, thereby synchronously collecting environmental data of the conveyor belt itself and its vicinity; analyzing the environmental data to determine the real-time operating status of the conveyor belt itself and the environmental safety status of the vicinity of the conveyor belt in the coal mine; and then, based on the determination result of the real-time operating status, instructing the host terminal to send a conveyor belt maintenance notification message; based on the determination result of the environmental safety status, instructing the host terminal to send a coal mine ventilation control instruction, thereby performing ventilation operations on the corresponding area. The inspection robot automatically collects and analyzes environmental data along the conveyor belt in the coal mine, thereby ensuring real-time and comprehensive troubleshooting of the conveyor belt and improving the safety of the working environment inside the coal mine.

[0004] When its inspection robot inspects underground, it can generally only perform underground detection based on the corresponding radar system, and generate relevant models from real-time detection data for output for external relevant personnel to view and determine whether there are abnormal areas in the well. However, the original judgment method can only be based on manpower for relevant judgments, and the underground pipelines are generally very long, so it will consume too much manpower when making judgments. There is no processing method for determining abnormal areas by itself to determine the abnormal model and display the determined abnormal model. The inspection robot needs to be enhanced. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an underground inspection robot with a radar system, which solves the problem of lacking a processing method for automatically determining abnormal areas to determine abnormal models.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an underground inspection robot with a radar system, comprising:

[0007] The radar detection end performs real-time detection of the front channel area of ​​the inspection robot and transmits the radar data confirmed by the real-time detection to the detection model generation end;

[0008] The detection model generating end generates a channel internal model of the channel area in real time according to the radar data generated by the radar detection end;

[0009] The abnormal model area determination end disassembles the model surface of the constructed channel internal model, determines the center point of the model surface based on the disassembled model surface, and then locks the abnormal model based on the numerical feature change between the center points, including a center point analysis and processing unit and an abnormal model confirmation unit;

[0010] The center point analysis and processing unit decomposes the generated channel internal model into several monomer model surfaces based on the inspection direction of the inspection robot, places the relevant monomer model surfaces in a two-dimensional coordinate system to lock the center point, and then transmits the several monomer model surfaces with the completed center point calibration to the abnormal model confirmation unit. The specific method is as follows:

[0011] According to the running rail on which the inspection robot slides, the inspection direction of the inspection robot is confirmed in real time. According to the confirmed inspection direction, a model surface perpendicular to the inspection direction is confirmed from the internal model of the channel. Starting from this model surface, the subsequent model surfaces that appear continuously in the internal model of the channel are disassembled in sequence to confirm several model surfaces to be processed;

[0012] Each model surface to be processed is placed in a set of two-dimensional coordinate systems, and based on the overall edge contour of the model surface to be processed, the point coordinates of several contour points in the overall edge contour are determined in the two-dimensional coordinate system, and then the determined several point coordinates are averaged to determine the mean coordinates, and the relevant points corresponding to the mean coordinates are locked in the two-dimensional coordinate system, and the relevant points are marked as the center points of the model surface to be processed;

[0013] The abnormal model confirmation unit performs point analysis on several monomer model surfaces that have completed center point calibration, determines the change relationship between the center points of adjacent monomer model surfaces, and determines the abnormal change node. Based on the determined abnormal change node and the area change between the monomer model surfaces, the abnormal model is determined in the channel internal model. The specific method is as follows:

[0014] Starting from the initial monomer model surface of the determined monomer model surfaces, determine the angle change between the center point of the initial monomer model surface and the center point of the adjacent monomer model surface: connect the two sets of center points to determine a set of angle straight lines, then construct a perpendicular line passing through the center point of the initial monomer model surface, and determine the angle J between the perpendicular line and the angle straight line. k , if J k >Y1, where k represents different adjacent center points, and Y1 is the preset angle value. The adjacent monomer model surface is marked as an abnormal model surface. If J k≤Y1, and then use the adjacent monomer model surface as the initial monomer model surface, and then change the angle of the center point of the adjacent monomer model surface, and so on, and confirm the subsequent abnormal model surfaces one by one;

[0015] Based on the determined abnormal model surface, determine the area parameters M of other monomer model surfaces following the abnormal model surface i , where i represents different monomer model faces, and the area parameter M i Compare it with the standard area MB of a group of single model faces in front of the abnormal model to determine its area parameter M i Satisfied: If Where X1 is a floating value:

[0016] If yes, other monomer model surfaces following the abnormal model surface are marked as subsidiary abnormal surfaces of the abnormal model surface;

