A visual processing method of a mine safety monitoring system

Abstract

The application provides a kind of mine safety monitoring system visual processing method, method includes: visual processing terminal receives the three-dimensional scene data that mine safety monitoring system sends, extracts, washes and standardization pretreatment to three-dimensional scene data, obtains the three-dimensional scene data after pretreatment;Secondly, the three-dimensional scene data is rendered by a variety of rendering techniques, and the three-dimensional scene data after rendering is obtained, and the rendering technique priority is set according to the different target features;Finally, the three-dimensional scene data after rendering is visualized by display terminal, and the risk early warning and scientific decision of mine safety monitoring are carried out.The application renders the three-dimensional scene data by a variety of rendering techniques, realizes the purpose of building intelligent safety supervision and management application, and realizes the monitoring, evaluation, early warning and trend analysis of the safety production risk of high-risk enterprises such as mine production by using big data and artificial intelligence and other technical means.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology for mine monitoring, and in particular to a visualization processing method for a mine safety monitoring system. Background Technology

[0002] Due to the complex environment, numerous equipment and personnel involved in mining production, high-standard safety monitoring is essential. Current safety monitoring systems utilize digital twin-based industrial internet platforms, enabling various interactive operations through event binding. These platforms provide visualization functions such as camera view control, point, line, and surface effects, interface data display, model attribute settings, and data binding. Furthermore, leveraging data visualization and digital twin technologies, they recreate the macro-distribution and micro-scenes of key mining enterprises on a computer, integrating various data to visually present the "twin" business scenarios of these key enterprises. However, mining safety monitoring systems involve a vast amount of high-precision terrain, oblique photography, and all building scene data, resulting in a large number and wide range of 3D scenes. Applications require the system to load and display an extremely large number of 3D models and scenes. Traditional software methods to improve rendering efficiency primarily focus on increasing the efficiency of a single rendering pass. This is achieved by reducing the number of triangles in various models within the scene, thereby improving rendering efficiency. Alternatively, reducing texture size and pixel size can reduce rendering time per pass. However, both of these methods significantly degrade image quality and are typically used for less important scene areas and models.

[0003] Existing technology one, application number: CN201511011484.0, discloses a networked integrated supervision platform for mine safety monitoring based on the Internet of Things. It includes wireless sensors for monitoring the mine environment, positioning devices for determining and tracking the location of personnel or equipment, wireless sensors and positioning devices connected to a router, the router connected to a data transmission device via a centralized coordinator, the data transmission device connected to a control platform via a communication network, and the control platform connected to a database for storage. It also includes video surveillance equipment connected to the control platform via a communication network. Although it is an integrated software platform for the "six major systems" of underground mines, and a networked integrated supervision platform for mine safety monitoring, it lacks the visualization processing of big data for mine safety monitoring, and cannot accurately reproduce mine scene data. It is currently one of the few platforms in China that has achieved the integration of major mine systems, multi-level risk, and comprehensive supervision.

[0004] The prior art two, application number: CN202110927631.8 discloses a mine safety hazard identification and early warning system based on big data analysis, comprising: a big data acquisition module for collecting safety hazard data of different sources and different formats; a big data standardization preprocessing module for extracting, cleaning and standardizing preprocessing the collected safety hazard data; a big data analysis module for analyzing the preprocessed data based on a data mining model library, and structurally storing the analysis results; a big data visualization module for visualizing the big data analysis results, facilitating risk early warning and scientific decision-making by safety managers. Although a wisdom safety analysis and early warning system is constructed with intelligent analysis, identification, mining and knowledge discovery as the core content, which can provide support for mine safety risk identification, hazard source classification, safety control strategy and safety emergency disposal; but the big data analysis results are not rendered, which affects the effect of the picture.

[0005] The prior art three, application number: CN202211165591.9 discloses a mine safety monitoring system mass video data adaptive streaming method based on mobile edge computing, comprising a center server, an edge server and a user device; first, a mobile edge framework of the mobile mine safety monitoring system client adaptive streaming is constructed, then the information of the client is collected through the edge server and input into the neural network model to determine the bit rate of the requested mine safety monitoring system video segment. The neural network model is trained by using the reinforcement learning algorithm, and the mine safety monitoring system mass video bit rate adaptive to the client is calculated and then distributed to the client; although the mobile edge computing and the mine safety monitoring system video adaptive streaming are combined, the overall QoE of the client and the QoE fairness of the client are improved, and the mine edge server plays an important role in video data monitoring and visualization to the client. However, the rendering process of the video data is lacking, which reduces the texture size and pixel size of the three-dimensional model and scene.

[0006] At present, the prior art one, the prior art two and the prior art three have low rendering efficiency of mine safety monitoring data, which reduces the texture size and pixel size and greatly reduces the picture effect. Therefore, the present application provides a visualization processing method of a mine safety monitoring system, which uses multiple methods to improve the rendering efficiency to optimize the engine of the mine safety monitoring system, and finally achieves the purpose of high rendering efficiency and fast speed of the mine scene. SUMMARY

[0007] In order to solve the above technical problems, the present application provides a visualization processing method of a mine safety monitoring system, comprising the following steps:

[0008] The visual processing terminal receives three-dimensional scene data sent by the mine safety monitoring system, extracts, cleans and standardizes the three-dimensional scene data for preprocessing, and obtains the preprocessed three-dimensional scene data;

[0009] The three-dimensional scene data is rendered by multiple rendering techniques, and the rendered three-dimensional scene data is obtained. The rendering techniques are prioritized according to different target features. The rendering techniques include centralized rendering, visibility culling rendering, multi-resolution model simplification rendering, texture optimization rendering, image drawing rendering and acceleration, and programmable graphics card rendering acceleration.

[0010] The display terminal visualizes the rendered three-dimensional scene data, and performs risk warning and scientific decision-making for mine safety monitoring.

[0011] Optionally, the process of extracting, cleaning and standardizing the three-dimensional scene data includes the following steps:

[0012] An instruction for receiving three-dimensional scene data is obtained. Based on the instruction, the three-dimensional scene data is input into a data extraction network to obtain the name of the target three-dimensional scene data and the corresponding parameter value, forming a target three-dimensional scene data set. The name of the target three-dimensional scene data includes the name of one or more three-dimensional scene data with the same meaning.

