Robot multi-layer cooperation method and system for grotto cultural relic protection and navigation

Abstract

The invention discloses a robot multi-layer cooperation method and system for grotto cultural relic protection and guide, and the method comprises the steps: collecting the multi-dimensional data of grotto cultural relics and surroundings through distributed sensing nodes, carrying out the classified storage, and calling a cultural relic defect hierarchical recognition and evaluation model to complete the preliminary screening and type judgment of defects, the method comprises the following steps: scheduling a detection robot to execute refined detection and defect level quantitative evaluation based on a multi-layer cooperative dynamic task allocation mechanism of the robot, starting a style migration model for a wall painting fading defect to carry out color reconstruction and style adaptation, and planning an optimal path through a dynamic avoidance algorithm based on environment obstacle information; and integrating various data to generate personalized navigation content and transmitting the personalized navigation content to the navigation robot to execute a task. Through multi-step cooperation and model algorithm support, intelligent linkage of defect identification, fading repair, path planning and navigation service is realized, and the cooperation efficiency of grotto cultural relic protection and navigation is improved.

Description

Technical Field

[0001] This invention relates to the field of cultural relic protection and tour guide technology, and in particular to a multi-layered collaborative method and system for robots for the protection and tour guide of grotto cultural relics. Background Technology

[0002] As important heritage sites carrying historical and cultural value, grottoes face the dual challenges of complex environments and diverse needs in their protection and guided tours. Grottoes are mostly located in outdoor natural environments, subject to long-term weathering and erosion, leading to surface defects and fading of murals. Furthermore, the complex internal spatial structure and irregular distribution of obstacles within grottoes greatly complicate conservation efforts. Traditional guided tours rely heavily on human explanations, making it difficult to provide personalized content based on the real-time conservation status of the artifacts. Moreover, manual conservation work is inefficient and lacks precision, failing to meet the refined conservation needs of large-scale grottoes. With the development of robotics and intelligent algorithms, applying multi-layered collaborative robot systems to grotto conservation and guided tours, enabling intelligent conservation operations and personalized guided tours, has become a key direction for addressing the current pain points in grotto conservation and guided tours.

[0003] Existing technologies in the field of grotto cultural relic protection and tour guidance have two significant shortcomings: First, there is a lack of systematic collaborative mechanisms for the identification and restoration of cultural relic defects. Existing defect identification models mostly operate independently and are not deeply integrated with the robot collaboration system. They cannot dynamically adjust identification parameters according to the grotto site environment, and the restoration models are difficult to accurately match the original style of the cultural relic, resulting in insufficient coordination between the restoration effect and the cultural relic itself. At the same time, there is a lack of refined assessment of defect levels, affecting the targeted nature of protection operations. Second, the efficiency of robot path planning and task allocation is low. Existing avoidance algorithms do not fully consider the dynamic obstacles inside the grotto and the collaborative needs of multi-level robot collaboration. Path planning lacks flexibility, and the tour guide content management platform does not effectively integrate cultural relic protection data and cultural background information, and cannot achieve linkage updates between protection data and tour guide content. This results in a lack of real-time and targeted tour guide content. In the process of robot collaboration, the task allocation is unreasonable, making it difficult to balance the efficiency of protection operations and the quality of tour guide services. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, the purpose of this invention is to provide a multi-layered collaborative method and system for robots in the protection and guidance of grotto cultural relics.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention provides a multi-layered collaborative robot method for the protection and guidance of grotto cultural relics, comprising the following steps: S1 collects surface texture features, spatial three-dimensional coordinates, and multi-dimensional data of the surrounding environment of the grotto cultural relics through distributed sensing nodes in the multi-layer collaborative robot system, and transmits the collected data to the intelligent management platform for tour guide content for data classification and storage. S2, invoke the cultural relic defect hierarchical identification and evaluation model to extract features and perform preliminary defect screening on the classified and stored data, and make a preliminary determination of the defect type through the multi-level identification rules built into the model; S3. Based on the preliminary judgment results, a task allocation model for multi-layered robot collaboration for the protection and guidance of grotto cultural relics is used to schedule a dedicated detection robot to perform refined defect detection. Combined with the defect feature parameters output by the model, a quantitative assessment of the mural fading defect level is completed. S4. For the mural fading defect areas confirmed by the assessment, the mural fading restoration style transfer model is activated, the original color data of cultural relics and historical style reference data collected by the robot are imported, and the color reconstruction and style adaptation of the fading areas are performed. S5, based on the internal spatial structure data of the grotto and real-time environmental obstacle information, runs a dynamic obstacle avoidance algorithm for the robot to plan the optimal path for robot protection operations and guided travel; S6 integrates defect assessment results, repair processing data, and grotto cultural background information through the content generation model of the intelligent management platform for tour guide content. Combined with the interactive response mechanism of multi-layered robot collaboration, it generates personalized tour guide data and transmits it to the tour guide robot to perform tour guide tasks.

[0006] Preferably, the expression of the hierarchical identification and evaluation model for cultural relic defects is: ,in, This is the assessment value for the level of defects in cultural relics. For the first Weighting coefficients for class defects For the first Feature extraction values ​​of class defects, For the first in multi-layer robot collaboration The confidence level of data collection at each sensing node. This is a parameter for adjusting the sensitivity of defect identification. For the first Texture complexity parameters for each cultural relic area This is the threshold for defect determination. Spatial correlation factor, Grotto cultural relics OK Column pixel correlation matrix, For the first OK The gradient change value of the column pixels.

[0007] Preferably, the expression for the style transfer model for mural fading restoration is: ,in, The restored color values, The original color retention factor, For the original color data of cultural relics, For style transfer transformation matrix, To integrate historical style with weight, For the first The weights of each historical style sample, For the first Historical style reference data, This is the fading correction factor. This is a quantitative value for the degree of fading. This is an element-wise product operation.

[0008] Preferably, the expression for the robot obstacle area dynamic avoidance algorithm is: in, The optimal path, The set of candidate paths, The path length weight is used for weighting. For path length, To avoid obstacles and optimize weights, For the first An obstacle to the path Influence factors For robot size parameters, For environmental adaptability coefficient, For path The corresponding environmental gradient change, These are the path coordination parameters for multi-layered robot collaboration.

