Method and system for generating a three-dimensional thermal map of a storage space based on image data processing

Abstract

This application relates to the field of equipment health management and maintenance, and discloses a method and system for generating three-dimensional storage space heatmaps based on image data processing. The method includes the following steps: S1, constructing a three-dimensional graph structure corresponding to the physical arrangement of the storage device, wherein the three-dimensional graph structure includes nodes and the adjacency relationships between nodes; S2, obtaining the wear values ​​of the nodes based on image acquisition data, and dividing the set of observed nodes and the set of unobserved nodes based on the wear values; S3, establishing a spatial function optimization model with fixed boundary conditions based on the wear values ​​in the set of observed nodes and the node connection relationships in the three-dimensional graph structure. This invention, by combining a spatial function optimization model of a three-dimensional graph structure with a numerical solution method, accurately predicts the wear status of various components of the equipment, improving the accuracy, real-time performance, and computational efficiency of equipment health management.

Description

Technical Field

[0001] This invention relates to the field of equipment health management and maintenance, specifically to a method and system for generating three-dimensional storage space heatmaps based on image data processing. Background Technology

[0002] In the daily operation of modern equipment, wear and tear on various mechanical and electronic devices is inevitable with prolonged use. Especially under high-load, high-intensity environments, wear and tear can lead to performance degradation and even malfunctions. Therefore, timely and accurate assessment of equipment wear status can effectively predict its service life and allow for proactive maintenance measures, preventing unexpected failures. This not only improves equipment efficiency and extends its lifespan but also reduces maintenance costs and avoids losses caused by downtime.

[0003] In existing technologies, equipment wear monitoring methods typically employ localized data acquisition or prediction based on simple mathematical models. By monitoring the state of certain components in real time using sensors and utilizing statistical models to predict wear, existing technologies can provide some degree of wear information for equipment. For example, some data analysis methods based on vibration, temperature, or pressure sensors can detect abnormal states of certain equipment components and provide early warnings of potential failure risks. These methods mainly rely on localized information and specific models, possessing a certain degree of practicality and the ability to detect problems early, thus reducing the risk of equipment failure.

[0004] However, existing technologies have some shortcomings, especially when dealing with complex equipment systems. These methods typically focus only on localized wear information and cannot comprehensively assess the wear status of the entire equipment. Sensor-based data can only reflect the condition of certain specific components, often lacking effective predictions for wear in other areas. Furthermore, existing prediction models are usually based on relatively simple assumptions, making it difficult to accurately describe the complex interrelationships between equipment components. In addition, traditional methods are computationally inefficient; processing large-scale data involves cumbersome and time-consuming calculations, making it difficult to meet the needs of real-time monitoring. These shortcomings significantly limit the application of traditional technologies in global equipment health management and accurate prediction. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for generating three-dimensional storage space heatmaps based on image data processing, which solves the problems of insufficient accuracy in equipment wear prediction, inability of local information to fully reflect equipment health status, and low computational efficiency in existing technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for generating a three-dimensional storage space heatmap based on image data processing, comprising the following steps:

[0007] S1. Construct a three-dimensional graph structure corresponding to the physical arrangement of storage devices, wherein the three-dimensional graph structure includes nodes and the adjacency relationships between nodes;

[0008] S2. Obtain the wear value of the node based on the image acquisition data, and divide the observed node set and the unobserved node set based on the wear value;

[0009] S3. Based on the wear values ​​in the observation node set and the node connection relationship in the three-dimensional graph structure, establish a spatial function optimization model with fixed boundary conditions. The spatial function optimization model uses the measure of the wear value difference between adjacent nodes as the objective expression.

[0010] S4. Based on the space function optimization model, the predicted wear value corresponding to the set of unobserved nodes is obtained by numerical solution.

[0011] S5. Normalize all node wear values ​​in the observed node set and the unobserved node set, and generate a three-dimensional heat map based on the normalization result.

[0012] Preferably, in step S1, constructing the three-dimensional graph structure corresponding to the physical arrangement of the storage devices includes:

[0013] Based on the spatial distribution image of the storage devices acquired by the image acquisition device, the physical location information of each storage unit in the three-dimensional coordinate system is extracted;

[0014] Each storage unit is mapped to a node in the graph structure, and the adjacency relationship between nodes is determined based on the relative positional relationship between storage units in the physical space.

