Gis-based coal mine intelligent operation management system and method

Abstract

The application discloses a GIS-based intelligent operation management system and method for coal mines, and relates to the technical field of image processing, and comprises: a ring band reinforcement module, which generates an initial ring band reinforcement image according to an original image of a scale disc; an observation optimization module, which obtains an interference index based on an initial ring band region of the initial ring band reinforcement image, and collects an optimized observation image according to a reinforcement learning method and the interference index; and a ring band mask module, which determines a ring-shaped search region of the optimized observation image, and performs pixel marking on the optimized observation image according to a ring band contour point set corresponding to the ring-shaped search region, so as to obtain a ring band mask image; and the application improves the accuracy of roof risk zoning of the coal mine.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a GIS-based intelligent operation and management system for coal mines. Background Technology

[0002] Coal mine roadway roofs are prone to delamination under mining disturbances. To ensure roadway safety, roof delamination meters are usually installed at the roadway roof location. The amount of roof delamination is displayed through a dial and pointer. Mines are gradually adopting Geographic Information Systems (GIS) to bind the spatial location of the roof delamination meters to the coal mine roadway sections for roof risk zoning display and operation management. In the actual underground environment, the roof delamination meters are subjected to high humidity and temperature fluctuations for a long time. A damp water film area distributed along the circumference of the dial is easily formed on the surface of the dial or the inside of its transparent protective component. This causes the scale lines and pointer to appear blurred, ghosted, or misaligned in the image, thus affecting the automatic reading based on the image.

[0003] When a circumferentially distributed wet film area exists on the dial surface, the geometric relationship in the dial image will be inconsistent with the actual structure, causing deviation in pointer angle recognition and thus affecting the accuracy of delamination measurement. When using delamination measurement for coal mine roadway segment location and roof risk status assessment in GIS scenarios, the above deviation will lead to misjudgment of roadway segment risk status, affecting inspection arrangements and operation management decisions. Existing technologies mostly rely on image acquisition under a single observation posture or manual verification, which makes it difficult to obtain reliable delamination measurement data suitable for GIS operation management in humid environments, and the problem of unstable risk status updates still exists. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that make it difficult to obtain reliable delamination data suitable for GIS operation and management in humid environments, and to propose a GIS-based intelligent operation and management system and method for coal mines.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution: A GIS-based intelligent operation and management system for coal mines includes: The ring reinforcement module generates an initial ring reinforcement image based on the original image of the dial; The observation optimization module obtains the interference index based on the initial annular region of the initial annular enhanced image, and acquires optimized observation images according to the reinforcement learning method and the interference index. The annular mask module determines the annular search region of the optimized observation image and, based on the set of annular contour points corresponding to the annular search region, performs pixel marking on the optimized observation image to obtain the annular mask image. The ring correction module determines the geometric transformation relationship based on the ring mask image and performs coordinate transformation on the pixels of the target ring region in the optimized observation image to obtain the correction dial image. The operation management module manages the roof risk status of coal mine roadway sections based on the delamination amount corresponding to the calibration dial image.

[0006] Preferably, the specific steps for generating the initial annular enhanced image are as follows: A camera device is installed in front of the scale of the top plate separation instrument; Set the observation trajectory according to the center of the dial; Select multiple observation postures on the observation trajectory; The original image of the dial is captured using a camera device at each observation posture; Determine the maximum and minimum grayscale values ​​of each pixel coordinate in the original image; The grayscale range of pixel coordinates is determined based on the maximum and minimum grayscale values. The grayscale range is normalized to obtain the initial annular enhanced image.

[0007] Preferably, the specific steps for obtaining the interference index are as follows: Image segmentation is performed on the initial annular enhancement image to obtain the initial annular region; At each observation posture, the observation image of the dial is acquired by a camera device; Identify the set of target pixels that overlap with the initial annular region in the observed image; Determine the first grayscale average value of the target pixel set; Based on the target pixel set, determine the remaining pixel set of the initial annular region; Determine the second grayscale average value of the remaining pixel set; The interference index is obtained based on the first grayscale average value and the second grayscale average value.

[0008] Preferably, the specific steps for acquiring and optimizing observation images are as follows: Based on the reinforcement learning method, the observation attitude is iteratively adjusted to obtain the target observation attitude; where the target observation attitude refers to the observation attitude corresponding to the smallest disturbance index. Under the target observation posture, optimized observation images of the dial are acquired through a camera device.

[0009] Preferably, the specific steps for determining the annular search region of the optimized observation image are as follows: Determine the grayscale distribution of the initial annular enhancement image; Determine the initial center coordinates, initial inner radius, and initial outer radius of the initial annular region based on the grayscale distribution. Based on the initial center coordinates, initial inner radius, and initial outer radius, the annular search area for optimizing the observed image is determined.

[0010] Preferably, the specific steps for obtaining the annular mask image are as follows: Edge detection is performed on the annular search region to obtain a set of annular contour points; based on the set of annular contour points, pixel marking is performed on the optimized observation image to obtain an annular mask image.

[0011] Preferably, determining the geometric transformation relationship based on the annular mask image includes: Based on the annular mask image, contour points are extracted from the optimized observation image to obtain the inner edge contour point set; An ellipse is fitted to the set of inner edge contour points to obtain the coordinates of the ellipse center, the length of the major axis, and the length of the minor axis. Based on the coordinates of the ellipse's center, the length of its major axis, and the length of its minor axis, determine the geometric transformation relationships.

[0012] Preferably, the specific steps for obtaining the correction dial image are as follows: Based on the annular mask image, determine the target annular region for optimizing the observation image; Based on the geometric transformation relationship, the coordinates of the pixels in the target annular region are transformed to obtain the correction dial image.

[0013] Preferably, the operation and management of the roof risk status of coal mine roadway sections includes: The outer contour of the correction dial image is detected to obtain the center pixel coordinates of the dial. The scale markings of the calibration dial image are identified to obtain the 0-scale direction vector; Extract the pointer skeleton line from the corrected dial image; Endpoint detection is performed on the pointer skeleton lines to obtain the pixel coordinates of the pointer tip; The pointer direction vector is determined based on the center pixel coordinates and the pointer tip pixel coordinates; Determine the angle between the 0-point direction vector and the pointer direction vector; The amount of delamination is determined based on the included angle and the calibration constant of the top plate delamination instrument; The section of the coal mine roadway where the roof separation instrument is located; Based on the amount of delamination, the risk status of the roof in coal mine roadway sections is managed through operation.

[0014] To address the above problems, this invention also provides a GIS-based intelligent operation and management method for coal mines, the method comprising: An initial ring-enhanced image is generated based on the original image of the dial; The interference index is obtained from the initial annular region based on the initial annular enhancement image, and optimized observation images are acquired according to the reinforcement learning method and the interference index. The annular search region of the optimized observation image is determined, and the pixel marking of the optimized observation image is performed based on the set of annular contour points corresponding to the annular search region to obtain the annular mask image. The geometric transformation relationship is determined based on the annular mask image, and the coordinate transformation is performed on the pixels of the target annular region of the optimized observation image to obtain the correction dial image; Based on the delamination amount corresponding to the calibration dial image, the roof risk status of the coal mine roadway section is managed operationally.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention, through the construction of a ring-band enhancement, observation optimization, ring-band masking, and ring-band correction processing flow, can identify and weaken the influence of the ring-shaped optical influence area on the geometric relationship of the image when there is moisture film interference on the dial. This allows the dial image to restore the spatial relationship consistent with the actual mechanical structure, thereby improving the accuracy of pointer angle recognition and delamination calculation, and solving the problem of unstable image readings in humid environments.