[0013] Check if there is duplicate data in the target three-dimensional scene data set. If there is, delete the duplicate data. Check if there is an abnormal value in the data set. If there is, delete the abnormal value. Obtain the target three-dimensional scene data set after deleting the duplicate data and the abnormal value. The abnormal value is caused by equipment failure, manual input error or abnormal event.

[0014] The target three-dimensional scene data set after deleting the duplicate data and the abnormal value is converted into a format recognizable by the visual processing terminal, and is grouped according to different types of three-dimensional scene data to obtain the preprocessed three-dimensional scene data.

[0015] Optionally, the process of setting the rendering technique priority according to the different target features includes the following steps:

[0016] According to the purpose of the multiple rendering techniques, the corresponding target features are defined.

[0017] The priority of the multiple rendering techniques is defined according to the different target features. The content is that the programmable graphics card rendering acceleration is the first priority, the image drawing rendering is the second priority, the centralized rendering is the third priority, the visibility culling rendering is the fourth priority, the multi-resolution model simplification rendering is the fifth priority, and the texture optimization rendering is the sixth priority.

[0018] Receiving three-dimensional scene data to be rendered, sorting according to priority, calling rendering technology through priority mobilization program, respectively rendering target features, when one rendering technology ends, automatically entering the next priority rendering technology;

[0019] When the three-dimensional scene data to be rendered is completely rendered, the visualization processing automatically sends a stop rendering instruction to the programmable rendering acceleration of the first priority graphics card, at this time, the corresponding programs of other priorities also stop working;

[0020] The rendered three-dimensional scene data is delivered to the display terminal for visual presentation processing.

[0021] Optionally, the target features of centralized rendering are: triangle vertex format, whether the vertex format is the same, whether the number of textures used is the same, whether each texture is the same, whether lighting is used, whether the material used is the same, and whether the drawing method is the same; the target features of visibility culling rendering are: triangle vertex transformation, removing the scene part that does not contribute to the final three-dimensional scene data rendering before the three-dimensional scene model coordinate conversion stage, and sending the remaining scene to the drawing pipeline; the target features of multi-resolution model simplification rendering are: simplifying small and unimportant parts and remote scenes in the three-dimensional scene model; the target features of texture optimization rendering are: cutting textures, using methods such as compressing multiple images into a single texture or eliminating small textures, and using high-resolution images for objects close to the observer, and using low-resolution images for objects gradually moving away from the observer; the target features of image drawing rendering and acceleration are: drawing a virtual scene, using discrete images collected by a camera or continuous videos collected by a camera as basic data, generating panoramic images through image processing and establishing a virtual environment with spatial manipulation ability based on spatial correlation; the target features of programmable rendering acceleration of the graphics card are: using triangle strip and triangle fan geometric representation methods to accelerate three-dimensional scene rendering.

[0022] Optionally, the processing process of image drawing rendering in the three-dimensional scene data includes the following steps:

[0023] Obtaining three-dimensional scene data to be rendered, under the support of programmable rendering acceleration of the first priority graphics card, using image drawing rendering for the three-dimensional scene data to be rendered;

[0024] Calling stored three-dimensional scene data from the visualization processing terminal to draw a virtual scene instead of all geometric information of the three-dimensional scene data to be rendered;

[0025] The discrete images captured by a camera or the continuous video captured by a video camera are used as basic data, and a panorama image is generated through image processing, and a virtual environment with spatial manipulation ability is established through spatial correlation, and the three-dimensional scene data to be rendered is rendered to obtain three-dimensional scene data after image rendering.

[0026] Optionally, the centralized rendering process in the three-dimensional scene data comprises the following steps:

[0027] The three-dimensional scene data after image rendering is received, and the three-dimensional scene data after image rendering is rendered using triangles under the support of programmable rendering acceleration of the first priority graphics card.

[0028] The triangles are classified according to the following standards: vertex format, number of textures, texture quality, light intensity, material, and rendering method.

[0029] The quality level of the three-dimensional scene data to be rendered is determined based on the resources of the graphics processor, and at least one triangle and at least two sub-three-dimensional scene data corresponding to the three-dimensional scene data to be rendered are generated according to the quality level, the triangles are classified according to the classification standard, and the same type of triangles are combined to obtain the combined same type of triangles.

[0030] The at least two sub-three-dimensional scene data to be rendered and the at least one combined same type of triangle are rendered to obtain the three-dimensional scene data after centralized rendering.

[0031] Optionally, the processing process of the visibility culling rendering on the three-dimensional scene data after centralized rendering comprises the following steps:

[0032] Each scene object model in the three-dimensional scene data after centralized rendering is obtained, and the part of the three-dimensional scene data without contribution value is culled to obtain the occlusion information of the scene object model after culling.

[0033] According to the occlusion information, the occlusion type of the scene object model after culling processing is determined, and the target scene object model is obtained based on the occlusion type and the scene object model after culling processing.

[0034] The target scene object model is rendered to obtain the three-dimensional scene data after visibility culling rendering.

[0035] Optionally, the contribution value is based on a feature correlation degree of the object model in the three-dimensional scene data and the background, the feature correlation network is used to calculate the correlation features of the object model features and the background features to obtain the correlation features corresponding to the object model features and the background features, the classification network is used to analyze the correlation features to determine the feature correlation degree corresponding to the object model and the background in the three-dimensional scene data, and then compared with the preset feature correlation degree threshold, if the preset feature correlation degree threshold is not reached, it is considered that there is no contribution value.

[0036] Optionally, the multi-resolution model simplification rendering process for the three-dimensional scene data after the visibility culling rendering includes the following steps:

[0037] The three-dimensional scene data after the visibility culling rendering is obtained, the objects in the three-dimensional scene data are determined, and the visible area and the visible polygons are obtained according to the shooting position of the camera;

[0038] The initial error control threshold of the point-to-point method is used to simulate the inaccurate critical state when the position and shape of the object are simplified;

[0039] The distance between the positions and shapes corresponding to the points is calculated, the corrected error control threshold is calculated through the initial error control threshold, and the region in the cuboid view volume is simplified according to the corrected error control threshold.

[0040] The rendering list is sent to the renderer for rendering to obtain the three-dimensional scene data after the multi-resolution model simplification rendering.