[0015] Preferably, in step S2, dividing the set of observed nodes and the set of unobserved nodes based on the wear value includes:

[0016] Extract currently available partial node wear value data from image data; the wear value includes the number of writes to the storage unit, error statistics, and percentage of lifetime usage.

[0017] Nodes with clear and reliable wear value data are marked as observation nodes, forming an observation node set;

[0018] The remaining nodes that failed to collect wear values ​​in the current period are marked as unobserved nodes, forming a set of unobserved nodes.

[0019] Preferably, the extraction of currently available partial node wear value data from image data includes:

[0020] Extract the location information of each storage unit from images acquired by image data sources or image acquisition devices, and determine the precise coordinates of the storage units through image analysis technology;

[0021] Using the interface of the backend storage system, for each storage unit, the wear value of each node is calculated and obtained based on historical records or real-time collected working parameters.

[0022] Preferably, in step S3, establishing a spatial function optimization model with fixed boundary conditions includes:

[0023] The wear values ​​in the set of observed nodes are used as the boundary conditions of the model, keeping the values ​​unchanged during the optimization process;

[0024] Based on the adjacency relationship between nodes in the graph structure, a variational energy expression of the wear function on the graph structure is defined, which is used to measure the difference in wear values ​​between adjacent nodes;

[0025] The optimization objective is to minimize the total energy across the entire graph structure, where the total energy is defined as:

[0026] ;

[0027] In the formula, Represents a node With nodes Connection weights between them; Represents a node The wear value; The total energy of the three-dimensional graph structure; Represents a node The wear value; For the summation symbol node.

[0028] Preferably, the variational energy expression of the wear function on the graph structure includes:

[0029] A local energy function is defined using the wear value of each node in the three-dimensional graph structure. The local energy function is used to represent the wear difference of each node relative to its neighboring nodes.

[0030] Construct a variational energy expression to quantify the energy difference between adjacent nodes in the graph;

[0031] The energy expression for the entire graph is formed by accumulating the local energy differences of each node to create a global energy function, and the optimization objective is to minimize this global energy function.

[0032] Preferably, in step S4, obtaining the predicted wear value corresponding to the set of unobserved nodes through numerical solution includes:

[0033] Based on the constructed spatial function optimization model, the linear system expression form corresponding to the graph Laplacian matrix is ​​extracted;

[0034] The linear system described above is divided into matrix blocks to separate the subsystems related to unobserved nodes;

[0035] The subsystem is solved using a sparse matrix solving algorithm, which includes the conjugate gradient method and sparse LU decomposition, to efficiently obtain the predicted wear values ​​of all unobserved nodes.

[0036] The numerical solution process is a linear, stable, and deterministic solution process.

[0037] Preferably, the linear system representation corresponding to the extracted graph Laplacian matrix includes:

[0038] In the constructed spatial function optimization model, the three-dimensional graph structure is discretized, and the corresponding Laplacian matrix is ​​extracted. The Laplacian matrix is ​​used to describe the relationship between nodes and the flow of information.

[0039] Each element of the Laplace matrix represents the connection strength between nodes and their relative relationships, and reflects the energy transfer process of each node.

[0040] Preferably, in step S5, the normalization process includes:

[0041] Find the minimum wear value for all nodes. With the maximum value ;

[0042] The wear value of each node Mapped to color The mapping relationship is as follows:

[0043] ;

[0044] In the formula, The normalized thermal color value; Represents a node The wear value; This represents the minimum wear value. Maximum wear value.

[0045] This invention also provides a three-dimensional storage space heatmap generation system based on image data processing, comprising:

[0046] The graph structure construction module constructs a three-dimensional graph structure based on the physical location of the storage device. Each storage unit corresponds to a node in the graph structure, and the adjacency relationship between nodes is determined based on their relative positions in physical space.

[0047] The data acquisition and node partitioning module is used to collect wear values ​​of some storage units from image data, and based on the collected wear values, divide the nodes into a set of observed nodes and a set of unobserved nodes.

[0048] The optimization model construction and calculation module establishes a harmonic energy optimization model with boundary conditions based on the wear values ​​and graph structure in the observation node set.

[0049] The heatmap generation and visualization module is used to normalize the wear values ​​of all nodes and map them into color values, thereby generating a three-dimensional heatmap.

[0050] The display and control module is used to display the generated 3D heat map and provide an interactive interface for users. Based on the set conditions, the system will generate real-time status prompts and maintenance suggestions.