[0016] 2. This invention uses reinforcement learning to iteratively optimize the observation posture, enabling the camera device to automatically tend towards an observation state with lower interference under various observation conditions. This avoids relying on single posture acquisition or manual verification, improves imaging stability and data reliability in complex downhole environments, and reduces the amplification effect of separation error in subsequent spatial aggregation from the source.

[0017] 3. The present invention further binds the geometrically corrected delamination data with the roadway spatial elements in the Geographic Information System (GIS) to achieve a precise correspondence between the delamination amount and the coal mine roadway segment. Based on the changes in the delamination amount, the roof risk status is dynamically updated and managed, thereby improving the accuracy of coal mine roof risk zoning and the timeliness of operation management. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a GIS-based intelligent operation and management method for coal mines, as provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] This embodiment provides a GIS-based intelligent operation and management system for coal mines, specifically including: The ring reinforcement module generates an initial ring reinforcement image based on the original image of the dial; In an embodiment of the present invention, the specific steps for generating the initial annular enhanced image are as follows: A camera device is installed in front of the scale of the top plate separation instrument; A roof delamination meter is a monitoring device installed on or connected to the roof structure of a coal mine roadway. It is used to obtain the relative displacement between different rock strata or support structures of the roadway roof. Through mechanical transmission, displacement conversion, or sensing, it converts the displacement changes inside or on the surface of the roof into readable display information to reflect the stability of the roof structure. The roof delamination meter includes a measuring component anchored to the roof and a scale or output interface for displaying the delamination amount. Its measurement results are used to assess the safety status of the roadway roof and provide basic data support for coal mine safety production and operation management.

[0021] The dial is a mechanical display component used to display the amount of delamination of the tunnel roof. It has a circular structure, and the center of the dial is the geometric center point of the dial, which is used as a reference for angle and position calculation. The camera device is an imaging device installed in front of the dial to acquire visual image information of the dial and its pointer.

[0022] First, determine the installation location of the roof separation meter in the coal mine roadway and clear the visible space in front of the dial to ensure no obstructions affect the imaging line of sight. Then, select an installation location with a fixed support structure and mechanically connect the camera device to the installation location via a bracket, ensuring the camera device is stably fixed in front of the dial and does not shift significantly with roadway vibrations. Next, adjust the spatial position of the camera device so that its imaging optical axis roughly points to the center area of ​​the dial. Check whether the dial is completely within the imaging frame by real-time display of the acquired images. After ensuring that the outer contour of the dial is fully displayed and the pointer and scale are clearly visible, tighten the fixing bolts of the camera device to prevent posture changes. At the same time, connect the power supply line and data transmission line of the camera device to complete the electrical and communication connections. Finally, conduct an image acquisition test to confirm that the dial is stably positioned in the imaging frame and that the image data can be output normally, thus completing the setup process of the camera device in front of the roof separation meter dial.

[0023] Set the observation trajectory according to the center of the dial; The observation trajectory is the spatial movement path formed when the position of the camera device relative to the dial changes, with the center of the dial as the spatial reference. It is used to limit the range of motion of the camera device.

[0024] The camera device acquires the original image of the dial and determines its outer contour. The set of pixels corresponding to the outer contour is then circularly fitted to obtain the center pixel coordinates of the dial. These center pixel coordinates are then converted into a spatial center position with the dial as a reference using the camera device's imaging calibration relationship. This spatial center position is then used as the geometric center of the observation trajectory. Next, the current spatial position and orientation of the camera device are acquired, and the optical axis of the camera device is aligned with the spatial center position while maintaining this alignment. Then, while ensuring the relative distance between the camera device and the dial changes continuously and the dial remains within the camera device's field of view, the camera device is moved around the spatial center position. The path formed by this spatial movement is the observation trajectory, which geometrically represents a spatial curve centered on the spatial center position. During the movement, the orientation of the camera device is continuously corrected to keep the dial center position near the center of the field of view in the imaging plane, thus obtaining the observation trajectory established around the dial center.

[0025] Select multiple observation postures on the observation trajectory; The observation posture refers to the specific position and orientation of the camera device on the observation trajectory. Each observation posture corresponds to an independent image acquisition condition.

[0026] The camera device is driven to different spatial positions relative to the dial along the observation trajectory. At each spatial position, the pose information of the camera device is acquired and the dial image is captured. The outer contour of the dial is detected in the image, and the pixel coordinates of the dial center in the image are calculated. At the same time, the imaging size of the outer contour of the dial and the imaging direction information of the dial plane are calculated to characterize the relationship between the current field of view coverage and the imaging angle. When the camera device moves continuously along the observation trajectory, it is continuously determined whether the dial center is located in the imaging plane and whether the dial outer contour is completely visible. If the dial outer contour is completely visible, the image corresponding to the current pose is used as a candidate observation pose sample. The candidate observation pose samples are traversed according to the distribution of their pose information on the observation trajectory. A set of samples is selected such that adjacent samples have differences in spatial position and orientation and the dial center remains stable and locatable in the image. The spatial position and orientation of the camera device corresponding to the sample are defined as multiple observation poses.

[0027] The original image of the dial is captured using a camera device at each observation posture; The original image refers to the dial image data directly acquired by the camera device under various observation postures. This image has not undergone image enhancement or numerical processing.

[0028] The camera is fixed in the spatial position and orientation corresponding to the observation posture to be captured, so that the imaging field of view of the camera covers the display area of ​​the dial and the outer contour of the dial is fully presented in the viewfinder. The lighting device is turned on and the lighting direction is adjusted to ensure that the dial surface receives stable illumination. The image acquisition function of the camera is started and an imaging exposure is completed. The acquired image data is stored in the form of a pixel array as the original image of the corresponding observation posture. At the same time, the correspondence between the original image and the observation posture is recorded. Then, the camera is driven to switch to the next observation posture in sequence and repeat the above imaging exposure and storage process until the original images of all observation postures are acquired, thereby obtaining the original image corresponding to each observation posture.

[0029] Determine the maximum and minimum grayscale values ​​of each pixel coordinate in the original image; Pixel coordinates are the location identifiers of the smallest spatial units that make up the original image. Each pixel coordinate corresponds to a specific brightness value. The maximum grayscale value refers to the highest brightness value that the same pixel coordinate appears in multiple original images, and the minimum grayscale value refers to the lowest brightness value corresponding to that pixel coordinate.

[0030] The original image sequence is uniformly aligned to ensure that the center of the dial in each original image is in the same position in the image coordinate system and that the outer contour of the dial corresponds to the same pixel area in each image. Then, the original image sequence is traversed using pixel coordinates as indices. For each pixel coordinate, the gray value of that pixel coordinate in each original image is read sequentially to form a gray value set. The gray value sets are compared and calculated to determine the largest gray value as the maximum gray value of that pixel coordinate and the smallest gray value as the minimum gray value of that pixel coordinate.

[0031] The grayscale range of pixel coordinates is determined based on the maximum and minimum grayscale values. Gray-scale range is the magnitude of brightness change reflected by the difference between the maximum and minimum gray-scale values ​​at the same pixel coordinates. This magnitude of change is used to characterize the response difference of the pixel under different observation conditions.

[0032] Using the pixel coordinates of the original image as the traversal index, each pixel position in the image is processed sequentially. For any pixel coordinate, the maximum and minimum gray values ​​corresponding to the pixel coordinate are read from the pre-obtained result data. The maximum gray value is used as the minuend and the minimum gray value is used as the subtrahend. Subtraction is performed to obtain the gray range of the pixel coordinate. The gray range is then recorded in a one-to-one correspondence with the pixel coordinate.