[0041] Optionally, the texture optimization rendering process for the three-dimensional scene data after the multi-resolution model simplification rendering includes the following steps:

[0042] The density clustering algorithm is used to cluster the textures in the fine three-dimensional scene data to obtain a first clustering result, and multiple three-dimensional scene data are compressed into a single texture;

[0043] Each point in the single texture is judged, if the current point is an unclassified point, the current point is assigned a classification number, and it is judged whether the remaining points belong to the same triangular face as the current point, if they belong to the same triangular face, regardless of the classification number of the remaining points in which class cluster of the first clustering result, the classification number of the remaining points is modified to the classification number of the current point, and a second clustering result is obtained.

[0044] The minimum circumscribed rectangle of each class cluster in the second clustering result is calculated to determine the effective texture.

[0045] The first visualization processing terminal of the application receives three-dimensional scene data sent by a mine safety monitoring system, extracts, cleans and standardizes the three-dimensional scene data for preprocessing, and obtains preprocessed three-dimensional scene data; secondly, the three-dimensional scene data is rendered by various rendering techniques, and the rendered three-dimensional scene data is obtained, and the rendering technique priority is set according to the different target features; the rendering techniques include centralized rendering, visibility culling rendering, multi-resolution model simplification rendering, texture optimization rendering, image rendering and acceleration, and programmable graphics card rendering acceleration; finally, the rendering terminal visualizes the rendered three-dimensional scene data, and carries out risk early warning and scientific decision-making of mine safety monitoring; the above scheme renders the three-dimensional scene data by various rendering techniques, sets the rendering technique priority according to the different target features, first meets the requirements of large and wide range of three-dimensional scene data in mine safety production site, realizes the efficiency of three-dimensional scene and model system loading and display, greatly reduces the number of display cards called by each frame of picture through centralized rendering, avoids unnecessary rendering resources and rendering state switching, and finally realizes the purpose of improving display efficiency, which can improve 30% to 50% compared with the prior art; the remaining three-dimensional scene is sent to the rendering pipeline through the visibility culling rendering, which can effectively reduce the complexity of the three-dimensional scene and the burden of the graphics pipeline, and effectively improve the efficiency of three-dimensional scene rendering; the multi-resolution model simplification rendering ensures that the content of the three-dimensional scene will not be severely distorted, improves the rendering speed of the entire three-dimensional scene, and maximizes the reduction of the number of triangles and vertices of the original three-dimensional model while maintaining the original three-dimensional model features as much as possible; the texture optimization rendering effectively improves the use efficiency of texture memory and the quality of three-dimensional scene rendering; through the image rendering and acceleration, the complex problem of traditional technology in the process of establishing complex space model is solved, and the efficiency of establishing panoramic map model is improved; through the programmable graphics card rendering acceleration technology, a large amount of graphics processing unit (GPU) transformation and illumination calculation time is saved, and the bandwidth occupation between central processing unit (CPU) and image hardware is reduced.

[0046] The embodiment renders three-dimensional scene data through various rendering technologies, obtains three-dimensional scene data after rendering processing, sets rendering technology priority according to different target characteristics, and realizes the purpose of constructing a smart safety supervision and management application. As the support of risk monitoring and sensing data of high-risk industry enterprises such as mine production, the application uses big data and artificial intelligence and other technical means to realize the monitoring, evaluation, early warning and trend analysis of the safety production risk of high-risk enterprises such as mine production, perfects the new mechanism of safety supervision driven by big data, promotes the implementation of the safety production subject responsibility of enterprises, promotes the deep integration of industrial internet and emergency management business, realizes the standardized and standardized safety risk grading control ability, provides support for key supervision, strict law enforcement and scientific policy making, and effectively curbs major accidents. The embodiment realizes the modernization of emergency management production by information technology, promotes the formation of a complete, clear and advanced emergency management information system, improves the monitoring and early warning, supervision and law enforcement, auxiliary command decision, rescue operation and social mobilization ability, and finally realizes the key technology research and development and engineering application of mine safety production, mine safety supervision and inspection or emergency rescue.

[0047] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the application. The objects and other advantages of the present application can be achieved and obtained by the structure particularly pointed out in the written description, claims, and drawings.

[0048] The technical solutions of the present application will be further described in detail below by means of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0049] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. In the drawings:

[0050] Figure 1 The flowchart of the visualization processing method of the mine safety monitoring system in embodiment 1 of the present application;

[0051] Figure 2 The process diagram of extracting, cleaning and standardizing preprocessing of three-dimensional scene data in embodiment 2 of the present application;

[0052] Figure 3 The process diagram of setting rendering technology priority according to different target characteristics in embodiment 3 of the present application;

[0053] Figure 4 The processing process diagram of image rendering in three-dimensional scene data in embodiment 4 of the present application;

[0054] Figure 5 This is a diagram illustrating the centralized rendering process of the three-dimensional scene data in Embodiment 5 of the present invention;

[0055] Figure 6 This is a diagram illustrating the process of visibility culling rendering on the centrally rendered 3D scene data in Embodiment 6 of the present invention.

[0056] Figure 7 This is a diagram illustrating the process of processing the 3D scene data after visibility culling rendering in the simplified rendering of the multi-resolution model in Embodiment 7 of the present invention.

[0057] Figure 8 This is a diagram illustrating the process of texture optimization rendering of simplified rendering of 3D scene data after multi-resolution model rendering in Embodiment 8 of the present invention. Detailed Implementation

[0058] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0059] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0060] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0061] Example 1: As Figure 1 As shown in the figure, this embodiment of the invention provides a visualization processing method for a mine safety monitoring system, including the following steps:

[0062] S100: The visualization processing terminal receives the three-dimensional scene data sent by the mine safety monitoring system, extracts, cleans and standardizes the three-dimensional scene data for preprocessing, and obtains the preprocessed three-dimensional scene data;

[0063] S200: The three-dimensional scene data is rendered by multiple rendering techniques to obtain the rendered three-dimensional scene data, and the rendering technique priority is set according to the different target characteristics; the rendering techniques include centralized rendering, visibility culling rendering, multi-resolution model simplification rendering, texture optimization rendering, image drawing rendering and acceleration, and programmable drawing acceleration of the graphics card;

[0064] S300: The display terminal visually presents the rendered three-dimensional scene data, and performs risk early warning and scientific decision-making for the mine safety monitoring;