[0009] Preferably, the content generation model expression of the intelligent management platform for tour guide content is: ,in, For personalized tour guide content data, For defect information weights, For the data of the cultural relic defect assessment results, A matrix relating the cultural background of grottoes. Weighting based on user preferences For the first Weighting coefficients for user preferences For the first User preference feature data A content distribution matrix for multi-layered robot collaboration. For environmental information fusion coefficient, This provides real-time data on the grotto environment.

[0010] Preferably, the task allocation model expression for the multi-layer collaborative robot system for the protection and guidance of grotto cultural relics is as follows: ,in, Assign a result matrix to the task. For the first Priority weights of tasks For the first The complexity parameter of the task. For the robot to the first The adaptability parameters of the task. For the first Communication bandwidth parameters of each distributed node For the first The communication delay gradient of each node, The coefficient for collaborative equilibrium. For the first The operational efficiency parameters of each robot The number of robots participating in the collaboration.

[0011] Preferably, step S3 includes the following sub-steps: S31, through the main control node in the robot multi-layer collaborative system, receives the preliminary screening results output by the cultural relic defect hierarchical identification and evaluation model, extracts the spatial coordinates, texture feature differences and gray-scale distribution features of the defect candidate area, and standardizes and organizes the extracted feature parameters according to the preset data format; S32, based on the standardized feature parameters, calls the task priority sorting module of the robot's multi-layer collaboration, constructs multi-dimensional sorting indicators according to the area, location importance and repair urgency of the defect area, and arranges the tasks that need to be performed for fine detection in order. S33. Based on the sorting results, the detection task is assigned to the corresponding detection robot through a dynamic task allocation mechanism. The feature parameters of the defect candidate area and the accuracy requirements of the detection operation are transmitted synchronously to ensure that the detection robot performs the defect detection operation according to the specified parameters.

[0012] Preferably, step S4 includes the following sub-steps: S41. Retrieve historical color archive data and style feature parameters of the corresponding murals from the intelligent management platform for tour content, and combine them with real-time image data of the faded areas collected by the robot to construct a reference dataset for color restoration. S42, input the reference dataset into the style transfer model for mural fading restoration, set the style transfer intensity parameter and color fusion ratio parameter of the model, start the iterative calculation process of the model, and gradually optimize the color reconstruction effect of the faded area; S43 extracts the restored image data output by the model, performs edge fusion processing with the surrounding area images collected in real time by the robot, eliminates the visual difference between the restored area and the original area, forms complete mural image data, and transmits it to the intelligent management platform for tour guide content for storage.

[0013] Preferably, step S5 includes the following sub-steps: S51 collects obstacle location data, spatial size data, and dynamic environmental change data inside the grotto through the environmental perception sensors on the robot, and transmits the collected data to the input layer of the robot's obstacle area dynamic avoidance algorithm. S52, the algorithm classifies the obstacle type and assesses the risk level of the input data, and determines the safety boundary parameters and passage priority of the path planning by combining the path planning constraints of the robot's multi-layer collaboration. S53, based on the classification evaluation results and constraint parameters, uses the algorithm's path search module to generate multiple candidate travel paths, and calculates the passage efficiency parameter and obstacle avoidance success rate parameter for each path; S54: Based on the calculation results, the optimal path is selected, and the path coordinate data and travel control parameters are transmitted to the robot's motion control module to guide the robot to perform tasks and guide movement according to the planned path.

[0014] The second aspect of this invention also provides a multi-layered collaborative robot system for the protection and guidance of grotto cultural relics. This system, applied to the aforementioned multi-layered collaborative robot method for the protection and guidance of grotto cultural relics, includes: a distributed data acquisition unit for multi-dimensional grotto cultural relics, an intelligent processing unit for hierarchical identification and assessment of cultural relic defects, a style transfer calculation unit for mural fading restoration, a dynamic obstacle avoidance path planning unit for robots, an intelligent integration and distribution unit for guide content, and a multi-layered collaborative control unit for robots; wherein... The multi-dimensional distributed data acquisition unit for grotto artifacts collects data on artifacts and the environment through distributed sensing nodes, and transmits the data to the intelligent processing unit for hierarchical identification and assessment of artifact defects. After identifying and assessing the data, the intelligent processing unit sends the defect data and location information to the style transfer calculation unit for mural fading restoration and the robot obstacle avoidance path planning unit, respectively. After completing the fading restoration, the style transfer calculation unit transmits the data to the intelligent integration and distribution unit for tour guide content. The robot obstacle avoidance path planning unit generates the optimal path and sends it to the multi-layer collaborative control unit for robots. The intelligent integration and distribution unit for tour guide content integrates various data to generate tour guide content, which is then transmitted to the multi-layer collaborative control unit for robots. The multi-layer collaborative control unit for robots receives the path data and tour guide content, coordinates the robots to perform protection operations and tour guide tasks, and achieves efficient collaboration between the protection and tour guide functions of grotto artifacts.

[0015] Beneficial effects: This invention proposes a multi-layered robot collaboration method and system for the protection and guidance of grotto cultural relics. By using the distributed sensing nodes of the multi-layered robot collaboration system, it achieves comprehensive collection of multi-dimensional data on grotto cultural relics and the environment. Combined with a hierarchical identification and evaluation model for cultural relic defects, it completes accurate defect identification and hierarchical quantitative evaluation, solving the problems of disconnect between defect identification and robot collaboration and insufficient evaluation refinement in existing technologies. At the same time, by using a style transfer model for mural fading restoration to import original color and historical style data, it achieves style-adaptive restoration of faded areas, making up for the shortcomings of traditional restoration effects and lack of coordination with cultural relics.