[0033] The grayscale range is normalized to obtain the initial annular enhanced image.

[0034] Normalization is a process of mapping grayscale differences to a uniform scale, so that the variation range of different pixels is within a comparable range. The initial ring-enhanced image refers to the image result formed after the grayscale differences have been normalized. The areas with significant brightness changes in this image correspond to the ring-shaped optical influence areas on the dial surface caused by factors such as condensation.

[0035] The grayscale range data corresponding to all pixel coordinates is traversed, and the grayscale range value corresponding to each pixel coordinate is read. During the traversal, the maximum and minimum values ​​among all grayscale ranges are determined by comparison. After obtaining the maximum and minimum values, the grayscale range data is processed again according to the pixel coordinate order. For any pixel coordinate, the minimum value is subtracted from the grayscale range corresponding to that pixel coordinate to obtain the translated grayscale difference value. Then, the translated grayscale difference value is divided by the difference between the maximum and minimum values ​​to obtain the normalized grayscale value within a uniform scale. Subsequently, the normalized grayscale values ​​are arranged and assigned values ​​according to the same pixel coordinate positions as the original image to form an image data structure with the same size as the original image. This image data structure is then output as an initial annular enhancement image, so that the brightness of each pixel in the image reflects the result of the grayscale change amplitude of the pixel under different observation postures after uniform scale transformation.

[0036] It should be noted that the annular optical influence area caused by factors such as condensation refers to the continuous or semi-continuous distribution of water film formed by the condensation of water vapor in the air under temperature changes on the surface of the dial of the top plate delamination instrument or its transparent protective component. The water film gathers in a band along the circumference of the dial, thus forming a humid area concentric with the circumference on the dial surface. This humid area refracts, reflects, and locally magnifies or shifts the incident light, causing changes in brightness, outline shift, or sharpness of the scale lines and pointer images passing through this area in the imaging plane. The band-shaped optical effect range formed by the water film is the annular optical influence area caused by factors such as condensation, and its spatial position is usually concentric with the center of the dial and extends around the circumference.

[0037] The observation optimization module obtains the interference index based on the initial annular region of the initial annular enhanced image, and acquires optimized observation images according to the reinforcement learning method and the interference index. In an embodiment of the present invention, the specific steps for obtaining the interference index are as follows: Image segmentation is performed on the initial annular enhancement image to obtain the initial annular region; Image segmentation refers to the process of dividing an image into several regions according to grayscale differences, used to distinguish the annular band region with relatively concentrated brightness changes from the initial annular enhanced image; the initial annular region refers to the set of pixels that are distributed in a ring and have obvious brightness changes in the initial annular enhanced image, and this region corresponds to the range of optical changes on the dial surface caused by condensation and other factors.

[0038] An initial annular enhanced image is acquired and its grayscale consistency is checked to confirm the integrity and usability of the image data. Then, grayscale values ​​are read pixel by pixel in the image to form grayscale distribution data. Based on the grayscale distribution, pixels with similar grayscale values ​​are divided into the same connected region, and pixels with significantly different grayscale values ​​are divided into different connected regions. Subsequently, connected regions that are spatially distributed in a ring and extend around the center of the scale are identified in all connected regions. The set of pixel coordinates corresponding to the connected region is determined as the initial annular region. At the same time, the correspondence between the pixel coordinates of the initial annular region and the initial annular enhanced image is recorded for subsequent position matching and overlap judgment in the observed image.

[0039] At each observation posture, the observation image of the dial is acquired by a camera device; An observation image refers to a scale image captured by a camera device under various observation postures, which reflects the imaging state of the scale in a specific spatial position and under lighting conditions.

[0040] The camera device is sequentially adjusted to the spatial position and orientation corresponding to each observation posture along the observation trajectory, so that the imaging field of view of the camera device covers the display area of ​​the dial and the outer contour of the dial is fully presented in the viewfinder. After maintaining stable lighting and checking that there are no obvious obstructions on the dial surface, the camera device is started to perform exposure imaging, acquire a dial observation image, and store the observation image as the image data corresponding to the current observation posture. At the same time, the correspondence between the observation image and the current observation posture is recorded. Then, the camera device is adjusted to the next observation posture and the above imaging and storage process is repeated until the dial observation images under all observation postures are acquired, thereby obtaining an observation image corresponding to each observation posture.

[0041] Identify the set of target pixels that overlap with the initial annular region in the observed image; The target pixel set refers to the set of pixels in the observed image that corresponds spatially to and overlaps with the initial annular region. This set represents the portion of the observed image affected by annular optics.

[0042] An observation image corresponding to the observation posture is acquired, and the pixel coordinate set of the initial annular region is obtained. Then, the outer contour of the dial is detected in the observation image, and the pixel coordinates of the dial center are determined. At the same time, the pixel coordinates of the dial center are determined in the initial annular enhanced image. A pixel coordinate alignment relationship is established by making the center pixel coordinates of the two images consistent. The pixel coordinate set of the initial annular region is mapped to the pixel coordinate system of the observation image according to this alignment relationship, and the pixel coordinate set corresponding to the annular region in the coordinate system of the observation image is obtained. Then, the pixel coordinate set corresponding to the annular region is traversed one by one in the observation image, and the pixel coordinates that are located within the effective imaging range of the observation image and whose gray values ​​can be read are selected. The selected pixel coordinates and their corresponding pixel gray values ​​are jointly determined as the target pixel set, thereby obtaining the overlapping pixel set that corresponds to the initial annular region in spatial position and is observable in the observation image.

[0043] Determine the first grayscale average value of the target pixel set; The first grayscale average value is the average brightness obtained by summing the grayscale values ​​of all pixels in the target pixel set and dividing by the number of pixels. It is used to represent the overall brightness level of the area.

[0044] Obtain the coordinates of all pixels in the target pixel set and their corresponding gray values. Iterate through the target pixel set and sum the gray values ​​of each pixel to obtain the total gray value. At the same time, count the number of pixels in the target pixel set. After the traversal is completed, divide the total gray value by the number of pixels to obtain the first average gray value.

[0045] Based on the target pixel set, determine the remaining pixel set of the initial annular region; The remaining pixel set refers to the pixel set after removing the target pixel set within the initial annular region. It represents the pixels in the same annular region that do not overlap with the currently observed image.

[0046] Obtain the pixel coordinate set of the initial annular region and the pixel coordinate set of the target pixel set in the observation image coordinate system. First, organize both using a consistent pixel coordinate representation method so that each pixel coordinate is uniquely identified by the same row and column index. Then, using the pixel coordinate set of the initial annular region as the traversal object, read each pixel coordinate in turn and perform an inclusion check on the pixel coordinate set of the target pixel set. If the pixel coordinate exists in the target pixel set, mark it as a removed pixel coordinate. If the pixel coordinate does not exist in the target pixel set, retain it as a remaining pixel coordinate. After the traversal is completed, collect all the retained remaining pixel coordinates to form a remaining pixel set.

[0047] Determine the second grayscale average value of the remaining pixel set; The second grayscale average value is the brightness value obtained by averaging the grayscale values ​​of all pixels in the remaining pixel set.

[0048] Obtain the observation image corresponding to the current observation posture and the pixel coordinate list of the remaining pixel set corresponding to the observation posture. Access the pixel gray value of the corresponding position in the observation image one by one according to the order of the pixel coordinate list, and accumulate the read pixel gray values ​​to form a gray sum. At the same time, count the number of pixels that have been accumulated. After completing the gray reading of all remaining pixel coordinates, divide the gray sum by the number of pixels to obtain the second gray average value.

[0049] The interference index is obtained based on the first grayscale average value and the second grayscale average value.