[0065] The working principle and beneficial effects of the technical solution are as follows: firstly, the embodiment visualizes the three-dimensional scene data received by the terminal from the mine safety monitoring system, extracts, cleans and standardizes the three-dimensional scene data for preprocessing, and obtains the preprocessed three-dimensional scene data; secondly, the three-dimensional scene data is rendered and processed through various rendering technologies, and the three-dimensional scene data after rendering and processing is obtained, and the rendering technology priority is set according to different target features; the rendering technology includes centralized rendering, visibility culling rendering, multi-resolution model simplification rendering, texture optimization rendering, image rendering and acceleration, and programmable graphics card rendering acceleration; finally, the terminal displays the three-dimensional scene data after rendering and processing, and visualizes the three-dimensional scene data, and carries out risk early warning and scientific decision-making of mine safety monitoring; the above scheme renders and processes the three-dimensional scene data through various rendering technologies, sets the rendering technology priority according to the different target features, firstly meets the requirements of large and wide range of three-dimensional scene data of mine safety production site, realizes the efficiency of three-dimensional scene and model system loading and display, greatly reduces the number of display cards called for each frame of picture through centralized rendering, avoids unnecessary rendering resources and rendering state switching, and finally realizes the purpose of improving display efficiency, which can improve 30% to 50% compared with the prior art; the remaining three-dimensional scene is sent to the drawing pipeline through the visibility culling rendering, which can effectively reduce the complexity of the three-dimensional scene and the burden of the graphics pipeline, and effectively improve the efficiency of three-dimensional scene drawing; the multi-resolution model simplification rendering ensures that the content of the three-dimensional scene will not be severely distorted, improves the rendering speed of the entire three-dimensional scene, and maximizes the reduction of the number of triangles and vertices of the original three-dimensional model while maintaining the characteristics of the original three-dimensional model as much as possible; the texture optimization rendering effectively improves the use efficiency of texture memory and the quality of three-dimensional scene rendering; through the drawing and acceleration of the image, the complex problem of the traditional technology in the process of establishing a complex space model is solved, and the establishment efficiency of the panoramic map model is improved; through the programmable graphics card rendering acceleration technology, a large amount of graphics processing unit (GPU) transformation and illumination calculation time is saved, and the bandwidth occupation between the central processing unit (CPU) and the image hardware is reduced.

[0066] The embodiment renders three-dimensional scene data by various rendering technologies, obtains rendered three-dimensional scene data, sets rendering technology priority according to different target characteristics, visualizes the rendered three-dimensional scene data, carries out risk early warning and scientific decision of mine safety monitoring, realizes the purpose of building a smart safety supervision and management application, supports risk monitoring and sensing data of high-risk industries such as mine production, uses big data and artificial intelligence and other technical means to realize monitoring, evaluation, early warning and trend analysis of safety production risks of high-risk enterprises such as mine production, perfects a new safety supervision mechanism driven by big data, promotes the implementation of the safety production subject responsibility of enterprises, promotes the deep integration of industrial internet and emergency management business, realizes the standardized and standardized safety risk grading control ability, provides support for key supervision, strict law enforcement and scientific policy-making, and effectively prevents major accidents. The embodiment realizes the use of information technology to promote the modernization of emergency management production, promotes the formation of a complete, clear and advanced emergency management information system, improves monitoring and early warning, supervision and law enforcement, auxiliary command decision-making, rescue combat and social mobilization capabilities, and finally realizes the research and development and engineering application of mine safety production, mine safety supervision and inspection or emergency rescue key technologies.

[0067] Embodiment 2: as shown in Figure 2 On the basis of embodiment 1, the process of extracting, cleaning and standardizing preprocessing of three-dimensional scene data provided by the embodiment of the application comprises the following steps:

[0068] S101: An instruction for receiving three-dimensional scene data is obtained, based on the instruction, the three-dimensional scene data is input into a data extraction network, the name of the target three-dimensional scene data and the corresponding parameter value are obtained, a target three-dimensional scene data set is formed, and the name of the target three-dimensional scene data includes the name of one or more three-dimensional scene data with the same meaning;

[0069] S102: It is checked whether there is repeated data in the target three-dimensional scene data set, if there is, the repeated data is deleted; it is checked whether there is an abnormal value in the data set, if there is, the abnormal value is deleted, and a target three-dimensional scene data set with repeated data and abnormal values deleted is obtained; the abnormal value is caused by equipment failure, manual input error or abnormal event;

[0070] S103: The target three-dimensional scene data set with repeated data and abnormal values deleted is converted into a format recognizable by a visualization processing terminal, and is grouped according to different types of three-dimensional scene data, and a preprocessed three-dimensional scene data is obtained;

[0071] The working principle and beneficial effects of the above technical solution are as follows: First, this embodiment obtains an instruction to receive 3D scene data. Based on the instruction, the 3D scene data is input into a data extraction network to obtain the name and corresponding parameter values ​​of the target 3D scene data, forming a target 3D scene dataset. The name of the target 3D scene data includes one or more names of 3D scene data with the same meaning. Second, it checks whether there is duplicate data in the target 3D scene dataset. If so, the duplicate data is deleted. It then checks whether there are outliers in the dataset. If so, the outliers are deleted, resulting in a target 3D scene dataset with duplicate data and outliers removed. Finally, the target 3D scene dataset with duplicate data and outliers removed is converted into a format recognizable by the visualization processing terminal and processed according to the 3D scene data... The data is grouped according to different types to obtain preprocessed 3D scene data. The above scheme extracts, cleans, and standardizes the 3D scene data. The extraction method obtains the name and corresponding parameter values ​​of the target 3D scene data, realizing the initial processing of 3D scene data and laying the data foundation for the visualization processing of 3D scene data. The extraction realizes the feature division of 3D scene data, which helps to improve the recognition rate. The cleaning method removes duplicate data and outliers in the 3D scene data. On the one hand, it reduces the pressure of visualization processing, and on the other hand, it moves the processing of abnormal data forward, which is an important guarantee for improving the efficiency of 3D scene data. The standardization method realizes the format conversion, improving the processing efficiency of the visualization processing terminal.