[0016] By combining a dynamic obstacle avoidance algorithm with a multi-layered collaborative dynamic task allocation mechanism, the system flexibly plans optimal paths and rationally allocates protection and guidance tasks, solving the problems of insufficient flexibility in existing path planning and unreasonable task allocation. The intelligent management platform for guidance content integrates defect assessment, repair data, and cultural background information to generate personalized guidance content, achieving synchronized updates between protection data and guidance content, overcoming the shortcomings of traditional guidance methods that lack real-time updates and specificity. The collaborative operation of all units in this system improves the accuracy and efficiency of cultural relic defect identification and repair, while ensuring smooth robot operation and guidance, achieving efficient collaboration between grotto cultural relic protection and guidance, and fully meeting the dual needs of refined protection and personalized guidance for large-scale grotto cultural relics. Attached Figure Description

[0017] Figure 1 A schematic diagram illustrating the overall steps of the multi-layered robot collaboration method for the protection and guidance of grotto cultural relics provided by this invention. Figure 2 This is a flowchart illustrating step S3 in the multi-layered collaborative robot method for the protection and guidance of grotto cultural relics provided by the present invention. Figure 3 This is a flowchart of step S4 in the multi-layered robot collaboration method for the protection and guidance of grotto cultural relics provided by the present invention. Figure 4 This is a flowchart of step S5 in the multi-layered robot collaboration method for the protection and guidance of grotto cultural relics provided by the present invention. Figure 5 This is a diagram showing the components of the multi-layered collaborative robot system for the protection and guidance of grotto cultural relics according to the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, a multi-layered collaborative robot method for the protection and guidance of grotto cultural relics includes the following steps: S1 collects surface texture features, spatial three-dimensional coordinates, and multi-dimensional data of the surrounding environment of the grotto cultural relics through distributed sensing nodes in the multi-layer collaborative robot system, and transmits the collected data to the intelligent management platform for tour guide content for data classification and storage. Specifically, step S1 utilizes at least 30 distributed sensing nodes deployed within a multi-layered robotic collaborative system to complete data acquisition and storage. These sensing nodes include high-resolution image sensors, 3D laser scanners, and environmental monitoring modules. The image sensors are set to a sampling frequency of 30 frames per second, the 3D laser scanner's measurement accuracy is controlled within 0.1 millimeters, and the environmental monitoring module simultaneously collects multi-dimensional data such as temperature, humidity, light intensity, and air quality. The data collected covers surface texture features of the grotto artifacts, including crack width, wear depth, and pigment adhesion. The spatial 3D coordinate system is established with a preset reference point at the grotto entrance as the origin. The acquisition range includes the artifact itself and environmental data within a 5-meter radius. Each distributed sensing node transmits the collected raw data in real-time to the intelligent management platform for tour guide content via a 5G communication module. The platform's built-in data classification and storage module categorizes the data into texture feature datasets, 3D coordinate datasets, and environmental parameter datasets, each with its own independent storage directory. A distributed database architecture is used to achieve parallel storage and rapid retrieval of data. Simultaneously, a data verification mechanism eliminates abnormal data that occurs during the acquisition process, ensuring the integrity and accuracy of the stored data and providing reliable data support for subsequent defect identification, repair, and tour guide content generation.

[0020] S2, invoke the cultural relic defect hierarchical identification and evaluation model to extract features and perform preliminary defect screening on the classified and stored data, and make a preliminary determination of the defect type through the multi-level identification rules built into the model; Specifically, step S2 activates the hierarchical identification and assessment model for cultural relic defects integrated into the intelligent management platform for guided tour content. This model incorporates a 128-dimensional feature extraction vector and 8 levels of defect identification rules, and performs targeted processing on the three types of classified datasets. The model first extracts features from the texture feature dataset, using a 16×16 pixel convolution operation to extract key features such as the gray-level co-occurrence matrix and edge gradients of the cultural relic surface. Simultaneously, it performs point cloud registration processing on the 3D coordinate dataset to obtain the spatial morphological features of the cultural relic surface. In the initial defect screening stage, the model employs a multi-threshold judgment mechanism, setting a crack width of 0.5 mm and a wear depth of 1 mm as the initial defect judgment thresholds. Combined with environmental parameter data such as humidity variation and light intensity fluctuations, areas exceeding the thresholds or exhibiting abnormal fluctuations are marked as candidate defect areas. Subsequently, the model uses its built-in 8-level recognition rules to compare the feature parameters of the defect candidate area with a preset feature library of common defect types in grotto cultural relics. The defect types include four categories: cracks, peeling, fading, and weathering. Each type of defect corresponds to a 3-level severity classification. The defect type and preliminary severity are determined by calculating the feature matching degree. During the determination process, the feature matching threshold of the model is set to 0.85 to ensure the accuracy of the preliminary screening results and provide a clear target guidance for subsequent refined exploration.

[0021] S3, based on the preliminary judgment results, utilizes the multi-layer collaborative task allocation model of robots to schedule dedicated detection robots to perform refined defect detection, and combines the defect feature parameters output by the model to complete the quantitative assessment of the defect level; Specifically, step S3, based on the preliminary judgment results of step S2, activates the multi-layer collaborative task allocation model of the robots. This mechanism incorporates a task priority evaluation algorithm, classifying five priority levels according to the locational importance of the defect candidate area, the initial severity of the defect, and the difficulty of detection. The system schedules 3-5 dedicated inspection robots to perform refined defect detection. These robots are equipped with high-precision ultrasonic detectors, infrared thermal imagers, and microscopic imaging modules. The ultrasonic detector has a detection depth range of 0-50 mm, the infrared thermal imager has a temperature resolution of 0.02℃, and the microscopic imaging module has a magnification of up to 200x. Based on the target coordinates issued by the dynamic task allocation mechanism, the inspection robots autonomously navigate to the defect candidate area. The ultrasonic detector acquires the internal structural features of the defect, the infrared thermal imager detects the temperature distribution differences inside the artifact, and the microscopic imaging module captures the microscopic morphology of the defect area. Meanwhile, the inspection robot receives defect feature parameters output by the cultural relic defect hierarchical identification and evaluation model, including the texture complexity, gray-level distribution variance, and spatial location coordinates of the defect candidate area. Combined with the refined data it collects, it calculates the quantitative evaluation value of the defect level using a weighted summation method. The evaluation indicators include defect area, depth, diffusion trend, and degree of impact on the structural stability of the cultural relic. Each indicator corresponds to a different weight coefficient. The final level of the defect is determined through quantitative evaluation, providing accurate technical parameter support for subsequent restoration work.