[0050] The interference index is obtained by subtracting the second gray-scale average from the first gray-scale average. The interference index is a quantitative value used to characterize the degree of influence of the annular optical influence area on the brightness distribution of the image during the dial imaging process. This value reflects the difference in brightness performance between the annular area affected by condensation, water film or surface medium changes and the surrounding area. When the annular optical influence area produces obvious refraction, reflection or local brightness shift to the scale line and pointer image, the value corresponding to the interference index changes, thereby indicating the strength of the influence of external optical factors on the dial imaging under the current observation conditions. The interference index is used to measure the degree of influence of the observation posture on the stability and readability of the dial image.

[0051] During imaging, the grayscale value of each pixel on the dial originates from the light intensity distribution after the incident light is reflected and refracted on its surface. There is a monotonic relationship between light intensity and grayscale value. When condensation or a water film is present on the dial surface, the light propagation path within the annular region changes, causing a shift in the light intensity distribution in that region relative to the unaffected region. According to the fundamental laws of optical reflection and refraction, if the medium state of a certain spatial region changes under the same incident light conditions, a stable difference will form between the reflected light intensity of that region and the adjacent unchanging region. Therefore, the grayscale values ​​of pixels within the annular region that overlap with the currently observed image are analyzed. By averaging, the overall brightness level of the optically affected area can be obtained. Averaging the pixel gray values ​​of the parts of the same annular region that do not overlap with the currently observed image can yield the brightness level of the relatively unaffected part. The difference between the two reflects the degree of overall light intensity shift caused by the change in optical propagation conditions under the observation posture. The larger the difference, the more obvious the optical disturbance caused by factors such as condensation on the imaging. Therefore, using the difference between the first gray value average and the second gray value average as an interference index, we can quantitatively characterize the strength of the effect of the annular optically affected area on the dial image based on the basic law of light intensity change.

[0052] In an embodiment of the present invention, the specific steps for acquiring and optimizing observation images are as follows: Based on the reinforcement learning method, the observation attitude is iteratively adjusted to obtain the target observation attitude; where the target observation attitude refers to the observation attitude corresponding to the smallest disturbance index. Reinforcement learning is an adaptive decision-making method based on trial and error and feedback. It obtains the values ​​of interference index under different observation postures and adjusts the spatial position and orientation of the camera device according to the changes of the interference index, thereby gradually approaching the observation state with less interference.

[0053] Iterative adjustment refers to the gradual correction of the spatial position and orientation of the camera device based on the interference index results corresponding to the current observation posture during multiple consecutive observations and evaluations, so that the subsequent observation posture changes in space.

[0054] The target observation attitude refers to the spatial position and orientation of the camera device with the lowest corresponding interference index value among all the observation attitudes tried. Under this attitude, the dial is less affected by the ring optical influence area.

[0055] First, determine the adjustable observation attitude parameters of the camera device. These parameters include the camera's pitch angle, yaw angle, and relative distance to the scale of the top plate exometry instrument. Then, treat the adjustment of one or more observation attitude parameters as a single action in reinforcement learning. Use the interference index corresponding to the acquired image after each adjustment as the feedback basis for reinforcement learning. Next, using the initial observation attitude as the starting point for iteration, drive the camera device to adjust the observation attitude parameters according to the action logic of reinforcement learning and acquire observation images of the scale of the top plate exometry instrument in each adjusted observation attitude. Simultaneously, calculate the corresponding interference index value for the observation images in each observation attitude according to the interference index calculation method. Then, use the change in the interference index value as the reward criterion for reinforcement learning. If the interference index value decreases after adjusting the observation attitude, the reward is awarded accordingly. If the interference index value is low, a positive reward is given for the action. If the interference index value increases after adjusting the observation posture, a negative reward is given for the action. If the interference index value does not change after adjusting the observation posture, a zero reward is given for the action. Then, based on the reward determination results, the observation posture parameters of the camera device are continuously adjusted iteratively. Each iteration uses the interference index value of the previous observation posture as a reference and corrects the parameters towards the observation posture with a lower interference index value. During the iteration process, the parameter combination of the observation posture after each adjustment and the corresponding interference index value are continuously recorded until multiple rounds of observation posture adjustment are completed and the interference index value no longer decreases significantly. Finally, the observation posture with the smallest corresponding interference index value is selected from all the observation postures that have been adjusted and recorded. This observation posture is the target observation posture.

[0056] Under the target observation posture, optimized observation images of the dial are acquired through a camera device.

[0057] Optimized observation images refer to dial images acquired by a camera device under the target observation posture. These images have lower optical disturbance effects compared to images under other observation postures and are used for subsequent geometric correction and delamination calculation.

[0058] The spatial position and orientation parameters of the camera device corresponding to the target observation posture are read, and the camera device is driven to move along the observation trajectory to the spatial position and adjust to the orientation so that the imaging optical axis of the camera device points to the center area of ​​the dial. Then, the dial image is acquired through real-time viewing, the outer contour of the dial is detected and it is confirmed that the dial is completely in the imaging field of view, and the pointer and scale mark are confirmed to be in the imaging area and that there are no objects obstructing the imaging optical path. Then, the illumination conditions are kept stable and an imaging exposure is performed to acquire the dial image data under the target observation posture. The image data is stored in the form of a pixel array and its corresponding observation posture mark information is recorded. Finally, the integrity check of the acquired image data is performed to confirm that the image data is readable and the dial area can be located. The dial image that passes the check is determined as the optimized observation image.

[0059] The annular mask module determines the annular search region of the optimized observation image and, based on the set of annular contour points corresponding to the annular search region, performs pixel marking on the optimized observation image to obtain the annular mask image. In an embodiment of the present invention, the specific steps for determining the annular search region of the optimized observation image are as follows: Determine the grayscale distribution of the initial annular enhancement image; Gray-scale distribution refers to the arrangement and range of brightness values ​​of each pixel in a gray-scale image within a spatial range. It reflects the magnitude and variation of light intensity received at different locations in the image. Locations with higher gray-scale values ​​correspond to areas with greater reflected light intensity on the imaging plane, while locations with lower gray-scale values ​​correspond to areas with weaker light intensity.

[0060] A full pixel-by-pixel traversal operation is performed on the initial annular enhancement image. The grayscale value corresponding to the row and column positions of each pixel in the image is read sequentially. The row and column position information of each pixel is associated with the corresponding grayscale value and stored to form a complete grayscale value and pixel position mapping dataset. This mapping dataset directly presents the spatial arrangement and value variation range of the brightness values ​​of each pixel in the entire image. The grayscale value is positively correlated with the intensity of reflected light received at the corresponding position on the imaging plane. The higher the grayscale value, the greater the intensity of reflected light received at the corresponding position, and the lower the grayscale value, the smaller the intensity of reflected light received at the corresponding position. This mapping dataset fully presents the intensity of light at different positions in the image and its variation law, providing a clear and quantifiable brightness distribution basis for the subsequent extraction of relevant geometric parameters of the initial annular region.

[0061] Determine the initial center coordinates, initial inner radius, and initial outer radius of the initial annular region based on the grayscale distribution. The initial center coordinates refer to the geometric center position of the annular region in the image coordinate system, which is used to represent the spatial reference point of the annular region in the image; the initial inner radius refers to the distance from the initial center coordinates to the inner boundary of the annular region, and the initial outer radius refers to the distance from the initial center coordinates to the outer boundary of the annular region. Together, they define the width range of the annular region in the image.