[0017] If not, the calibration of the attached abnormal surface is stopped;

[0018] Combine the abnormal model surface and the subsequent related subsidiary abnormal surfaces to confirm the abnormal model;

[0019] The value difference analysis processing end performs correlation analysis on the center point of the determined abnormal model, and based on the center point change characteristics of the monomer model surface inside the abnormal model, identifies whether the obstacles inside such abnormal model are standard specifications obstacles, and performs relevant calibration; the specific method is:

[0020] The angle between the center points of adjacent monomer model surfaces in the mutation model is confirmed and calibrated as J k , where k represents different adjacent center points, if J k >Y1, the segments between the corresponding monomer model surfaces are marked as abnormal change segments. If Jk≤Y1, no relevant calibration is performed;

[0021] Record the total number of change segments G between adjacent monomer model surfaces in the abnormal model, and then record the number of abnormal change segments S, and determine the proportion value ZB = S ÷ G;

[0022] If ZB≤30%, it means that the obstacles in this abnormal model are standard obstacles, and they are directly displayed on the display terminal;

[0023] If ZB>30%, it means that the obstacle in the abnormal model is a non-standard obstacle, and it is directly displayed through the display terminal.

[0024] The present invention provides an underground inspection robot with a radar system. Compared with the prior art, it has the following beneficial effects:

[0025] The present invention performs relevant detection in the channel based on the radar system, and then generates a relevant channel model belonging to the internal channel, and then decomposes the channel model into a plurality of model surfaces based on the model characteristics, and determines the center point between each model surface, and determines the corresponding abnormal model surface based on the specific change of the center point between each model surface, and performs overall determination of the corresponding abnormal model in the entire channel model based on the area parameters of the abnormal model surface and subsequent other model surfaces. This determination processing method can quickly and effectively complete the determination process of the corresponding abnormal model, and its determination rate is faster and has a wider range of use;

[0026] Based on the determined abnormal model, the change characteristics of the internal center point of the abnormal model are confirmed to identify whether there are standard obstacles or non-standard obstacles in such abnormal model. Based on the specific identification results, a comprehensive assessment is conducted to facilitate unified management by subsequent relevant management parties. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] Example 1

[0030] See also Figure 1 The present application provides an underground inspection robot with a radar system, including a radar detection end, a detection model generation end, an abnormal model area determination end, a model library, a value difference analysis processing end and an output end, wherein the radar detection end is electrically connected to an input node of the detection model generation end, and the model library and the detection model generation end are both electrically connected to an input node of the abnormal model area determination end, wherein the abnormal model area determination end, the value difference analysis processing end and the output end are electrically connected in sequence from the output node to the input node;

[0031] The abnormal model area determination end includes a central point analysis and processing unit and an abnormal model body confirmation unit, and the central point analysis and processing unit is electrically connected to an input node of the abnormal model body confirmation unit;

[0032] Among them, the radar detection end is set in the front end area of ​​the inspection robot, and the inspection robot performs running inspection according to the preset running rail. It is set in the upper area of ​​the corresponding underground channel, performs real-time detection of the front end channel area of ​​the inspection robot, and transmits the radar data confirmed by the real-time detection to the detection model generation end;

[0033] The detection model generating end generates the channel internal model of the channel area in real time according to the radar data generated by the radar detection end. Since the method of determining the channel internal model by radar data is relatively common in the prior art, it will not be described in detail here. Generally, the pre-processed data is converted into a point cloud form, each point represents a measurement point on the channel surface, including its spatial position information, and a surface reconstruction algorithm, such as Poisson reconstruction, Marching Cubes algorithm, etc., is used to build a preliminary surface model of the channel according to the point cloud data, and the preliminary generated model is optimized, including smoothing, removing small defects and errors, filling holes, etc., to improve the quality and accuracy of the model;

[0034] Among them, the abnormal model area determination end disassembles the model surface of the constructed channel internal model, determines the center point of the model surface based on the disassembled model surface, and then locks the abnormal model based on the change of numerical characteristics between the center points. The specific execution end is:

[0035] The center point analysis and processing unit decomposes the generated channel internal model into several monomer model surfaces based on the inspection direction of the inspection robot, places the relevant monomer model surfaces in a two-dimensional coordinate system to lock the center point, and then transmits the several monomer model surfaces with the center point calibration to the abnormal model confirmation unit. The specific processing method is as follows:

[0036] According to the running rail on which the inspection robot slides, the inspection direction of the inspection robot is confirmed in real time (that is, the facing direction, which is generally perpendicular to the channel to be detected). According to the confirmed inspection direction, the model surface perpendicular to the inspection direction is confirmed from the internal model of the channel (if the initial model surface of the internal model of the channel is not perpendicular to the inspection direction, it will not be considered until a completely vertical model surface is determined, because the determined internal model of the channel may construct surrounding debris or irrelevant objects, thereby generating redundant model bodies and corresponding redundant model surfaces). Starting from this model surface, the subsequent model surfaces of the internal model of the channel are disassembled in turn to confirm several model surfaces to be processed (a body is composed of several surfaces, so a model body can be directly disassembled into several model surfaces);

[0037] Each model surface to be processed is placed in a set of two-dimensional coordinate systems, and based on the overall edge contour of the model surface to be processed, the point coordinates of several contour points in the overall edge contour are determined in the two-dimensional coordinate system, and then the determined several point coordinates are averaged to determine the mean coordinates, and the relevant points corresponding to the mean coordinates are locked in the two-dimensional coordinate system, and the relevant points are marked as the center points of the model surface to be processed;

[0038] Example: Assume that there is an inspection robot sliding on a running rail, and its inspection direction is determined to be due east.

[0039] From the internal model of the channel, find the model surface A that is perpendicular to the east direction. Assume that the previous model surfaces B and C are not perpendicular to the inspection direction, so they are not considered until the perpendicular model surface A is found.

[0040] Starting from model surface A, the subsequent model surfaces that appear continuously, such as model surfaces D, E, and F, are disassembled in sequence and determined as model surfaces to be processed.

[0041] For the model surface D to be processed, place it in a two-dimensional coordinate system. Assume that the overall edge contour of the model surface D contains the point coordinates of several contour points such as points (1,2), (3,4), (5,6), and (7,8).

[0042] Calculate the mean of these point coordinates. For example, if the mean of the horizontal coordinate is 4 and the mean of the vertical coordinate is 5, then the mean coordinate is (4,5). Lock the point corresponding to this mean coordinate in the two-dimensional coordinate system and mark it as the center point of the model surface D.

[0043] In the same way, model faces E, F and other model faces to be processed are processed to determine their respective center points;

[0044] The abnormal model confirmation unit performs point analysis on several monomer model surfaces that have completed center point calibration, determines the change relationship between the center points of adjacent monomer model surfaces, and determines the abnormal change node. Based on the determined abnormal change node and the area change between the monomer model surfaces, the abnormal model is determined in the channel internal model. The specific method of determining the abnormal model is:

[0045] Starting from the initial monomer model surface of the determined monomer model surfaces, determine the angle change between the center point of the initial monomer model surface and the center point of the adjacent monomer model surface: connect the two sets of center points to determine a set of angle straight lines, then construct a perpendicular line passing through the center point of the initial monomer model surface, and determine the angle J between the perpendicular line and the angle straight line. k , if J k>Y1, where k represents different adjacent center points, and Y1 is a preset angle value, the specific value of which is determined by the operator based on experience, and is generally a very small value of 5°. The adjacent monomer model surface is marked as an abnormal model surface. If J k ≤Y1, and then use the adjacent monomer model surface as the initial monomer model surface, and then change the angle of the center point of the adjacent monomer model surface, and so on, and confirm the subsequent abnormal model surfaces one by one;

[0046] Based on the determined abnormal model surface, determine the area parameters M of other monomer model surfaces following the abnormal model surface i , where i represents different monomer model faces, and the area parameter M i Compare it with the standard area MB of a group of single model faces in front of the abnormal model to determine its area parameter M i Satisfied: If Where X1 is a floating value, which is prepared in advance by the operator. If yes, the other monomer model surfaces following this abnormal model surface are calibrated as the subsidiary abnormal surfaces of this abnormal model surface. If no, the calibration of the subsidiary abnormal surfaces is stopped.

[0047] Combine the abnormal model surface and the subsequent related subsidiary abnormal surfaces to confirm the abnormal model;

[0048] Example: Suppose there is a series of monomer model faces arranged in sequence, and we start from the first monomer model face A.

[0049] The center point of surface A is O1, and the center point of its adjacent single-body model surface B is O2. Connect O1 and O2 to get an angle line, and construct a perpendicular line passing through O1. Assuming that the measured angle J1 between the perpendicular line and the angle line is 8°, which is greater than the preset angle value of 5°, B is calibrated as an abnormal model surface.

[0050] Then, the B surface is used as the initial single-body model surface to examine its angle with the adjacent C surface. Assuming that the angle J2 is 3°, which is less than 5°, the C surface is continued to be used as the initial single-body model surface to examine the angle with the next adjacent surface.