[0022] S4. For the mural fading defect areas confirmed by the assessment, the mural fading restoration style transfer model is activated, the original color data of cultural relics and historical style reference data collected by the robot are imported, and the color reconstruction and style adaptation of the fading areas are performed. Specifically, step S4, targeting the fading defects in the murals identified in step S3, initiates the mural fading restoration style transfer model. This model includes a color reconstruction module and a style adaptation module, supporting restoration processing for different types of pigment fading. First, the image acquisition robot within the multi-layered collaborative robot system re-acquires real-time image data of the fading area. During acquisition, the light intensity is controlled to remain between 500-800 lux to ensure the authenticity of the color data. Simultaneously, the original color archive data and historical style reference data of the mural are retrieved from the historical database of the intelligent management platform for tour content. The original color data includes the RGB three-color channel value range of the pigments, while the historical style reference data includes information such as the color matching ratios and painting techniques of murals from the same period in the grotto. The real-time acquired image data of the fading area, the original color data, and the historical style reference data are imported into the model's color reconstruction module. The number of color reconstruction iterations is set to 50, with a learning rate adjusted to 0.001 for each iteration. Basic color restoration of the fading area is achieved through pixel-by-pixel color calibration. The style adaptation module is then activated. Based on the color matching ratios and painting techniques in the historical style reference data, the restored colors are style-optimized to ensure that the color transitions and brushstroke styles of the restored area are consistent with the overall mural. During the restoration process, the color differences between the restored area and the surrounding normal areas are compared in real time to keep the difference within 5% and ensure the naturalness and harmony of the restoration effect. Finally, complete restored image data is generated and transmitted to the intelligent management platform for tour guide content.

[0023] S5, based on the internal spatial structure data of the grotto and real-time environmental obstacle information, runs a dynamic obstacle avoidance algorithm for the robot to plan the optimal path for robot protection operations and guided travel; Specifically, step S5 uses the grotto's internal spatial structure data stored in the intelligent management platform for guided tour content, along with real-time environmental obstacle information collected by the robot, to run a dynamic obstacle avoidance algorithm for precise path planning. The grotto's internal spatial structure data includes parameters such as passage width, height, turning angles, and the location of artifact displays. Passage width data covers all traversable areas within the grotto, with a minimum recorded width of 1.2 meters. Turning angles range from 30 to 120 degrees. The robot uses its onboard LiDAR and vision sensors to collect real-time environmental obstacle information. The LiDAR sensor has a detection range of 0.5-10 meters and a scanning frequency of 10 Hz, enabling it to identify static obstacles such as pillars and walls, as well as dynamic obstacles such as visitors and temporary facilities. It also acquires parameters such as obstacle size, location, and movement speed. The algorithm first fuses the spatial structure data and obstacle information to construct a 3D environmental map of the grotto's interior, setting a safe distance parameter of 0.3 meters for path planning, meaning the minimum distance between the robot and obstacles and artifacts is no less than 0.3 meters. Subsequently, a path search algorithm was used to generate multiple candidate paths. The length, travel time, and obstacle avoidance difficulty of each path were used as evaluation indicators. The travel time calculation was combined with the robot's movement speed parameter, with the robot's movement speed set to 0.2 meters per second during protection operations and 0.5 meters per second during guided tours. The optimal path was selected through comprehensive evaluation of multiple indicators. The path planning process simultaneously considered the collaborative needs of multiple robot layers to avoid path conflicts between different robots and ensure the smoothness of protection operations and guided tours. The planned path data included a series of continuous coordinate points, with the distance between adjacent coordinate points set to 0.1 meters.

[0024] S6 integrates defect assessment results, repair and processing data, and grotto cultural background information through the intelligent management platform for tour content. Combined with the multi-layered collaborative interactive response mechanism of the robot, it generates personalized tour data and transmits it to the tour robot to perform tour tasks.

[0025] Specifically, step S6 integrates and processes multi-source data through the intelligent management platform for guided tour content. The integrated data includes the defect level quantitative assessment results output from step S3, the repair processing data generated in step S4, and pre-stored grotto cultural background information. This information includes historical period, construction techniques, artistic style, and cultural connotations, with each piece of information associated with the corresponding cultural relic area coordinates. The platform's built-in personalized guided tour content generation module, combined with a multi-layered collaborative interactive response mechanism, receives real-time data such as the number of visitors and their locations transmitted by the guided tour robot, while also acquiring interest and preference information sent by visitors through their terminal devices. Based on the data integration results and interactive response data, the generation module uses a content matching algorithm to associate the defect assessment results, repair process data, and corresponding cultural background information. It adjusts the level of detail in the information presentation for different visitor groups, offering three content levels: basic, advanced, and professional. The basic version includes core cultural information and the main defect repair status; the advanced version adds explanations of the techniques; and the professional version supplements detailed defect assessment parameters and repair technical details. The generated personalized tour guide data is transmitted to the tour guide robot via a wireless communication module. The voice broadcast module on the tour guide robot outputs tour guide content synchronously with the display screen. The voice broadcast speed is controlled at 200 words per minute, and the display screen resolution is 1920×1080 pixels. At the same time, the tour guide robot moves according to the planned optimal path, and automatically triggers the corresponding tour guide content when passing through key cultural relic areas, realizing a deep integration of protection data and tour guide services, and improving the professionalism and pertinence of the tour.

[0026] The expression for the hierarchical identification and assessment model for cultural relic defects is as follows: ,in, This is the assessment value for the level of defects in cultural relics. For the first Weighting coefficients for class defects For the first Feature extraction values ​​of class defects, For the first in multi-layer robot collaboration The confidence level of data collection at each sensing node. This is a parameter for adjusting the sensitivity of defect identification. For the first Texture complexity parameters for each cultural relic area This is the threshold for defect determination. Spatial correlation factor, Grotto cultural relics OK Column pixel correlation matrix, For the first OK The gradient change value of the column pixels.

[0027] Specifically, the hierarchical identification and assessment model for cultural relic defects achieves precise quantification of defect levels through multi-dimensional parameter collaborative computation. The implementation process relies heavily on the distributed sensing data of a multi-layered robotic collaborative system and the model's built-in weight configuration mechanism. During model execution, weights are first assigned based on the impact of different defect types on the structural stability and historical value of the grotto artifacts, with structural defects receiving higher weights than surface decoration defects. For each distributed sensing node's collected data on artifact surface texture, spatial morphology, and environmental correlation, key feature information of the corresponding defects is extracted, and the feature extraction results are calculated. These results are then preliminarily weighted based on the sensing node's acquisition confidence level, which is determined by a combination of sensor accuracy and adaptation to the grotto environment. Simultaneously, the complexity of the artifact surface texture and a defect judgment benchmark value are introduced. Texture complexity is derived by analyzing pixel gradient changes on the artifact surface, while the judgment benchmark value is formulated based on the grotto artifact's protection level, historical age, and cultural value attributes. Defect features in adjacent areas are associated through spatial correlation parameters. The pixel correlation matrix and gradient change values ​​are extracted from the acquired images using professional image processing techniques, with the matrix dimensions consistent with the resolution of the acquired images. The model outputs a comprehensive defect level assessment result through the collaborative calculation of various parameters. This result is directly used for the priority ranking of subsequent defect processing and the robot task allocation process, enabling the defect processing work to take corresponding measures for defects of different severity levels, ensuring the pertinence and rationality of the grotto cultural relic protection work.