[0062] The coordinates of all pixels in the annular connected region are subjected to circular fitting. The geometric center of the annular region is calculated using the least squares method. This geometric center is the initial center coordinate. Using the initial center coordinate as the reference point, radial distances are calculated to the inner and outer boundaries of the annular connected region. All pixels on the inner and outer boundaries of the annular connected region are traversed, and the straight-line distance from each inner boundary pixel to the initial center coordinate is calculated. The minimum value is taken as the initial inner radius. At the same time, all pixels on the outer boundary of the annular connected region are traversed, and the straight-line distance from each outer boundary pixel to the initial center coordinate is calculated. The maximum value is taken as the initial outer radius. The initial inner radius and the initial outer radius together define the width range of the initial annular region in the image.

[0063] Based on the initial center coordinates, initial inner radius, and initial outer radius, the annular search area for optimizing the observed image is determined.

[0064] The annular search region refers to the ring-shaped pixel range in the optimized observation image, centered on the initial circle center coordinates and bounded by the initial inner radius and the initial outer radius. This region is used to define the spatial range for subsequent precise positioning of the annular contour.

[0065] In the optimized observation image, the outer contour of the dial is detected to determine the pixel coordinates of the dial's center. These center pixel coordinates are used as the spatial reference point of the optimized observation image. Subsequently, the determined initial center coordinates in the initial annular enhancement image are obtained and aligned with the center pixel coordinates of the optimized observation image to establish a pixel coordinate correspondence between the initial annular enhancement image and the optimized observation image. Then, in the optimized observation image, with the aligned initial center coordinates as the center, the distance from each pixel coordinate to the center coordinate is calculated. Pixel coordinates whose distance is greater than or equal to the initial inner radius and less than or equal to the initial outer radius are determined to be located within the annular search area. The pixel coordinates that meet the distance condition are collected to form the set of pixel coordinates of the annular search area of ​​the optimized observation image, thereby obtaining the annular search area used to limit the range of subsequent edge detection and contour point extraction.

[0066] In an embodiment of the present invention, the specific steps for obtaining the annular mask image are as follows: Edge detection is performed on the annular search region to obtain the set of annular contour points; Edge detection refers to the process of identifying locations in an image where brightness changes drastically. Brightness changes correspond to locations where the light reflection or refraction state of an object's surface changes significantly. The annular contour point set refers to the set of pixels corresponding to the brightness change boundaries detected within the annular search area. This set forms a band-shaped boundary distributed around the center of the scale in space, used to represent the actual boundary position of the annular area.

[0067] The process involves acquiring an optimized observation image and a set of pixel coordinates for the determined annular search region, and then performing subsequent processing only on pixels within the annular search region. Next, the grayscale value of each pixel within the annular search region is read sequentially, and the grayscale difference between that pixel and its neighboring pixels is calculated. By comparing the grayscale differences of neighboring pixels, locations with significant brightness changes are identified, and pixels with grayscale differences greater than the average grayscale difference level of their surrounding pixels are determined as edge candidate pixels. Then, connectivity analysis is performed on all edge candidate pixels to filter out pixels that are spatially distributed in a band around the center of the scale and are continuously arranged. This continuously arranged set of pixels is determined as the annular contour point set.

[0068] Based on the set of annular contour points, the optimized observation image is pixel-marked to obtain the annular mask image.

[0069] Pixel marking refers to distinguishing and recording the pixel positions corresponding to the set of points on the annular contour in the optimized observation image, so that the pixels in this part are different from other pixels in the image in terms of numerical representation; the annular mask image refers to the image data formed based on the pixel marking results, in which pixels belonging to the annular region are numerically distinguished from pixels not belonging to the annular region, and are used to represent the spatial distribution range of the optically affected area on the dial surface.

[0070] A mask image carrier with the same size as the pixel coordinate range of the optimized observation image is established, and all pixels of the mask image carrier are initialized to a uniform background marker value. Then, using the pixel coordinates in the annular contour point set as the starting boundary, the annular contour point set is connected and filled in the mask image carrier to determine the band-shaped closed region enclosed by the inner and outer contours. Pixels within this band-shaped closed region are assigned a uniform annular marker value, while pixels outside the band-shaped closed region are kept to the background marker value. Then, the mask image carrier is checked for integrity to confirm that the band-shaped closed region is a continuous region and that its pixel coordinates are consistent with the corresponding pixel positions in the optimized observation image. Finally, the mask image carrier is output as an annular mask image, so that the pixel positions corresponding to the annular marker values ​​in the annular mask image represent the annular regions in the optimized observation image, and the pixel positions corresponding to the background marker values ​​represent the non-annular regions in the optimized observation image.

[0071] The ring correction module determines the geometric transformation relationship based on the ring mask image and performs coordinate transformation on the pixels of the target ring region in the optimized observation image to obtain the correction dial image. In an embodiment of the present invention, determining the geometric transformation relationship based on the annular mask image includes: Based on the annular mask image, contour points are extracted from the optimized observation image to obtain the inner edge contour point set; Contour point extraction refers to the process of identifying the boundary positions where brightness or texture changes significantly along the boundary of the annular region in the optimized observation image, thereby obtaining the boundary pixel coordinates; the inner edge contour point set refers to the set of boundary pixel coordinates continuously distributed on the boundary of the annular region near the center of the dial, used to characterize the geometry of the inner boundary of the annular region.

[0072] The pixel coordinates marked as annular regions in the annular mask image are traversed, and the corresponding grayscale value in the optimized observation image is read at the pixel position of each annular region. Then, the neighborhood of each pixel within the annular region is checked. When there is a grayscale change between a pixel in an annular region and its neighboring pixels in the direction of the dial center, and the neighboring pixels do not belong to the annular region, the pixel in the annular region is determined as the inner boundary pixel of the annular region. Next, the pixel coordinates of all pixels determined as the inner boundary pixels of the annular region are summarized, and isolated pixels are removed by connectivity judgment, so that the remaining pixel coordinates form a continuously distributed boundary curve in space. Finally, the set of continuously distributed pixel coordinates located at the inner boundary position of the annular region is determined as the inner edge contour point set.

[0073] An ellipse is fitted to the set of inner edge contour points to obtain the coordinates of the ellipse center, the length of the major axis, and the length of the minor axis. Ellipse fitting refers to the process of approximating the boundary shape represented by the set of inner edge contour points using elliptical geometry. It reflects how the originally nearly circular ring boundary appears as an ellipse in the image due to the tilt of the camera's viewing angle or the projection of the image. The ellipse center coordinates refer to the geometric center position of the fitted ellipse in the image coordinate system, corresponding to the position of the dial center in the projected image. The length of the ellipse's major axis is the length of the longest diameter of the ellipse, and the length of the ellipse's minor axis is the length of the shortest diameter perpendicular to the major axis. Both reflect the degree of stretching and compression of the ellipse shape and are related to the projection compression caused by the imaging viewing angle.

[0074] The pixel coordinates of all contour points in the inner edge contour point set are taken, and the pixel coordinates are deduplicated and valid to ensure that each contour point corresponds to a valid pixel position in the image. Then, the pixel coordinates of all contour points are used as fitting input to construct a parameter set to describe the shape of the ellipse. This parameter set includes at least the position of the ellipse center in the image coordinate system and the scale of the ellipse in two orthogonal directions. The fitting objective is to minimize the overall deviation of each contour point from the boundary of the ellipse. The parameter set is iteratively updated until the matching degree between the ellipse boundary and the contour point set no longer improves, thereby obtaining an ellipse corresponding to the geometry of the inner edge contour point set. After obtaining the fitted ellipse, the position of the ellipse center in the image coordinate system is read from the parameter set as the ellipse center coordinates, and the maximum and minimum scales in the two orthogonal directions of the ellipse are read as the lengths of the major and minor axes of the ellipse. The correspondence between the ellipse center coordinates, the lengths of the major and minor axes of the ellipse and the inner edge contour point set is established and recorded for subsequent determination of geometric transformation relationships.