[0051] Assume that surface D is determined to be an abnormal model surface, and then measure the area parameter M1 of the subsequent single model surface E of surface D. Assume that the standard area MB of the single model surface before the abnormal model surface D is 10 square centimeters, and the floating value X1 is 2 square centimeters. If M1 is 7 square centimeters, which is not in the range of [8,12], then E is calibrated as an auxiliary abnormal surface of surface D. If M1 is 9 square centimeters, stop calibrating the auxiliary abnormal surface.

[0052] Finally, the abnormal model surface D and the attached abnormal surface E are combined to confirm the abnormal model;

[0053] Specifically, when debris accumulates or partially collapses and overflows in the corresponding channel, the center point of the first model surface will cause a drastic angle change. After the angle of such center point changes, the center point of the subsequent auxiliary abnormal surface may be the same as the center point angle of the abnormal model surface. In order to better determine the abnormal model, it is necessary to confirm the area to lock the corresponding abnormal area, so that the corresponding abnormal model can be confirmed.

[0054] Example 2

[0055] The value difference analysis processing end performs a correlation analysis of the center point of the determined abnormal model, and identifies whether the obstacle inside the abnormal model is a standard obstacle (for example, a stack of boxes is a standard obstacle) based on the center point change characteristics of the monomer model surface inside the abnormal model, and performs relevant calibration. The specific method of identification is as follows:

[0056] The angle between the center points of adjacent monomer model surfaces in the mutation model is confirmed and calibrated as J k , where k represents different adjacent center points, if J k >Y1, the segments between the corresponding monomer model faces are marked as abnormal change segments, otherwise, no relevant marking is performed;

[0057] Record the total number of change segments G between adjacent monomer model surfaces in the abnormal model, and then record the number of abnormal change segments S, and determine the proportion value ZB = S ÷ G;

[0058] If ZB≤30%, it means that the obstacles in this abnormal model are standard obstacles, and they are directly displayed on the display terminal;

[0059] If ZB>30%, it means that the obstacles in the abnormal model are non-standard obstacles, and they are directly displayed on the display terminal;

[0060] Example: Assume that a mutation model contains 5 adjacent monomer model faces, marked as A, B, C, D, and E respectively.

[0061] First, confirm that the angle between the center points of A and B is J1, the angle between the center points of B and C is J2, the angle between the center points of C and D is J3, and the angle between the center points of D and E is J4.

[0062] Assume that the preset angle value Y1 is 20°, J1=15°, J2=25°, J3=18°, and J4=12°.

[0063] Because J2 is greater than 20°, the segment between B and C is calibrated as an abnormal change segment, and other angles less than 20° are not calibrated.

[0064] Next, the total number of recorded change segments G is 4 (A and B, B and C, C and D, D and E), and the number of abnormal change segments S is 1 (B and C).

[0065] Calculate the proportion ZB = S ÷ G = 1 ÷ 4 = 0.25.

[0066] Since ZB=0.25 is less than 30%, it means that the obstacles in this abnormal model are obstacles of standard specifications, and are directly displayed through the display terminal.

[0067] Assume another variant model, which also has 5 adjacent monomer model faces, with angles of J1 = 28°, J2 = 35°, J3 = 10°, and J4 = 40°.

[0068] This time, the number of abnormally changed segments S is 3 (segments corresponding to J1, J2, and J4), and the total number of changed segments G is still 4.

[0069] Calculate the proportion ZB = 3 ÷ 4 = 0.75.

[0070] Because ZB=0.75 is greater than 30%, it means that the obstacle in this abnormal model is a non-standard obstacle, and is directly displayed through the display terminal.

[0071] Specifically, when the obstacle in the corresponding mutation model is a standard obstacle, the picture change characteristics inside it generally do not change too much. Only the center point at the beginning has a large change, and there is no large difference in the subsequent center point changes. Therefore, a comprehensive determination can be made based on the relevant changes between the corresponding center point values, which is convenient for external management parties to carry out relevant management.