[0028] The expression for the style transfer model of mural fading restoration is as follows: ,in, The restored color values, The original color retention factor, For the original color data of cultural relics, For style transfer transformation matrix, To integrate historical style with weight, For the first The weights of each historical style sample, For the first Historical style reference data, This is the fading correction factor. This is a quantitative value for the degree of fading. This is an element-wise product operation.

[0029] The mural fading restoration style transfer model achieves precise restoration and style adaptation of faded areas through dynamic adjustment of multiple parameters. The implementation process requires comprehensive integration of real-time data collected by the robot and historical data stored on the intelligent management platform for tour content. During model operation, the original color retention ratio and historical style integration ratio are configured first. The original color retention ratio is set to maximize the approximation of the original color state of the artifact, while the historical style integration ratio is used to reasonably incorporate the stylistic characteristics of grotto murals from the same period as the object being restored. The style transfer conversion rules are constructed based on the color distribution patterns, painting techniques, and color matching logic of historical murals, forming a corresponding adaptation relationship with the number of image color channels. Each historical style sample is configured with exclusive associated parameters, determined based on the similarity between the sample and the current object being restored in terms of creation period, painting style, and pigment type. The fading correction parameters are dynamically adjusted based on the quantification results of the fading degree, which are obtained by comparing the color difference between the faded area and the surrounding normal area. The model uses an element-to-element fusion operation to process color data. First, it completes the basic color reconstruction of the faded area through the collaborative operation of the original color data and style transfer conversion rules. Then, it optimizes the style details by combining historical style reference data. Finally, it adjusts the overall restoration effect by adjusting the fading correction parameters to ensure that the color transition between the restored area and the surrounding normal area is natural and the style is consistent. The restored data is directly stored in the intelligent management platform for tour guide content, which not only provides complete cultural relic image data for tour guide display, but also provides accurate original data support for subsequent protection status tracking and secondary restoration.

[0030] The expression for the robot's dynamic obstacle avoidance algorithm is as follows: in, The optimal path, The set of candidate paths, The path length weight is used for weighting. For path length, To avoid obstacles and optimize weights, For the first An obstacle to the path Influence factors For robot size parameters, For environmental adaptability coefficient, For path The corresponding environmental gradient change, These are the path coordination parameters for multi-layered robot collaboration.

[0031] Specifically, the robot obstacle avoidance dynamic algorithm achieves optimal planning of robot operation and tour guide paths through multi-objective optimization calculations. The implementation process closely integrates the spatial structure data of the grotto's interior, real-time environmental obstacle information, and the multi-layered collaborative requirements of the robots. During algorithm execution, it first integrates basic spatial structure data such as the width, height, and turning angles of the grotto's internal passageways, and simultaneously receives obstacle information collected by the robot's onboard LiDAR and vision sensors, including key parameters such as the obstacle's specific location, size, and motion status. By setting an optimization objective that correlates path length with obstacle avoidance, and combining this with the robot's own dimensional parameters, it determines the required passage space, ensuring a safe distance between the robot and obstacles and the cultural relics themselves. Environmental adaptation parameters are introduced to adapt to complex environmental conditions such as changes in lighting and ground flatness within the grotto, and the calculation of environmental gradient changes along the path reflects the degree of environmental impact on movement. Simultaneously, path coordination parameters for multi-layered robot collaboration are incorporated to coordinate the movement routes of multiple robots, avoiding path conflicts between robots with different functions during operation or tour guide activities. The algorithm evaluates and calculates all candidate paths from multiple dimensions, and selects the optimal route in terms of path length, passage efficiency, obstacle avoidance safety, and collaborative coordination. The planning results include continuous coordinate guidance and movement control commands, which are directly transmitted to the robot's motion control module to guide the robot to accurately complete protection operations and guided movements, ensuring the smoothness and safety of the entire process.

[0032] The content generation model expression of the intelligent management platform for tour guide content is as follows: ,in, For personalized tour guide content data, For defect information weights, For the data of the cultural relic defect assessment results, A matrix relating the cultural background of grottoes. Weighting based on user preferences For the first Weighting coefficients for user preferences For the first User preference feature data A content distribution matrix for multi-layered robot collaboration. For environmental information fusion coefficient, This provides real-time data on the grotto environment.

[0033] Specifically, the intelligent management platform for guided tour content uses a content generation model that achieves precise content generation through deep fusion of multi-source data and a personalized adaptation mechanism. The implementation process integrates cultural relic protection data and cultural background information, fully considering the interactive response needs of multi-layered robot collaboration. During model execution, it first retrieves protection-related information such as the assessment results of cultural relic defects and mural restoration data, and matches them with pre-stored background data on the grottoes' historical dates, construction techniques, artistic styles, and cultural connotations to establish a mapping relationship between protection data and cultural information. Simultaneously, it receives real-time interactive data transmitted by the guided tour robots, including visitor locations, movement trajectories, and interest information fed back through terminal devices, adjusting the focus and level of detail in the guided tour content accordingly. By configuring the association weights between defect information and user preferences, it ensures that the guided tour content includes core protection information while catering to the information needs of different visitors. A content distribution rule for multi-layered robot collaboration is constructed, rationally allocating guided tour content transmission tasks based on the number, distribution, and operational status of participating guided tour robots to ensure synchronous content output. By integrating real-time environmental data from the grottoes, such as current visitor density and lighting conditions, the model dynamically adjusts the presentation of guided tour content, including adjusting voice broadcast volume and screen brightness. Through multi-parameter collaborative operation, the model generates personalized guided tour content, achieving real-time linkage and updates between conservation data and guided tour content. This allows visitors to simultaneously learn about the grotto culture and obtain information on the status of cultural relic conservation, enhancing the professionalism and relevance of the guided tours.