[0072] Example 3: As Figure 3 As shown, based on Example 1, the process of setting rendering technology priorities according to different target features provided in this embodiment of the invention includes the following steps:

[0073] S201: According to the purpose of the plurality of rendering techniques, the corresponding target features are defined, wherein the target features of the centralized rendering are: triangle vertex format, whether the vertex formats are the same, whether the number of used textures is the same, whether each texture is the same, whether illumination is used, whether the used materials are the same, whether the drawing modes are the same, etc.; the target features of the visibility culling rendering are: triangle vertex transformation, removing the scene part without contribution to the final three-dimensional scene data rendering before the conversion from the three-dimensional scene model coordinates, and sending the remaining scene to the drawing pipeline; the target features of the multi-resolution model simplification rendering are: simplifying the small and unimportant parts and the remote scene in the three-dimensional scene model; the target features of the texture optimization rendering are: cutting textures, using methods such as compressing multiple images into a single texture or eliminating small textures, and using a high-resolution image for objects close to the observer, and using a low-resolution image for objects gradually far away from the observer; the target features of the image drawing rendering and acceleration are: drawing a virtual scene, using discrete images collected by a camera or continuous videos collected by a camera as basic data, generating panoramic images through image processing, and establishing a virtual environment with spatial manipulation ability through spatial correlation; and the target features of the GPU programmable drawing acceleration are: using triangle strip and triangle fan geometric representation methods to accelerate three-dimensional scene rendering.

[0074] S202: The priority of the plurality of rendering techniques is defined according to the differences in the target features, and the content is that the GPU programmable drawing acceleration is the first priority, the image drawing rendering is the second priority, the centralized rendering is the third priority, the visibility culling rendering is the fourth priority, the multi-resolution model simplification rendering is the fifth priority, and the texture optimization rendering is the sixth priority.

[0075] S203: The three-dimensional scene data to be rendered is received, the priority is sorted, the rendering techniques are called through the priority mobilization program, the rendering processing of the target features is performed respectively, and when a certain rendering technique ends, the rendering technique of the next priority is automatically entered;

[0076] S204: When the three-dimensional scene data to be rendered is completely rendered, the visualization processing automatically sends a stop rendering instruction to the GPU programmable drawing acceleration of the first priority, at which time the corresponding programs of other priorities also stop working.

[0077] S205: The three-dimensional scene data rendered is delivered to the display terminal, and the visualization rendering processing is performed by the display terminal.

[0078] The working principle and beneficial effects of the technical solution are as follows: first, according to the purposes of the multiple rendering technologies, the corresponding target features are defined, wherein the target features of the centralized rendering are: triangle vertex format, whether the vertex formats are the same, whether the number of used textures is the same, whether each texture is the same, whether illumination is used, whether the used materials are the same, whether the drawing methods are the same, and the like; the target features of the visibility culling rendering are: triangle vertex transformation, removing the scene parts that do not contribute to the final three-dimensional scene data rendering before the conversion from the three-dimensional scene model coordinates, and sending the remaining scene to the drawing pipeline; the target features of the multi-resolution model simplification rendering are: simplifying the small and unimportant parts and the remote scene in the three-dimensional scene model; the target features of the texture optimization rendering are: cutting textures, using multiple images to compress into a single texture or eliminating small textures, and the like, and high-resolution images are used for objects close to the observer, and low-resolution images are used for objects gradually far away from the observer; the target features of the image drawing rendering and acceleration are: drawing a virtual scene, using discrete images collected by a camera or continuous videos collected by a camera as basic data, generating panoramic images through image processing, and establishing a virtual environment with spatial manipulation ability through spatial correlation; the target features of the programmable graphics card drawing acceleration are: using triangle strip and triangle fan geometric representation methods to accelerate three-dimensional scene rendering; second, priorities of the multiple rendering technologies are defined according to the differences of the target features, and the content is that the programmable graphics card drawing acceleration is the first priority, the image drawing rendering is the second priority, the centralized rendering is the third priority, the visibility culling rendering is the fourth priority, the multi-resolution model simplification rendering is the fifth priority, and the texture optimization rendering is the sixth priority; third, the three-dimensional scene data to be rendered is received, the priorities are sorted, the rendering technologies are called through the priority call program, the rendering processing of the target features is respectively performed, and when a certain rendering technology ends, the rendering technology of the next priority is automatically entered; then, when the three-dimensional scene data to be rendered is completely rendered, the visual processing automatically sends a stop rendering instruction to the programmable graphics card drawing acceleration of the first priority, at this time, the corresponding programs of other priorities also stop working; finally, the rendered three-dimensional scene data is delivered to a display terminal, and visual presentation processing is performed by the display terminal; the above scheme defines different priorities for the multiple rendering technologies, realizes the overall planning of different rendering technologies, guarantees the accurate processing of various target features, provides a reliable data basis for the visual presentation of the three-dimensional scene data, the first priority provides hardware and software support for the rendering technologies of other priorities, guarantees the stable operation of the rendering technologies of other priorities, and improves the rendering effects of the respective rendering technologies; meanwhile, through the priority design, the stability of the corresponding programs of the rendering technologies is improved, and the orderly performance of various rendering technologies is effectively guaranteed.

[0079] Embodiment 4: as Figure 4As shown, on the basis of Embodiment 1, the processing procedure of image rendering of the three-dimensional scene data provided by the present embodiment comprises the following steps:

[0080] S206: Obtain the three-dimensional scene data to be rendered, and render the three-dimensional scene data to be rendered by using image rendering under the support of programmable rendering acceleration of the first priority graphics card;

[0081] S207: Retrieve the stored three-dimensional scene data from the visualization processing terminal, and use the three-dimensional scene data to draw a virtual scene instead of all geometric information of the three-dimensional scene data to be rendered;

[0082] S208: Use the discrete images collected by the camera or the continuous videos collected by the video camera as the basic data, generate panoramic images by image processing, and establish a virtual environment with spatial manipulation ability by spatial correlation of the panoramic images, render the three-dimensional scene data to be rendered, and obtain the three-dimensional scene data after image rendering;

[0083] The working principle and beneficial effects of the above technical solution are as follows: firstly, the three-dimensional scene data to be rendered is obtained, and the three-dimensional scene data to be rendered is rendered by using image rendering under the support of programmable rendering acceleration of the first priority graphics card; secondly, the stored three-dimensional scene data is retrieved from the visualization processing terminal, and the three-dimensional scene data is used to draw a virtual scene instead of all geometric information of the three-dimensional scene data to be rendered; finally, the discrete images collected by the camera or the continuous videos collected by the video camera are used as the basic data, panoramic images are generated by image processing, and a virtual environment with spatial manipulation ability is established by spatial correlation of the panoramic images, the three-dimensional scene data to be rendered is rendered, and the three-dimensional scene data after image rendering is obtained; the above scheme uses a series of image samples to draw a virtual scene instead of part or all geometric information of the scene, uses the discrete images collected by the camera or the continuous videos collected by the video camera as the basic data, generates panoramic images by image processing, and establishes a virtual environment with spatial manipulation ability by spatial correlation of the panoramic images; an information environment with immersive feeling and perfect interaction ability for users is constructed, and the limitation of the traditional virtual reality system based on computer graphics is solved, that is, the modeling process of complex space is quite tedious; the virtual information space constructed based on real scene images avoids complex scene modeling and rendering, and directly uses real scene images obtained by the camera or the video camera to construct the virtual information space; the panoramic image model is a representative image-based virtual reality model, and in the traditional panoramic image model, the entire scene is composed of a panoramic photo connected at the head and tail, and in some existing virtual roaming systems, hot spots can be specified on the photo to jump to other panoramic image models.