[0072] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0073] The above embodiments are only used to illustrate the technical method 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 the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An underground inspection robot with a radar system, characterized in that: include: The radar detection end performs real-time detection of the front channel area of ​​the inspection robot and transmits the radar data confirmed by the real-time detection to the detection model generation end; The detection model generating end generates a channel internal model of the channel area in real time according to the radar data generated by the radar detection end; At the end of determining the abnormal model area, the model surface of the constructed internal channel model is disassembled, the center point of the model surface is determined based on the disassembled model surface, and then the abnormal model is locked based on the change of the numerical characteristics between the center points; The value difference analysis processing end performs correlation analysis on the center point of the determined abnormal model, and based on the center point change characteristics of the monomer model surface inside the abnormal model, it identifies whether the obstacles inside such abnormal model are standard specifications obstacles, and performs relevant calibration; The specific execution unit of the abnormal model area determination end includes a central point analysis and processing unit and an abnormal model confirmation unit; The center point analysis and processing unit decomposes the generated channel internal model into several monomer model surfaces based on the inspection direction of the inspection robot, places the relevant monomer model surfaces in a two-dimensional coordinate system to lock the center point, and then transmits the several monomer model surfaces with the completed center point calibration to the abnormal model confirmation unit; The abnormal model confirmation unit performs point analysis on several monomer model surfaces that have completed center point calibration, determines the change relationship between the center points of adjacent monomer model surfaces, and determines the abnormal change nodes. Based on the determined abnormal change nodes and the area changes between the monomer model surfaces, the abnormal model is determined in the channel internal model.

2. The underground inspection robot with a radar system according to claim 1, characterized in that: The specific method of the center point analysis and processing unit for performing center point calibration on a plurality of monomer model surfaces is as follows: According to the running rail on which the inspection robot slides, the inspection direction of the inspection robot is confirmed in real time. According to the confirmed inspection direction, a model surface perpendicular to the inspection direction is confirmed from the internal model of the channel. Starting from this model surface, the subsequent model surfaces that appear continuously in the internal model of the channel are disassembled in sequence to confirm several model surfaces to be processed; Each model surface to be processed is placed in a set of two-dimensional coordinate systems. Based on the overall edge contour of the model surface to be processed, the point coordinates of several contour points within its overall edge contour are determined in the two-dimensional coordinate system. Then, the determined point coordinates are averaged to determine the mean coordinates. The relevant points corresponding to the mean coordinates are locked in the two-dimensional coordinate system, and the relevant points are marked as the center points of the model surface to be processed.

3. The underground inspection robot with a radar system according to claim 2, characterized in that: The variant pattern confirmation unit determines the variant pattern in the following specific manner: Starting from the initial monomer model surface of the determined monomer model surfaces, determine the angle change between the center point of the initial monomer model surface and the center point of the adjacent monomer model surface: connect the two sets of center points to determine a set of angle straight lines, then construct a perpendicular line passing through the center point of the initial monomer model surface, and determine the angle J between the perpendicular line and the angle straight line. k , if J k >Y1, where k represents different adjacent center points, and Y1 is the preset angle value. The adjacent monomer model surface is marked as an abnormal model surface. If J k ≤Y1, and then use the adjacent monomer model surface as the initial monomer model surface, and then change the angle of the center point of the adjacent monomer model surface, and so on, and confirm the subsequent abnormal model surfaces one by one; Based on the determined abnormal model surface, determine the area parameters M of other monomer model surfaces following the abnormal model surface i , where i represents different monomer model surfaces, and the area parameter M i Compare it with the standard area MB of a group of single model faces in front of the abnormal model to determine its area parameter M i Satisfaction: If M i ∉[MB-X1,MB+X1], where X1 is a floating value: If yes, other monomer model surfaces following the abnormal model surface are marked as subsidiary abnormal surfaces of the abnormal model surface; If not, the calibration of the attached abnormal surface is stopped; Combine the abnormal model surface and the subsequent related subsidiary abnormal surfaces to confirm the abnormal model.

4. The underground inspection robot with a radar system according to claim 1, characterized in that: The specific method of the value difference analysis processing end for identifying whether the obstacle inside such abnormal model is a standard obstacle is as follows: The angle between the center points of adjacent monomer model surfaces in the mutation model is confirmed and calibrated as J k , where k represents different adjacent center points, if J k >Y1, mark the segments between the corresponding monomer model surfaces as abnormal change segments; Record the total number of change segments G between adjacent monomer model surfaces in the abnormal model, and then record the number of abnormal change segments S, and determine the proportion value ZB=S÷G; If ZB≤30%, it means that the obstacles in the abnormal model are standard obstacles, and are directly displayed through the display terminal.

5. The underground inspection robot with a radar system according to claim 4, characterized in that: If the proportion ZB>30%, it means that the obstacle in the abnormal model is a non-standard obstacle, and is directly displayed through the display terminal.

6. The underground inspection robot with a radar system according to claim 4, characterized in that: When Jk≤Y1, no relevant calibration is performed.

Images