[0034] The task allocation model expression for multi-layered collaborative robots in the protection and guidance of grotto cultural relics is as follows: ,in, Assign a result matrix to the task. For the first Priority weights of tasks For the first The complexity parameter of the task. For the robot to the first The adaptability parameters of the task. For the first Communication bandwidth parameters of each distributed node For the first The communication delay gradient of each node, The coefficient for collaborative equilibrium. For the first The operational efficiency parameters of each robot The number of robots participating in the collaboration.

[0035] Specifically, a multi-layered collaborative task allocation model for grotto cultural relic protection and tour guidance is developed. This model achieves efficient task allocation through multi-dimensional task-robot compatibility assessment, relying on resource status and task requirement data from the multi-layered collaborative robot system. During model execution, various tasks to be assigned, such as relic defect identification, restoration work, and tour guidance services, are received, clarifying the specific operational requirements, technical difficulty, and priority level of each task. Simultaneously, parameters such as the functional configuration, operational capabilities, current working status, and remaining resources of all participating robots are collected to establish a robot compatibility assessment index system. Combining communication status parameters such as communication bandwidth and data transmission latency of distributed nodes, data transmission efficiency during task allocation is calculated to ensure the timeliness of task instructions and data interaction. A collaborative balancing parameter is introduced to coordinate the workload of multiple robots, avoiding situations where some robots are overloaded while others are idle. By calculating the compatibility between each task and each robot, and combining task priority and communication efficiency parameters for comprehensive evaluation, a multi-objective optimization approach is used to determine the optimal task allocation scheme. The allocation results are presented in matrix form, clearly defining the specific tasks, execution order, and cooperation requirements for each robot. The main control node for multi-layered robot cooperation issues task instructions to the corresponding robots, achieving precise task allocation and efficient execution, and ensuring the coordinated advancement of various tasks related to the protection and guidance of grotto cultural relics.

[0036] like Figure 2 As shown, step S3 includes the following sub-steps: S31, receiving the preliminary screening results output by the hierarchical identification and evaluation model of cultural relic defects through the main control node in the robot multi-layer collaborative system, extracting the spatial coordinates, texture feature differences, and grayscale distribution features of the defect candidate areas, and standardizing the extracted feature parameters according to the preset data format; S32, based on the standardized feature parameters, calling the task priority sorting module of the robot multi-layer collaborative system, constructing multi-dimensional sorting indicators according to the area, location importance, and urgency of repair of the defect area, and sequentially arranging the tasks that need to be performed for refined detection; S33, according to the sorting results, allocating the detection tasks to the corresponding detection robots through a dynamic task allocation mechanism, synchronously transmitting the feature parameters of the defect candidate areas and the accuracy requirements of the detection operation, ensuring that the detection robots perform defect detection operations according to the specified parameters.

[0037] Specifically, step S3, through multi-layered robot collaboration in task allocation and refined detection, achieves a quantitative assessment of defect levels. This is executed in the following steps: S31 The main control node of the multi-layered robot collaboration system receives the preliminary screening results output by the cultural relic defect level identification and assessment model, extracts the spatial coordinates, texture feature differences, and grayscale distribution features of the defect candidate areas, and standardizes the extracted 128-dimensional feature parameters according to a preset data format standard to ensure that the parameter format is consistent and compatible with subsequent processing procedures; S32 Based on the standardized feature parameters, the task priority ranking module of the multi-layered robot collaboration is called to construct a parameter ranking system that includes the defect area, location importance, and other parameters. A three-dimensional ranking index system based on the urgency of defects and repairs is used to sequentially arrange tasks requiring refined detection, forming a task queue of no more than 20 tasks. Based on the ranking results, S33 uses a dynamic task allocation mechanism to assign each detection task to 3-5 dedicated detection robots. Simultaneously, the characteristic parameters of the defect candidate area and the accuracy requirements of the detection operation are transmitted. The detection robots start detection functions such as ultrasonic detection, infrared thermal imaging, and microscopic imaging according to the specified parameters, collect the depth, range, and microscopic morphology data of the defect area, and complete the quantitative assessment of the defect level by combining the characteristic parameters output by the model, forming an assessment report including defect level, impact range, and treatment suggestions.

[0038] like Figure 3 As shown, step S4 includes the following sub-steps: S41, retrieve the historical color archive data and style feature parameters of the corresponding mural from the intelligent management platform for tour content, and combine them with the real-time image data of the faded area collected by the robot to construct a reference dataset for color restoration; S42, input the reference dataset into the style transfer model for mural fading restoration, set the style transfer intensity parameters and color fusion ratio parameters of the model, start the iterative calculation process of the model, and gradually optimize the color reconstruction effect of the faded area; S43, extract the restored image data output by the model, perform edge fusion processing with the surrounding area images collected by the robot in real time, eliminate the visual difference between the restored area and the original area, form complete mural image data, and transmit it to the intelligent management platform for tour content for storage.

[0039] Specifically, step S4 involves using a style transfer model for mural fading restoration to achieve precise restoration of the faded areas. This is executed in the following steps: S41 retrieves 50 sets of historical color archive data and 20 sets of style feature parameters for the corresponding mural from the historical database of the intelligent management platform for tour content. Combined with real-time image data of the faded areas collected by the robot, and after removing abnormal data, a reference dataset for color restoration is constructed. This dataset includes RGB three-color channel values, color distribution patterns, and brushstroke style characteristics. S42 inputs the reference dataset into the mural fading restoration style transfer model, sets the style transfer intensity parameters and color fusion ratio parameters of the model, and starts the process. The model's iterative calculation process involves 50 iterations, with parameters dynamically adjusted during each iteration to optimize the restoration effect and gradually reduce the difference between the restored and original colors. The S43 extracts the restored image data from the model's output and stitches it together with images of the surrounding area collected in real time by the robot using an edge fusion algorithm. By adjusting the color transition parameters of the fused area, the visual difference between the restored and original areas is eliminated, ensuring a natural and coherent transition. After forming complete mural image data, it is transmitted to the intelligent management platform for guided tour content via a 5G communication module for categorized storage, providing data support for subsequent guided tour display and protection tracking.

[0040] like Figure 4 As shown, step S5 includes the following sub-steps: S51, using the environmental perception sensors mounted on the robot to collect obstacle location data, spatial size data, and dynamic environmental change data inside the grotto, and transmitting the collected data to the input layer of the robot's obstacle area dynamic avoidance algorithm; S52, the algorithm classifies the input data into obstacle types and assesses risk levels, and, combined with the path planning constraints of the robot's multi-layered collaboration, determines the safety boundary parameters and passage priority of the path planning; S53, based on the classification and assessment results and constraint parameters, the algorithm's path search module generates multiple candidate travel paths, and calculates the passage efficiency parameters and obstacle avoidance success rate parameters for each path; S54, based on the calculation results, the optimal path is selected, and the path coordinate data and travel control parameters are transmitted to the robot's motion control module to guide the robot to perform operations and guide movement according to the planned path.