[0075] Based on the coordinates of the ellipse's center, the length of its major axis, and the length of its minor axis, determine the geometric transformation relationships.

[0076] Geometric transformation relationships refer to coordinate transformation relationships used to scale and translate image coordinates, so that the boundary of the elliptical ring is close to a circle after coordinate transformation, thereby reducing the impact of deformation caused by imaging projection on dial reading processing.

[0077] The coordinates of the ellipse center are obtained and used as the reference origin for coordinate transformation. The coordinates of any pixel in the optimized observation image are converted into relative coordinates based on the coordinates of the ellipse center, ensuring that the position remains unchanged at the ellipse center during subsequent transformations. Then, the lengths of the ellipse's major and minor axes are obtained, and the scale compensation relationship in the major and minor axis directions is determined based on their ratio. This causes the coordinate scale corresponding to the minor axis direction to expand according to the ratio of the major and minor axis lengths, or to compress according to the ratio of the major and minor axis lengths, thus making the ellipse approach a uniform scale after coordinate scale transformation. Next, the direction of the scale compensation relationship in the image coordinate system is determined. The direction in the image coordinate system consistent with the ellipse's major axis direction is taken as the first coordinate direction, and the direction perpendicular to it is taken as the second coordinate direction. Corresponding scale transformations are applied to the first and second coordinate directions respectively. Finally, the scale transformation results of the relative coordinates are converted back to absolute coordinates expressed in the original image coordinate system, forming a mapping rule from the original pixel coordinates to the transformed pixel coordinates, and this mapping rule is determined as a geometric transformation relationship.

[0078] In an embodiment of the present invention, the specific steps for obtaining the correction dial image are as follows: Based on the annular mask image, determine the target annular region for optimizing the observation image; The target annular region refers to the set of strip-shaped pixels in the optimized observation image that corresponds one-to-one with the region identified by the annular mask image in pixel coordinates, and is used to represent the spatial range that needs to be geometrically corrected.

[0079] The process involves acquiring an optimized observation image and a corresponding annular mask image with pixel coordinates, ensuring consistency in spatial dimensions and pixel coordinate systems. The annular mask image is then traversed pixel-by-pixel, with each pixel's marker value recorded. When a pixel's marker value indicates it belongs to the annular region, its corresponding pixel coordinates in the optimized observation image are recorded as target pixel coordinates. All recorded target pixel coordinates are then aggregated according to their spatial positions in the optimized observation image, forming a set of pixel coordinates for the target annular region. A correspondence between this set and the optimized observation image is established. Finally, the completeness of the resulting set of pixel coordinates is checked to confirm its continuous, banded spatial distribution within the dial area. This set of pixel coordinates is then identified as the target annular region of the optimized observation image.

[0080] Based on the geometric transformation relationship, the coordinates of the pixels in the target annular region are transformed to obtain the correction dial image.

[0081] Coordinate transformation refers to the process of remapping the spatial position of each pixel in the target annular region according to the geometric transformation relationship. Corrected dial image refers to the image result after compensating for the geometric deformation caused by the tilt of the camera device's viewing angle, spatial position offset, or changes in projection relationship during the dial imaging process. In this image, the circular structure of the dial presents a geometric shape consistent with the actual physical structure on the imaging plane, so that the distribution of scale lines, pointer position, and spatial relationship between the center of the dial are consistent with the actual dial structure. This ensures that the angular relationship in the image corresponds to the angular relationship in the actual mechanical structure, providing an accurate geometric basis for subsequent calculation of the delamination amount based on the pointer direction.

[0082] An output image carrier is established using the pixel coordinate range of the optimized observation image as the data carrier for the correction dial image. Then, the pixel coordinate set of the target annular region is traversed, and the original pixel coordinates and corresponding grayscale values ​​of each target pixel in the optimized observation image are read one by one. The original pixel coordinates are substituted into geometric transformation relationships to calculate the transformed pixel coordinates of the target pixel. Next, the pixel grayscale values ​​are written into the output image carrier according to the transformed pixel coordinates. When multiple original pixels are mapped to the same transformed pixel coordinates after coordinate transformation, the corresponding multiple pixel grayscale values ​​are averaged and written to the transformed pixel coordinate position. When a transformed pixel coordinate is not written by any original pixel, the pixel grayscale value corresponding to that transformed pixel coordinate position in the optimized observation image is read and written into the output image carrier to maintain image background continuity. Finally, after completing the coordinate transformation and grayscale writing of all target pixels, the output image carrier is determined as the correction dial image, so that the geometric shape of the target annular region in the correction dial image is corrected according to geometric transformation relationships, while maintaining the imaging content of other areas of the dial consistent with the original optimized observation image.

[0083] When a ring of condensed water vapor appears on the dial surface, this thin layer of water acts like a transparent curved glass, altering the light path. This causes the images of the scale lines and pointer to be distorted and ghosted at the points where the water band passes. This manifests as local lines appearing thicker, more curved, with blurred edges, or pulled off-center. Consequently, the circumferential relationship of the dial in the image no longer matches the actual dial, and the angle of the pointer relative to the zero mark is locally distorted in the image. If the angle calculation is performed directly without first obtaining a corrected image of the dial, this visual offset caused by the water band may be mistaken for the actual pointer rotation, leading to errors in the calculation of the separation amount.

[0084] It should be noted that the water vapor condensation band refers to the strip-shaped liquid water layer formed when water vapor in the air condenses on the surface of the dial or the inside of its transparent protective component under the influence of temperature reduction or surface temperature difference. This liquid water layer is continuously or semi-continuously distributed along the circumference of the dial and extends around the center of the dial. Its formation is related to the printing ink of the dial, the coating boundary, the difference in surface roughness, and the edge wetting effect, which makes it easier for water droplets or water films to accumulate and remain at a certain approximately fixed radius. As a transparent medium, the water vapor condensation band will change the light propagation path through its area, causing brightness changes, edge blurring, positional shift, or ghosting phenomena in the imaging of the scale lines and pointers, thereby affecting the readability and geometric consistency of the dial image.

[0085] The operation management module manages the roof risk status of coal mine roadway sections based on the delamination amount corresponding to the calibration dial image.

[0086] In embodiments of the present invention, operational management of the roof risk status of coal mine roadway sections includes: The outer contour of the correction dial image is detected to obtain the center pixel coordinates of the dial. Outer contour detection refers to the process of identifying the pixel distribution of the outermost circular boundary of the dial in the image; the center pixel coordinates refer to the geometric center position of the circular boundary of the dial in the image coordinate system.

[0087] The image of the calibrated dial is acquired and its pixel-by-pixel grayscale is read. The boundary position between the dial and the background in terms of brightness or texture is identified. A set of outer contour pixels of the dial is formed along the boundary position. The connectivity of the outer contour pixel set is checked to remove discrete noise points and retain continuous closed contour curves. Then, a circular fitting process is performed on the outer contour pixel set to minimize the overall deviation between the fitted circle and the outer contour pixel set. The position of the circle center in the image coordinate system is read from the fitted circle result as the circle center pixel coordinates.

[0088] The scale markings of the calibration dial image are identified to obtain the 0-scale direction vector; Gradient recognition refers to the process of identifying the scale lines and numerical marks on the surface of a dial to determine their spatial orientation; the zero-gradient direction vector refers to the direction from the center pixel coordinates to the zero-gradient mark position on the dial.