[0051] This invention provides a method and system for generating three-dimensional storage space heatmaps based on image data processing. It has the following beneficial effects:

[0052] 1. This invention employs a technical solution based on a three-dimensional graph structure and a spatial function optimization model, achieving higher accuracy in wear value prediction. Compared to existing prediction methods based on local or simplified models, this invention, by comprehensively considering wear data from both observed and unobserved nodes, can more fully reflect the actual wear condition of the equipment. It solves the problem of insufficient prediction accuracy in traditional methods, making wear prediction more accurate and reliable.

[0053] 2. This invention introduces a technical solution that combines unified normalization processing with three-dimensional heat map generation, enhancing decision support capabilities for equipment maintenance. By generating intuitive heat maps, maintenance personnel can quickly identify severely worn areas of equipment and perform timely maintenance. Compared to existing technologies that rely solely on two-dimensional data or local analysis, this invention provides a more spatially aware global perspective, effectively solving the problem that traditional methods cannot comprehensively assess the overall condition of equipment.

[0054] 3. This invention successfully improves the efficiency and stability of the system when processing large-scale data by combining numerical solutions with optimization models. Compared with existing technologies that rely on manual data processing or inefficient algorithms, this invention can efficiently perform calculations on large-scale node sets, quickly predict and normalize wear values, avoid performance bottlenecks in the processing, and improve the system's response speed and data processing capabilities.

[0055] 4. This invention employs a scalable spatial function optimization model and numerical solution method, exhibiting strong extensibility and adaptability to monitoring needs in different equipment or working environments. Compared to existing solutions that struggle to adapt to diverse scenarios, this invention flexibly addresses wear prediction and monitoring for various equipment and operating conditions, resolving the limitation of applicability in existing technologies and making long-term equipment prediction and health monitoring more feasible. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0057] Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Please see the appendix Figure 1 This invention provides a method for generating a three-dimensional storage space heatmap based on image data processing, comprising the following steps:

[0060] S1. Construct a three-dimensional graph structure corresponding to the physical arrangement of storage devices, wherein the three-dimensional graph structure includes nodes and the adjacency relationships between nodes;

[0061] The 3D graph structure is the foundation of the entire system, providing a crucial framework and data support for subsequent wear value prediction, optimization calculations, and heatmap generation. Through this graph structure, the system can accurately represent the spatial relationships between storage cells, and use this as the basis for processing and calculating wear values. Specifically, the 3D graph structure consists of nodes and the adjacency relationships between them; nodes correspond to storage cells, and adjacency relationships reflect the spatial proximity of the storage cells.

[0062] The steps for constructing the 3D graph structure include: First, acquiring spatial distribution images of the storage devices using an image acquisition device, and extracting the physical location information of each storage unit; then, mapping each storage unit to a node in the graph structure, and determining the adjacency relationships between nodes based on the physical relative positions of the storage units. These processes not only provide a foundation for subsequent data acquisition and node partitioning, but also lay the groundwork for the construction of the optimization model and the generation of heatmaps.

[0063] In one possible implementation, the image acquisition device captures the spatial distribution of the storage devices using high-precision scanning or image processing equipment. This image data includes the precise location coordinates of the storage cells and their relative arrangement in physical space. This information is crucial for constructing an accurate 3D graph structure. To ensure that the graph structure accurately reflects the physical arrangement of the storage devices, a 3D coordinate system is typically used to calibrate these locations, and image analysis techniques are used to further correct and extract the actual location of each storage cell.

[0064] Specifically, the construction of nodes depends on the physical location of the storage cells. Each storage cell has a defined position in a three-dimensional coordinate system, and these storage cells are arranged according to a physical layout. Therefore, the system maps each storage cell to a node in a three-dimensional graph structure. The connections between nodes are defined based on the relative positions of these storage cells in space. Typically, adjacent storage cells are connected by an edge in the graph, reflecting their spatial proximity.

[0065] In some embodiments, the definition of adjacency relationships may depend on the geometric distance or physical proximity of storage cells. Two storage cells are defined as adjacent nodes in the graph structure if they are close to each other or if there is a direct physical connection between them. In practice, the definition of adjacency relationships can be adjusted based on the actual physical structure between storage cells. For example, adjacency relationships between different levels can be defined according to the device's storage hierarchy, thereby optimizing the construction of the graph structure.