[0084] Embodiment 5: as Figure 5As shown, on the basis of Embodiment 4, the centralized rendering processing procedure in the three-dimensional scene data provided by the present embodiment comprises the following steps:

[0085] S209: receiving the three-dimensional scene data after image drawing rendering, and rendering the three-dimensional scene data after image drawing rendering by using triangles under the support of the programmable drawing acceleration of the first priority graphics card;

[0086] S210: classifying the triangles according to a standard, and the standard is vertex format, texture quantity, texture texture, illumination intensity, material and drawing mode;

[0087] S211: determining the quality level of the three-dimensional scene data to be rendered based on the resources of the graphics processor, generating at least one triangle and at least two sub-three-dimensional scene data to be rendered corresponding to the three-dimensional scene data to be rendered according to the quality level, classifying the triangles according to the standard, and combining the triangles of the same category to obtain the combined triangles of the same category;

[0088] S212: rendering the at least two sub-three-dimensional scene data to be rendered and the at least one combined triangle of the same category to obtain the three-dimensional scene data after centralized rendering;

[0089] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first receives the three-dimensional scene data after image rendering. With the support of the first priority graphics card programmable rendering acceleration, the three-dimensional scene data after image rendering is rendered using triangles. Secondly, the triangles are classified according to the following standards: vertex format, number of textures, texture quality, lighting intensity, material, and rendering method. Then, based on the resources of the graphics processor, the quality level of the three-dimensional scene data to be rendered is determined. Based on the quality level, at least one triangle and at least two sub-three-dimensional scene data to be rendered are generated corresponding to the three-dimensional scene data to be rendered. The triangles are classified according to the classification standards, and triangles of the same category are combined to obtain combined triangles of the same category. Finally, at least two sub-three-dimensional scene data and at least one triangle are rendered. The combined triangles of the same category; the above scheme classifies and recombines the triangles used in each rendering based on whether the vertex format is the same, whether the number of textures used is the same, whether each texture is the same, whether lighting is used, whether the materials used are the same, and whether the drawing method is the same, which greatly reduces the number of times the graphics processor calls the display card for each frame and avoids unnecessary switching of rendering resources and rendering states; ultimately achieving the goal of improving efficiency; since the 3D scene data in mine safety monitoring contains a large number of equipment models and pipeline models, which are numerous and repetitive, centralized rendering technology can realize the simultaneous rendering of multiple identical models in one rendering, thereby effectively improving rendering efficiency. Compared with existing technologies, centralized rendering technology can improve rendering efficiency by 30%-50% in the system.

[0090] Example 6: As Figure 6 As shown, based on Example 5, the visibility culling rendering process provided in this embodiment of the invention for processing the centrally rendered 3D scene data includes the following steps:

[0091] S213: Obtain the object model of each scene in the centrally rendered 3D scene data, and remove the 3D scene data parts that do not contribute value to obtain the occlusion information of the object models after removal. The contribution value is based on the feature correlation degree between the object model and the background in the 3D scene data. The feature correlation network is used to calculate the correlation features between the object model features and the background features to obtain the correlation features corresponding to the object model features and the background features. Based on the classification network, the correlation features are analyzed to determine the feature correlation degree between the object model and the background in the 3D scene data. Then, it is compared with the preset feature correlation degree threshold. If the preset feature correlation degree threshold is not reached, it is considered that there is no contribution value.

[0092] S214: Based on the occlusion information, determine the occlusion type of the scene object model after culling, and obtain the target scene object model based on the occlusion type and the scene object model after culling.

[0093] S215: Render the object model of the target scene to obtain the 3D scene data after visibility culling rendering;

[0094] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first acquires each scene object model in the centrally rendered 3D scene data, and removes the 3D scene data parts that do not contribute value, obtaining the occlusion information of the scene object models after removal. The contribution value is based on the feature correlation degree between the object model and the background in the 3D scene data. The feature correlation network is used to calculate the correlation features between the object model features and the background features to obtain the correlation features corresponding to the object model features and the background features. Based on the classification network, the correlation features are analyzed to determine the feature correlation degree between the object model and the background in the 3D scene data, and then compared with the preset feature correlation degree threshold. If the preset feature correlation degree threshold is not reached, it is considered that there is no contribution value. Secondly, based on the occlusion information, the scene after removal is determined. The occlusion type of the object model is determined, and based on the occlusion type and the scene object model after culling, the target scene object model is obtained. Then, the target scene object model is rendered to obtain the 3D scene data after visibility culling rendering. As can be seen from the vertex transformation of the triangle, by using visibility culling to remove scene parts that do not contribute to the final image rendering before the 3D model coordinate transformation stage, and then sending the remaining scene to the rendering path, the complexity of the 3D scene of the mine and the burden on the graphics pipeline can be effectively reduced. It is a very effective method to improve the rendering efficiency of the 3D scene of the mine. The method and results of this embodiment are also applicable to other parts, such as performing inaccurate calculations on invisible parts during collision detection. Visibility culling is aimed at quickly removing invisible geometric elements.

[0095] Example 7: Figure 7 As shown in Example 6, the multi-resolution model provided in this embodiment of the invention simplifies the rendering process of the 3D scene data after visibility culling rendering, and includes the following steps:

[0096] S216: Obtain the 3D scene data after visibility culling rendering, determine the objects in the 3D scene data, and obtain the visible area and visible polygons based on the camera's shooting position;

[0097] S217: Simulate the inaccurate critical state when the position and shape of the simulated object are simplified based on the initial error control threshold of the point-to-point method;

[0098] S218: Based on the distance between points corresponding to the position and shape, calculate the corrected error control threshold through the initial error control threshold, and simplify the region within the cuboid view using the corrected error control threshold as the standard.