[0089] The coordinates of the center pixel are obtained and the circumference is traversed around the center in the correction dial image. The grayscale changes of the areas where the scale lines and scale numbers are located are read along the circumference to determine the candidate positions of the scale markers. Shape and connectivity analysis is performed on the candidate positions to distinguish between the scale line areas and the number areas. The area containing the zero scale number is located and the position of the zero scale line adjacent to the number is determined. The representative pixel position of the zero scale line in the image is selected as the zero scale reference point. Then, a direction representation is constructed from the center pixel coordinates to the zero scale reference point, and this direction representation is determined as the zero scale direction vector so as to determine the angle between it and the pointer direction vector and to calculate the delamination amount.

[0090] Extract the pointer skeleton line from the corrected dial image; The pointer skeleton line refers to the set of center line pixels obtained after refining the pointer shape in the dial image.

[0091] A calibration dial image is acquired, and the effective area of ​​the dial is defined with the center pixel coordinates as a reference. Within the effective area, the grayscale values ​​of each pixel are read, and the elongated pointer region is identified. The identification process is accomplished by comparing the differences in grayscale and connectivity between the pointer region and the dial background. After obtaining the set of pixel coordinates of the pointer region, a connectivity check is performed on the set of pixel coordinates, and discrete pixels that are not connected to the pointer region are removed. Then, a layer-by-layer shrinkage process is performed within the remaining pointer region, so that the width of the pointer region gradually converges to the width of a single pixel while maintaining the topological connectivity of the pointer, until a single-pixel connected center line that can run through the length direction of the pointer is obtained. The set of pixel coordinates corresponding to the single-pixel connected center line is determined as the pointer skeleton line.

[0092] Endpoint detection is performed on the pointer skeleton lines to obtain the pixel coordinates of the pointer tip; The pointer tip pixel coordinates refer to the pixel position of the pointer skeleton line at the end furthest from the center; the pointer direction vector refers to the direction formed by the center pixel coordinates pointing to the pointer tip pixel coordinates.

[0093] Obtain all pixel coordinates of the pointer skeleton line and count the neighborhood connectivity of each pixel coordinate. Determine the pixel coordinates of the pointer skeleton line that are connected to only one adjacent skeleton pixel as endpoint candidate pixels, and gather all endpoint candidate pixels to form an endpoint candidate set. Then obtain the center pixel coordinates, calculate the distance from each endpoint candidate pixel in the endpoint candidate set to the center pixel coordinates, and determine the endpoint candidate pixels with larger distances as the pointer tip pixel coordinates at the end furthest from the center.

[0094] The pointer direction vector is determined based on the center pixel coordinates and the pointer tip pixel coordinates; The pointer direction vector refers to the spatial direction formed in the dial image, starting from the center pixel coordinates of the dial and ending at the pointer tip pixel coordinates. It reflects the pointing relationship of the pointer relative to the center of the dial in the current imaging plane. This direction is used to describe the degree of angular offset of the pointer relative to the zero mark position and serves as an important geometric basis for converting the pointer position in the image into the actual delamination amount.

[0095] Using the center pixel coordinates as the starting point of the direction vector and the pointer tip pixel coordinates as the ending point of the direction vector, the coordinate difference between the ending point coordinates and the starting point coordinates in the image coordinate system is calculated to obtain the direction component from the center of the circle to the pointer tip, and the direction component is recorded as the pointer direction vector.

[0096] Determine the angle between the 0-point direction vector and the pointer direction vector; The included angle refers to the angular difference between the zero-scale direction vector and the pointer direction vector.

[0097] Obtain the zero-scale direction vector and the pointer direction vector, calculate the length of the two direction vectors respectively, and convert the two direction vectors into a direction representation of unit length. Then, calculate the dot product of the two unit direction representations based on the correspondence between the dot product of the vectors and the included angle, and combine the dot product result with the length relationship of the two unit direction representations to obtain the included angle between the two directions.

[0098] The amount of delamination is determined based on the included angle and the calibration constant of the top plate delamination instrument; The calibration constant refers to the proportional relationship parameter between the angle change and the actual delamination displacement in the mechanical structure of the roof delamination instrument; the delamination amount refers to the amount of separation displacement generated between the roadway roof and the surrounding rock.

[0099] The calibration constant corresponding to the top plate delamination instrument is read from the equipment information of the top plate delamination instrument. The calibration constant is used to characterize the proportional relationship between the pointer angle change and the delamination displacement. Then, the included angle value is multiplied by the calibration constant to obtain the delamination amount corresponding to the included angle.

[0100] The section of the coal mine roadway where the roof separation instrument is located; A coal mine roadway section refers to a specific roadway segment in the underground mining space of a coal mine, divided according to its spatial location.

[0101] The equipment identification information of the roof separation instrument and the spatial coordinate data bound to it are obtained. The spatial coordinate data is expressed using the unified coordinate reference system used by the coal mine geographic information system (GIS). Subsequently, the roadway spatial layer data is retrieved from the GIS spatial database. This roadway spatial layer data includes roadway centerline element layers, roadway boundary element layers, and roadway segment division element layers. Spatial overlay analysis is performed in the GIS platform, spatially matching the spatial coordinate data with the roadway centerline elements, calculating the spatial distance from the spatial coordinates to each roadway centerline element, and identifying elements with smaller distances located within the roadway boundary range. The centerline element within the enclosure is used as the corresponding roadway. Then, linear positioning calculations are performed in the GIS based on the centerline of the corresponding roadway to determine the mileage position of the spatial coordinates in the roadway direction. Based on the start and end mileage information recorded in the roadway segment division element layer, the specific roadway segment containing the mileage position is determined, and this roadway segment is identified as the coal mine roadway segment where the roof separation meter is located. Finally, in the attribute table of the GIS, the attribute association relationship between the coal mine roadway segment element and the separation amount value is established, so that the separation amount is attached as attribute data to the corresponding roadway segment element for spatial visualization in the GIS risk zoning map and subsequent operation and management processing of roof risk status.

[0102] Based on the amount of delamination, the risk status of the roof in coal mine roadway sections is managed through operation.

[0103] Roof risk status refers to the safety level of the roadway roof, which reflects the stability of the roadway roof based on the amount of delamination. The roadway roof refers to the surrounding rock structure above the roadway after the excavation of the underground roadway in a coal mine. It consists of rock strata or coal-rock composite layers above the top boundary of the coal seam. After the roadway is formed, this part of the rock mass loses its original stress balance and bears the redistribution of ground pressure, making it a key stress-bearing part for the stability of the roadway structure. Operation and management refers to the management behavior of scheduling, supporting and maintaining, or issuing early warnings for coal mine roadway sections based on the roof risk status.

[0104] Information on mine geological conditions, support structures, mining depth, historical delamination volume, and historical treatment results associated with the coal mine roadway section was retrieved. A time-series record of the delamination volume for this roadway section was established under the same coordinate reference system. Subsequently, using the historical time-series record of the delamination volume as the operating baseline for this roadway section, the deviation of the current delamination volume from this operating baseline and the rate of change per unit time were calculated. Simultaneously, the fluctuation amplitude and persistence of the delamination volume sequence were statistically analyzed to distinguish between short-term disturbances and continuous evolution. Then, the deviation, rate of change, and fluctuation persistence were compared with the geological conditions and support conditions of the roadway section to obtain the correlation with the... The stability assessment results are matched with the roadway section conditions, and the roof risk status is generated accordingly. When the delamination amount remains near the historical baseline and the rate of change is stable, it is judged as normal. When the delamination amount is continuously rising relative to the baseline and the rate of change shows that roof separation is developing, it is judged as a warning. When the delamination amount deviates significantly from the baseline and the rate and persistence of change indicate that roof separation is intensifying and accompanied by a rapid evolution trend, it is judged as dangerous. Finally, the assessment results, along with the current delamination amount, rate of change, and summary of assessment basis, are written into the roadway section spatial element attributes of the geographic information system (GIS) to realize the dynamic updating of the roof risk status of coal mine roadway sections.