[0066] Furthermore, the adjacency relationships in the graph structure can be optimized through further calculations. In some embodiments, the weights between nodes can be defined by calculating the physical distance between storage units. The magnitude of the weights can affect the relative relationships between nodes in the graph structure, thus playing an important role in subsequent energy optimization models. This method of weight calculation provides more precise control during the optimization process, making the model more adaptable to the actual physical conditions of the devices.

[0067] Building upon this foundation, the constructed 3D graph structure includes not only nodes but also the connection information between them. This connection information plays a crucial role in subsequent steps such as node partitioning, optimization calculations, and heatmap generation. Specifically, each node in the graph represents a storage unit, while the edges between nodes represent the relative spatial relationships of these storage units. In this way, the graph structure can clearly present the spatial layout of storage devices and provide strong data support for wear and tear prediction.

[0068] S2. Obtain the wear value of the node based on the image acquisition data, and divide the observed node set and the unobserved node set based on the wear value;

[0069] Acquiring image data and calculating node wear values ​​are crucial steps. The aforementioned steps in constructing the 3D graph structure have laid the foundation for subsequent wear value calculations. Each storage cell corresponds to a node in the 3D graph structure, and the wear values ​​of these nodes are key data for assessing the health status of the storage device. In this embodiment, image acquisition data is used not only to obtain the physical layout of the storage device but also to extract wear information for each storage cell. Based on these wear values, nodes are divided into a set of observed nodes and a set of unobserved nodes. This process provides the necessary input for subsequent optimization calculations and heatmap generation.

[0070] Specifically, the image acquisition device is responsible for acquiring image data from the storage device and extracting wear data for each storage unit using image analysis technology. This wear data can include write counts, error statistics, and percentage of lifetime usage. This data provides the core basis for subsequent node partitioning and optimization. Node partitioning lays the foundation for subsequent optimization calculations because, in the optimization model, the set of observed nodes will be used as known boundary conditions as input, while unobserved nodes need to be predicted by the optimization model.

[0071] In this embodiment, the image acquisition device may include a high-resolution scanner, a vision sensor, etc., for acquiring detailed image data of the storage device. The acquired image data includes the location coordinates of the storage unit, as well as the current state and operating parameters of the unit (e.g., write count, error statistics, lifetime usage percentage). From this data, the system can extract the wear value of each node and, based on the availability of the wear value, divide the nodes into a set of observed nodes and a set of unobserved nodes.

[0072] Typically, image acquisition devices obtain data containing wear data, but wear data for some storage units may not be acquired due to various reasons (such as scanning range, equipment malfunction, etc.). Therefore, in one possible implementation, the system verifies the validity of the wear data for each storage unit and assigns valid wear values ​​to the observed node set. Nodes for which wear values ​​could not be acquired are classified into the unobserved node set. Specifically, the wear values ​​of unobserved nodes are estimated and predicted using a subsequent optimization model.

[0073] As an alternative, during node partitioning, the system might use a set of preset rules to determine node validity. For example, if the wear value of a storage cell is below a set threshold, the data is considered invalid, and the node is automatically classified as an unobserved node. Another possibility is that if data is lost or unavailable, the node is directly marked as an unobserved node. This approach ensures that only verified wear value data enters the observation node set, guaranteeing data accuracy.

[0074] Specifically, the construction of the observation node set is based on wear data extracted from image data. The system obtains the location information of each storage unit from the image data source or image acquisition device, and determines the precise coordinates of the storage units through image analysis technology. These precise coordinates, along with historical or real-time acquired operating parameters, are used to calculate the wear value of each storage unit. The wear value can be calculated based on the following formula:

[0075] ;

[0076] In the formula, Represents a node The wear value; For storage units The number of writes; For storage units Error statistics, For storage units The percentage of lifespan used. By weighting these data, the actual wear value of each node can be obtained.

[0077] Once the wear values ​​of each node are obtained, the system divides the nodes into an observed node set and an unobserved node set based on this data. In some embodiments, the observed node set includes nodes for which reliable wear data is available, and this data is directly used for subsequent optimization calculations. The unobserved node set includes nodes for which wear values ​​were not obtained in the current period, and these nodes require prediction using a spatial optimization model.

[0078] S3. Based on the wear values ​​in the observation node set and the node connection relationship in the three-dimensional graph structure, establish a spatial function optimization model with fixed boundary conditions. The spatial function optimization model uses the measure of the wear value difference between adjacent nodes as the objective expression.