[0099] S219: The rendering list is sent to the renderer for rendering, and the simplified rendering of the multi-resolution model is obtained as 3D scene data.

[0100] The working principle and beneficial effects of the above technical solution are as follows: This embodiment obtains the 3D scene data after visibility culling rendering, determines the objects in the 3D scene data, and obtains the visible area and visible polygons according to the camera's shooting position; simulates the inaccurate critical state when the position and shape of the object are simplified according to the initial error control threshold of the point-to-point method; calculates the corrected error control threshold according to the distance between the points corresponding to the position and shape respectively, and simplifies the region within the cuboid view using the corrected error control threshold as the standard; the rendering list is sent to the renderer for rendering, and the 3D scene data after multi-resolution model simplification rendering is obtained; the above solution simplifies the small, unimportant parts and distant scenes in the 3D scene data model, while ensuring that the scene content is not seriously distorted; its purpose is to generate detailed hierarchical models of objects in the scene and improve the rendering speed of the entire mine 3D scene; the automatic processing technology can also reduce the workload of manually editing large-scale 3D scenes; since most of the mesh model is represented by triangular patches, and even if the original model is a polygonal mesh, it can be triangulated, the mesh model simplification is to minimize the number of triangles and vertices of the original model while maintaining the original model features as much as possible.

[0101] Example 8: As Figure 8 As shown in Example 7, the texture optimization rendering process provided in this embodiment of the invention for processing 3D scene data after simplified rendering of multi-resolution models includes the following steps:

[0102] S220: Cluster the textures in the fine 3D scene data based on the density clustering algorithm to obtain the first clustering result, and compress multiple 3D scene data into a single texture.

[0103] S221: For each point in a single texture, if the current point is an unclassified point, assign a classification number to the current point and determine whether the remaining points belong to the same triangular facet. If they belong to the same triangular facet, regardless of which cluster the remaining points are in the first clustering result, modify the classification number of the remaining points to the classification number of the current point to obtain the second clustering result.

[0104] S222: Calculate the minimum bounding rectangle of each cluster in the second clustering result to determine the effective texture;

[0105] The working principle and beneficial effects of the technical solution are as follows: the embodiment first performs clustering on the texture in the fine three-dimensional scene data based on a density clustering algorithm to obtain a first clustering result, and compresses the plurality of three-dimensional scene data into a single texture; secondly, each point in the single texture is judged, if the current point is an unclassified point, the current point is assigned a classification number, and whether the remaining points belong to the same triangular patch as the current point is judged, if they belong to the same triangular patch, the classification number of the remaining points is modified to the classification number of the current point regardless of which cluster of the first clustering result the remaining points are located in, to obtain a second clustering result; finally, the minimum circumscribed rectangle of each cluster in the second clustering result is calculated to determine the effective texture; the above scheme can effectively improve the use efficiency of texture memory by compressing a plurality of images into a single texture, and high-resolution images are used for objects close to the observer, and when the objects gradually move away from the observer, low-resolution images are used, which can improve the quality of scene rendering. The clipping texture also provides texture paging loading from the disk, and there is no need to always save all texture data in the memory; for paging scheduling, the clipping texture uses an image buffer in the main memory, the image buffer always maintains the same size and is updated with the movement of the viewpoint on the clipping texture; the texture to be used needs to be pre-filtered before being used by the clipping texture, and is segmented on the disk.

[0106] In the embodiment 1, the graphics card programmable rendering acceleration adopts a rendering model; the rendering model is used to calculate the operation capacity of the graphics card of the graphics card programmable rendering acceleration;

[0107] The function expression of the rendering model is:

[0108]

[0109] wherein, I(s) represents the visualization result of the texture optimization rendering of the three-dimensional scene data s, I0 represents the visualization result when the three-dimensional scene data s is input into the graphics card programmable rendering acceleration, s represents the three-dimensional scene data, τ(t) represents the coefficient of the operation capacity of the graphics card, the part other than I0 is the full-speed operation capacity of the graphics card, and the expression of the full-speed operation capacity a is:

[0110]

[0111] wherein, if τ=0, it indicates that the operation capacity of the graphics card is normally exerted, and if τ is infinite, it indicates that the operation capacity of the graphics card is exerted by 100%;

[0112] The working principle and beneficial effects of the technical solution are as follows: the programmable drawing acceleration of the graphics card in the embodiment adopts a drawing model; the drawing model is used to calculate the operation capacity of the graphics card of the programmable drawing acceleration of the graphics card, so that the accelerated drawing technology is greatly improved; through the trial, it is found that the organization structure of the optimized geometric representation can save a large amount of transformation and illumination calculation on the graphics processor, and reduce the occupation of the bandwidth between the central processing unit and the graphics hardware; the effective organization of the geometric data for acceleration can be applied to most polygon or triangle representation scenes; the geometric representation method of the triangle strip and the triangle fan can accelerate the three-dimensional scene rendering.