[0105] The steps for implementing operational management of the roof risk status of coal mine roadway sections include: First, acquiring spatial elements of the target coal mine roadway section in a Geographic Information System (GIS) and reading the roof risk status level of these spatial elements; simultaneously retrieving records of delamination volume, risk status change records, and personnel and equipment distribution information associated with the roadway section; then, generating an operational management instruction set corresponding to the roadway section based on the roof risk status level. This instruction set includes instructions for adjusting roadway section operations, adjusting inspection frequency, arranging support and maintenance work, and scheduling on-site verification. The operational management instruction set is then written into the mine production scheduling data to form tasks to be executed; finally, the tasks to be executed are... The task is linked with the spatial elements of the roadway section in the GIS interface to ensure that the spatial location of the task is consistent with the scope of the roadway section, and the task content is issued to the on-site execution terminal to complete the task assignment. Then, during the task execution, the execution status information and re-measured delamination data are continuously received from the field. Based on the returned data, the delamination record of the roadway section is updated and the latest value of the roof risk status level is updated. At the same time, the task execution result is written into the risk treatment record to form a traceable closed loop. Finally, based on the updated roof risk status level, the risk zoning display of the roadway section is refreshed in the GIS, and the current allowed operation status, inspection status and support status of the roadway section are synchronized to the mine dispatch management interface, thereby completing the dynamic operation management of the roof risk status of the coal mine roadway section.

[0106] To address the aforementioned problems, this invention also provides a GIS-based intelligent operation and management method for coal mines, see [link to relevant documentation]. Figure 1 Specifically, including: An initial ring-enhanced image is generated based on the original image of the dial; The interference index is obtained from the initial annular region based on the initial annular enhancement image, and optimized observation images are acquired according to the reinforcement learning method and the interference index. The annular search region of the optimized observation image is determined, and the pixel marking of the optimized observation image is performed based on the set of annular contour points corresponding to the annular search region to obtain the annular mask image. The geometric transformation relationship is determined based on the annular mask image, and the coordinate transformation is performed on the pixels of the target annular region of the optimized observation image to obtain the correction dial image; Based on the delamination amount corresponding to the calibration dial image, the roof risk status of the coal mine roadway section is managed operationally.

[0107] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A GIS-based intelligent operation and management system for coal mines, characterized in that, include: The ring reinforcement module generates an initial ring reinforcement image based on the original image of the dial; The observation optimization module obtains the interference index based on the initial annular region of the initial annular enhanced image, and acquires optimized observation images according to the reinforcement learning method and the interference index. The annular mask module determines the annular search region of the optimized observation image and, based on the set of annular contour points corresponding to the annular search region, performs pixel marking on the optimized observation image to obtain the annular mask image. The ring correction module determines the geometric transformation relationship based on the ring mask image and performs coordinate transformation on the pixels of the target ring region in the optimized observation image to obtain the correction dial image. The operation management module manages the roof risk status of coal mine roadway sections based on the delamination amount corresponding to the calibration dial image.

2. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, The specific steps for generating the initial annular enhancement image are as follows: A camera device is installed in front of the scale of the top plate separation instrument; Set the observation trajectory according to the center of the dial; Select multiple observation postures on the observation trajectory; The original image of the dial is captured using a camera device at each observation posture; Determine the maximum and minimum grayscale values ​​of each pixel coordinate in the original image; The grayscale range of pixel coordinates is determined based on the maximum and minimum grayscale values. The grayscale range is normalized to obtain the initial annular enhanced image.

3. The GIS-based intelligent operation and management system for coal mines according to claim 2, characterized in that, The specific steps to obtain the interference index are as follows: Image segmentation is performed on the initial annular enhancement image to obtain the initial annular region; At each observation posture, the observation image of the dial is acquired by a camera device; Identify the set of target pixels that overlap with the initial annular region in the observed image; Determine the first grayscale average value of the target pixel set; Based on the target pixel set, determine the remaining pixel set of the initial annular region; Determine the second grayscale average value of the remaining pixel set; The interference index is obtained based on the first grayscale average value and the second grayscale average value.

4. The GIS-based intelligent operation and management system for coal mines according to claim 2, characterized in that, The specific steps for acquiring and optimizing observation images are as follows: Based on the reinforcement learning method, the observation attitude is iteratively adjusted to obtain the target observation attitude; where the target observation attitude refers to the observation attitude corresponding to the smallest disturbance index. Under the target observation posture, optimized observation images of the dial are acquired through a camera device.

5. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, The specific steps for determining the circular search region for optimizing the observed image are as follows: Determine the grayscale distribution of the initial annular enhancement image; Determine the initial center coordinates, initial inner radius, and initial outer radius of the initial annular region based on the grayscale distribution. Based on the initial center coordinates, initial inner radius, and initial outer radius, the annular search area for optimizing the observed image is determined.

6. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, The specific steps to obtain the annular mask image are as follows: Edge detection is performed on the annular search region to obtain a set of annular contour points; based on the set of annular contour points, pixel marking is performed on the optimized observation image to obtain an annular mask image.

7. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, Determining geometric transformation relationships based on the annular mask image includes: Based on the annular mask image, contour points are extracted from the optimized observation image to obtain the inner edge contour point set; An ellipse is fitted to the set of inner edge contour points to obtain the coordinates of the ellipse center, the length of the major axis, and the length of the minor axis. Based on the coordinates of the ellipse's center, the length of its major axis, and the length of its minor axis, determine the geometric transformation relationships.

8. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, The specific steps to obtain the correction dial image are as follows: Based on the annular mask image, determine the target annular region for optimizing the observation image; Based on the geometric transformation relationship, the coordinates of the pixels in the target annular region are transformed to obtain the correction dial image.

9. The GIS-based intelligent operation and management system for coal mines according to claim 1, characterized in that, Operational management of the roof risk status of coal mine roadway sections includes: The outer contour of the correction dial image is detected to obtain the center pixel coordinates of the dial. The scale markings of the calibration dial image are identified to obtain the 0-scale direction vector; Extract the pointer skeleton line from the corrected dial image; Endpoint detection is performed on the pointer skeleton lines to obtain the pixel coordinates of the pointer tip; The pointer direction vector is determined based on the center pixel coordinates and the pointer tip pixel coordinates; Determine the angle between the 0-point direction vector and the pointer direction vector; The amount of delamination is determined based on the included angle and the calibration constant of the top plate delamination instrument; The section of the coal mine roadway where the roof separation instrument is located; Based on the amount of delamination, the risk status of the roof in coal mine roadway sections is managed through operation.

10. A GIS-based intelligent operation and management method for coal mines, characterized in that: The method includes: An initial ring-enhanced image is generated based on the original image of the dial; The interference index is obtained from the initial annular region based on the initial annular enhancement image, and optimized observation images are acquired according to the reinforcement learning method and the interference index. The annular search region of the optimized observation image is determined, and the pixel marking of the optimized observation image is performed based on the set of annular contour points corresponding to the annular search region to obtain the annular mask image. The geometric transformation relationship is determined based on the annular mask image, and the coordinate transformation is performed on the pixels of the target annular region of the optimized observation image to obtain the correction dial image; Based on the delamination amount corresponding to the calibration dial image, the roof risk status of the coal mine roadway section is managed operationally.

Images