[0041] Specifically, step S5 involves achieving optimal path planning through a robot obstacle area dynamic avoidance algorithm, executed in the following steps: S51: The robot's environmental perception sensor group collects obstacle location data, spatial size data, and dynamic environmental change data within the grotto at a frequency of 30 frames per second. The lidar sensor has a detection range of 0.5-10 meters, and the visual sensor has a resolution of 1920×1080 pixels. All collected data is transmitted in real-time to the input layer of the robot obstacle area dynamic avoidance algorithm via a data transmission module. S52: The algorithm preprocesses the input data, classifying obstacles into two main categories: static obstacles and dynamic obstacles. Based on parameters such as obstacle size, location, and movement speed, it assesses the risk level, classifying obstacles into three risk levels. In step S53, based on the classification evaluation results and constraint parameters, the algorithm's path search module generates 15-20 candidate travel paths, and calculates the travel efficiency parameter and obstacle avoidance success rate parameter for each path. The travel efficiency parameter is calculated by combining the robot's moving speed and path length. In step S54, a multi-objective decision algorithm is used to select the optimal path based on the calculation results. The path coordinate data and travel control parameters are transmitted to the robot's motion control module. The control module adjusts the robot's travel direction and speed according to the path parameters, guiding the robot to perform tasks and guide movement according to the planned path, ensuring work efficiency while avoiding obstacles.

[0042] like Figure 5As shown, a multi-layered collaborative robot system for the protection and guidance of grotto cultural relics is presented. This system applies a multi-layered collaborative robot method for the protection and guidance of grotto cultural relics, including: a distributed data acquisition unit for multi-dimensional grotto cultural relics, an intelligent processing unit for hierarchical identification and assessment of cultural relic defects, a style transfer calculation unit for mural fading restoration, a robot obstacle avoidance path planning unit, an intelligent integration and distribution unit for guide content, and a multi-layered collaborative control unit. The distributed data acquisition unit collects data on cultural relics and the environment through distributed sensing nodes and transmits the data to the intelligent processing unit for hierarchical identification and assessment of cultural relic defects. The intelligent processing unit then identifies defects in the data. After hierarchical evaluation, defect data and location information are sent to the mural fading restoration style transfer calculation unit and the robot obstacle area dynamic avoidance path planning unit, respectively. After completing the fading restoration, the mural fading restoration style transfer calculation unit transmits the data to the intelligent integration and distribution unit of the tour guide content. The robot obstacle area dynamic avoidance path planning unit generates the optimal path and sends it to the robot multi-layer collaborative control unit. The intelligent integration and distribution unit of the tour guide content integrates various data to generate tour guide content and transmits it to the robot multi-layer collaborative control unit. The robot multi-layer collaborative control unit receives the path data and tour guide content, coordinates each robot to perform protection operations and tour guide tasks, and achieves efficient collaboration between the protection and tour guide of grotto cultural relics.

[0043] This paper presents a multi-layered collaborative robotic method and system for the protection and guidance of grotto cultural relics. Through a multi-layered collaborative architecture, it achieves intelligent management throughout the entire process of protection and guidance. Relying on the linkage of distributed sensing nodes and an intelligent management platform, it enables comprehensive collection and classified storage of data on cultural relics and the environment. Combined with a dedicated identification and evaluation model, it completes hierarchical judgment and quantitative assessment of defects, completely changing the problem of disconnect between defect identification and robotic operation in traditional technologies, making protection work more targeted. Simultaneously, by importing original color and historical style data through a style transfer model, it achieves precise restoration of faded murals, solving the shortcoming of insufficient coordination between traditional restoration effects and the original cultural relics, ensuring that the restored cultural relics maintain their original historical appearance.

[0044] In terms of robot collaboration and guided tour services, this system achieves optimal planning of conservation operations and guided tour routes through the synergy of a dynamic task allocation mechanism and an obstacle avoidance algorithm. This effectively overcomes the shortcomings of existing technologies, such as poor path planning flexibility and unreasonable task allocation, thus improving the efficiency and safety of robot operations. The intelligent management platform for guided tour content integrates defect assessment, repair data, and cultural background information, generating personalized guided tour content based on user preferences. This enables real-time linkage and updates between conservation data and guided tour content, solving the problems of traditional guided tours lacking specificity and real-time updates. The collaborative operation of various units and models not only enhances the precision of cultural relic conservation but also optimizes the personalized experience of guided tour services, comprehensively compensating for the deficiencies of existing technologies in the field of grotto cultural relic conservation and guided tours.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-layered collaborative robot method for the protection and guidance of grotto cultural relics, characterized in that, Includes the following steps: S1 collects surface texture features, spatial three-dimensional coordinates, and multi-dimensional data of the surrounding environment of the grotto cultural relics through distributed sensing nodes in the multi-layer collaborative robot system, and transmits the collected data to the intelligent management platform for tour guide content for data classification and storage. S2, invoke the cultural relic defect hierarchical identification and evaluation model to extract features and perform preliminary defect screening on the classified and stored data, and make a preliminary determination of the defect type through the multi-level identification rules built into the model; S3. Based on the preliminary judgment results, a task allocation model for multi-layered robot collaboration for the protection and guidance of grotto cultural relics is used to schedule a dedicated detection robot to perform refined defect detection. Combined with the defect feature parameters output by the model, a quantitative assessment of the mural fading defect level is completed. S4. For the mural fading defect areas confirmed by the assessment, the mural fading restoration style transfer model is activated, the original color data of cultural relics and historical style reference data collected by the robot are imported, and the color reconstruction and style adaptation of the fading areas are performed. S5, based on the internal spatial structure data of the grotto and real-time environmental obstacle information, runs a dynamic obstacle avoidance algorithm for the robot to plan the optimal path for robot protection operations and guided travel; S6 integrates defect assessment results, repair processing data, and grotto cultural background information through the content generation model of the intelligent management platform for tour guide content. Combined with the interactive response mechanism of multi-layered robot collaboration, it generates personalized tour guide data and transmits it to the tour guide robot to perform tour guide tasks.

2. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, In step S2, the expression for the hierarchical identification and evaluation model for cultural relic defects is: ,in, This is the assessment value for the level of defects in cultural relics. For the first Weighting coefficients for class defects For the first Feature extraction values ​​of class defects, For the first in multi-layer robot collaboration The confidence level of data collection at each sensing node. This is a parameter for adjusting the sensitivity of defect identification. For the first Texture complexity parameters for each cultural relic area This is the threshold for defect determination. Spatial correlation factor, Grotto cultural relics OK Column pixel correlation matrix, For the first OK The gradient change value of the column pixels.

3. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, In step S4, the expression for the style transfer model for mural fading restoration is: ,in, The restored color values, The original color retention factor, For the original color data of cultural relics, For style transfer transformation matrix, To integrate historical style with weight, For the first The weights of each historical style sample, For the first Historical style reference data, This is the fading correction factor. This is a quantitative value for the degree of fading. This is an element-wise product operation.

4. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, In step S5, the expression for the robot's dynamic obstacle avoidance algorithm is: in, The optimal path, The set of candidate paths, The path length weight is used for weighting. For path length, To avoid obstacles and optimize weights, For the first An obstacle to the path Influence factors For robot size parameters, For environmental adaptability coefficient, For path The corresponding environmental gradient change, These are the path coordination parameters for multi-layered robot collaboration.

5. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, In step S6, the content generation model expression of the intelligent management platform for tour guide content is: ,in, For personalized tour guide content data, For defect information weights, For the data of the cultural relic defect assessment results, A matrix relating the cultural background of grottoes. Weighting based on user preferences For the first Weighting coefficients for user preferences For the first User preference feature data A content distribution matrix for multi-layered robot collaboration. For environmental information fusion coefficient, This provides real-time data on the grotto environment.

6. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, The expression for the multi-layer collaborative task allocation model of robots for the protection and guidance of grotto cultural relics is as follows: ,in, Assign a result matrix to the task. For the first Priority weights of tasks For the first The complexity parameter of the task. For the robot to the first The adaptability parameters of the task. For the first Communication bandwidth parameters of each distributed node For the first The communication delay gradient of each node, The coefficient for collaborative equilibrium. For the first The operational efficiency parameters of each robot The number of robots participating in the collaboration.

7. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, S3 includes the following steps: S31, through the main control node in the robot multi-layer collaborative system, receives the preliminary screening results output by the cultural relic defect hierarchical identification and evaluation model, extracts the spatial coordinates, texture feature differences and gray-scale distribution features of the defect candidate area, and standardizes and organizes the extracted feature parameters according to the preset data format; S32, based on the standardized feature parameters, calls the task priority sorting module of the robot's multi-layer collaboration, constructs multi-dimensional sorting indicators according to the area, location importance and repair urgency of the defect area, and arranges the tasks that need to be performed for fine detection in order. S33. Based on the sorting results, the detection task is assigned to the corresponding detection robot through a dynamic task allocation mechanism. The feature parameters of the defect candidate area and the accuracy requirements of the detection operation are transmitted synchronously to ensure that the detection robot performs the defect detection operation according to the specified parameters.

8. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, S4 includes the following steps: S41. Retrieve historical color archive data and style feature parameters of the corresponding murals from the intelligent management platform for tour content, and combine them with real-time image data of the faded areas collected by the robot to construct a reference dataset for color restoration. S42, input the reference dataset into the style transfer model for mural fading restoration, set the style transfer intensity parameter and color fusion ratio parameter of the model, start the iterative calculation process of the model, and gradually optimize the color reconstruction effect of the faded area; S43 extracts the restored image data output by the model, performs edge fusion processing with the surrounding area images collected in real time by the robot, eliminates the visual difference between the restored area and the original area, forms complete mural image data, and transmits it to the intelligent management platform for tour guide content for storage.

9. The multi-layered collaborative robot method for the protection and guidance of grotto cultural relics according to claim 1, characterized in that, S5 includes the following steps: S51 collects obstacle location data, spatial size data, and dynamic environmental change data inside the grotto through the environmental perception sensors on the robot, and transmits the collected data to the input layer of the robot's obstacle area dynamic avoidance algorithm. S52, the algorithm classifies the obstacle type and assesses the risk level of the input data, and determines the safety boundary parameters and passage priority of the path planning by combining the path planning constraints of the robot's multi-layer collaboration. S53, based on the classification evaluation results and constraint parameters, uses the algorithm's path search module to generate multiple candidate travel paths, and calculates the passage efficiency parameter and obstacle avoidance success rate parameter for each path; S54: Based on the calculation results, the optimal path is selected, and the path coordinate data and travel control parameters are transmitted to the robot's motion control module to guide the robot to perform tasks and guide movement according to the planned path.

10. A multi-layered collaborative robot system for the protection and guidance of grotto cultural relics, characterized in that: This system is applied to the multi-layered collaborative robot method for the protection and guidance of grotto cultural relics as described in claim 1, comprising: a distributed data acquisition unit for multi-dimensional grotto cultural relics, an intelligent processing unit for hierarchical identification and assessment of cultural relic defects, a style transfer calculation unit for mural fading restoration, a robot obstacle avoidance path planning unit, an intelligent integration and distribution unit for guide content, and a multi-layered collaborative control unit for robots; wherein... The multi-dimensional distributed data acquisition unit for grotto artifacts collects data on artifacts and the environment through distributed sensing nodes, and transmits the data to the intelligent processing unit for hierarchical identification and assessment of artifact defects. After identifying and assessing the data, the intelligent processing unit sends the defect data and location information to the style transfer calculation unit for mural fading restoration and the robot obstacle avoidance path planning unit, respectively. After completing the fading restoration, the style transfer calculation unit transmits the data to the intelligent integration and distribution unit for tour guide content. The robot obstacle avoidance path planning unit generates the optimal path and sends it to the multi-layer collaborative control unit for robots. The intelligent integration and distribution unit for tour guide content integrates various data to generate tour guide content, which is then transmitted to the multi-layer collaborative control unit for robots. The multi-layer collaborative control unit for robots receives the path data and tour guide content, coordinates the robots to perform protection operations and tour guide tasks, and achieves efficient collaboration between the protection and tour guide functions of grotto artifacts.

Images