[0113] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A visualization processing method for a mine safety monitoring system, characterized in that, Includes the following steps: The visualization processing terminal receives 3D scene data sent by the mine safety monitoring system, extracts, cleans, and standardizes the 3D scene data to obtain preprocessed 3D scene data. The 3D scene data is rendered using a variety of rendering techniques to obtain the rendered 3D scene data. The priority of the rendering techniques is set according to the different target features. Define corresponding target features based on the purpose of various rendering techniques; The system prioritizes various rendering techniques based on different target features. The priorities are as follows: GPU programmable rendering acceleration is the first priority, image rendering is the second priority, concentrated rendering is the third priority, visibility culling rendering is the fourth priority, multi-resolution model simplification rendering is the fifth priority, and texture optimization rendering is the sixth priority. The system receives 3D scene data to be rendered, sorts it according to priority, and retrieves the rendering techniques according to the priority. The system then performs rendering processing on the target features respectively. Once one rendering technique is completed, the system automatically moves to the next priority rendering technique. Once all the 3D scene data to be rendered is completed, the visualization process automatically sends a stop rendering command to the highest priority programmable rendering acceleration of the graphics card. At this time, the corresponding programs of other priorities also stop working. The rendered 3D scene data is then sent to the display terminal for visualization processing. Rendering techniques include centralized rendering, visibility culling rendering, multi-resolution model simplification rendering, texture optimization rendering, image rendering and acceleration, and graphics card programmable rendering acceleration; The target features of centralized rendering are: triangle vertex format, whether the vertex format is the same, whether the number of textures used is the same, whether each texture is the same, whether lighting is used, whether the materials used are the same, and whether the drawing method is the same. The target features of visibility culling rendering are: triangle vertex transformation, removing scene parts that do not contribute to the final 3D scene data rendering before the 3D scene model coordinate transformation stage, and sending the remaining scene to the rendering pipeline. The target features of multi-resolution model simplification rendering are: simplifying small, unimportant parts and distant scenes in the 3D scene model. The target features of texture optimization rendering are: texture clipping, using multiple images to compress into a single texture or eliminating fine textures, with high-resolution images close to the observer's object, and low-resolution images used when the object moves away from the observer. The target features of image rendering and acceleration are: rendering virtual scenes, using discrete images captured by a camera or continuous video captured by a video camera as basic data, generating panoramic images through image processing, and establishing a virtual environment with spatial manipulation capabilities through spatial correlation; the target features of graphics card programmable rendering acceleration are: using geometric representation methods of triangle strips and triangle fans to accelerate the rendering of three-dimensional scenes. The display terminal visualizes the rendered 3D scene data, enabling risk warnings and scientific decision-making for mine safety monitoring.

2. The visualization processing method of the mine safety monitoring system as described in claim 1, characterized in that, The process of extracting, cleaning, and standardizing preprocessing 3D scene data includes the following steps: The system receives instructions to receive 3D scene data. Based on these instructions, the 3D scene data is input into the data extraction network to obtain the name of the target 3D scene data and its corresponding parameter values, thus forming a target 3D scene dataset. The name of the target 3D scene data includes one or more names of 3D scene data with the same meaning. Check if there is duplicate data in the target 3D scene dataset. If so, delete the duplicate data. Check if there are outliers in the dataset. If so, delete the outliers. The result is a target 3D scene dataset with duplicate data and outliers removed. Outliers are caused by equipment failure, manual input errors, or abnormal events. The target 3D scene dataset, after removing duplicate data and outliers, is converted into a format recognizable by the visualization processing terminal and grouped according to different types of 3D scene data to obtain preprocessed 3D scene data.

3. The visualization processing method for the mine safety monitoring system as described in claim 1, characterized in that, The image rendering process in 3D scene data includes the following steps: The system acquires the 3D scene data to be rendered and, with the support of the first-priority programmable rendering acceleration of the graphics card, renders the 3D scene data using image rendering. The stored 3D scene data is retrieved from the visualization processing terminal and used to replace all the geometric information of the 3D scene data to be rendered to draw a virtual scene. Using discrete images captured by a camera or continuous video captured by a video camera as basic data, panoramic images are generated through image processing, and spatial correlation is established to create a virtual environment with spatial manipulation capabilities. The 3D scene data to be rendered is then rendered to obtain the 3D scene data after image rendering.

4. The visualization processing method for the mine safety monitoring system as described in claim 3, characterized in that, The centralized rendering process in 3D scene data includes the following steps: Receive the 3D scene data after image rendering, and with the support of the first priority graphics card programmable rendering acceleration, render the 3D scene data after image rendering using triangles; Triangles are classified according to the following criteria: vertex format, number of textures, texture quality, lighting intensity, material, and rendering method. Based on the resources of the graphics processor, the quality level of the 3D scene data to be rendered is determined. The quality level is combined to generate at least one triangle and at least two sub-3D scene data to be rendered, which correspond to the 3D scene data to be rendered. The triangles are classified according to the classification criteria, and triangles of the same category are combined to obtain a combined triangle of the same category. Render at least two sub-3D scene data to be rendered and at least one combined triangle of the same type to obtain the centrally rendered 3D scene data.

5. The visualization processing method for the mine safety monitoring system as described in claim 4, characterized in that, Visibility culling rendering is a process of processing the 3D scene data after centralized rendering, which includes the following steps: Obtain the occlusion information of each scene object model in the centrally rendered 3D scene data, and remove the 3D scene data parts that do not contribute value. Based on the occlusion information, determine the occlusion type of the scene object model after culling, and obtain the target scene object model based on the occlusion type and the scene object model after culling. The object model of the target scene is rendered to obtain the 3D scene data after visibility culling rendering.

6. The visualization processing method for the mine safety monitoring system as described in claim 5, characterized in that, The contribution value is based on the feature correlation between the object model and the background in the 3D scene data. The feature correlation network is used to calculate the correlation features between the object model features and the background features to obtain the correlation features corresponding to the object model features and the background features. Based on the classification network, the correlation features are analyzed to determine the feature correlation between the object model and the background in the 3D scene data. Then, it is compared with the preset feature correlation threshold. If the preset feature correlation threshold is not reached, it is considered that there is no contribution value.

7. The visualization processing method for the mine safety monitoring system as described in claim 5, characterized in that, Multi-resolution models simplify the processing of 3D scene data after visibility culling rendering, and include the following steps: Obtain the 3D scene data after visibility culling rendering, determine the objects in the 3D scene data, and obtain the visible area and visible polygons based on the camera's shooting position; The initial error control threshold of the point-to-point method is used to simulate the inaccurate critical state when the position and shape of the object are simplified. Based on the distance between points corresponding to the location and shape, the corrected error control threshold is calculated using the initial error control threshold, and the region within the cuboid view is simplified using the corrected error control threshold as the standard. The rendering list is sent to the renderer for rendering, resulting in simplified 3D scene data after multi-resolution model rendering.

8. The visualization processing method for the mine safety monitoring system as described in claim 7, characterized in that, Texture-optimized rendering involves processing 3D scene data after simplified rendering of multi-resolution models, and includes the following steps: Based on the density clustering algorithm, the textures in the fine 3D scene data are clustered to obtain the first clustering result, and multiple 3D scene data are compressed into a single texture. For each point in a single texture, if the current point is an unclassified point, then assign a classification number to the current point and determine whether the remaining points belong to the same triangular facet. If they belong to the same triangular facet, then regardless of which cluster the remaining points are in the first clustering result, modify the classification number of the remaining points to the classification number of the current point to obtain the second clustering result. Calculate the minimum bounding rectangle of each cluster in the second clustering result to determine the effective texture.

Images