[0079] By incorporating wear values ​​from the observed node set into a 3D graph structure and combining this with the connectivity between nodes, the system can establish a spatial function optimization model with fixed boundary conditions. The goal of this optimization model is to minimize the difference in wear values ​​between adjacent nodes, thereby calculating the wear values ​​of unobserved nodes. This model provides an important means of optimizing the wear distribution of the entire storage device. Through the optimization process of this model, the system can effectively predict the wear state of unobserved nodes, providing accurate data support for heatmap generation and equipment maintenance.

[0080] Using spatial optimization methods, the wear values ​​of all nodes can be further calculated based on the wear values ​​in the observed node set and the node connectivity in the 3D graph structure. This process relies on the establishment of an optimization model, specifically by establishing a spatial function optimization model with fixed boundary conditions and using the measure of the difference in wear values ​​between adjacent nodes as the objective to solve for the wear distribution of the entire graph structure.

[0081] In this embodiment, when constructing the optimization model, the system first determines the wear values ​​in the set of observed nodes. These wear values ​​serve as known boundary conditions in the spatial function optimization model. The wear value of each observed node is fixed as the initial condition of the model. Through these known boundary conditions, the goal of the optimization model is to ensure that the wear values ​​of unobserved nodes can be inferred throughout the entire graph structure while minimizing the wear differences with adjacent nodes.

[0082] Generally, the optimization objective can be expressed as a variational energy function, where the energy level is proportional to the difference in wear values ​​between nodes. The total energy is defined as:

[0083] ;

[0084] In the formula, Represents a node With nodes Connection weights between them; Represents a node The wear value; The total energy of the three-dimensional graph structure; Represents a node The wear value; For the summation symbol node.

[0085] Alternatively, when defining the objective function, the connection weights can be used. The weight can be determined based on the physical distance or connection strength between storage units. For example, if two nodes are very close in physical space, a higher weight value can represent a stronger relationship between them. Conversely, the connection weight between nodes that are far apart can be set to a lower value. The formula for calculating the weight can be similar to the following:

[0086] ;

[0087] in, Represents a node and nodes The physical distance between them; Represents a node With nodes The connection weights between nodes are considered. In this way, the optimization process prioritizes the differences in wear values ​​among neighboring nodes, making the wear values ​​between adjacent nodes more balanced, thereby calculating reasonable wear values ​​for unobserved nodes.

[0088] Specifically, this optimization model not only considers the wear differences between adjacent nodes, but also further optimizes the wear distribution of the entire graph by adjusting the weights of each node in the graph and its adjacent nodes. Each unobserved node in the optimization process receives a predicted value in the model's calculations, and these predicted values ​​are kept in balance with the wear values ​​of its adjacent nodes to ensure that the wear distribution on the entire 3D graph structure is reasonable and consistent with reality.

[0089] In one possible implementation, the model's boundary conditions are set based on the acquired wear values. Each node in the observed node set has a fixed wear value, which serves as input to the optimization model. The wear values ​​of unobserved nodes need to be solved numerically. During this optimization process, the system iteratively calculates and solves a system of linear equations to continuously update the wear value of each node until an optimal stable state is reached.

[0090] Alternatively, the optimization model can be solved using sparse matrix solving algorithms. For example, efficient matrix solving methods such as the conjugate gradient method or sparse LU decomposition can maintain high accuracy while ensuring computational speed. These numerical methods can quickly process large-scale graph structures, improving the overall computational efficiency of the system.

[0091] S4. Based on the space function optimization model, the predicted wear value corresponding to the set of unobserved nodes is obtained by numerical solution.

[0092] The aforementioned spatial function optimization model provides a theoretical basis for solving the wear values ​​of unobserved nodes. This optimization model, by considering the wear values ​​of observed nodes and the connectivity between nodes, aims to calculate the wear state of all nodes by minimizing the wear differences between adjacent nodes. Specifically, after establishing the model, the next step is to predict the wear values ​​of unobserved nodes using numerical methods. Through this process, the system can reasonably predict the wear of unobserved nodes based on known observation data, thus forming a complete wear distribution of the storage device.

[0093] In this embodiment, a numerical solution algorithm based on the aforementioned spatial function optimization model is used to calculate the predicted wear values ​​for the corresponding set of unobserved nodes. First, the system transforms the spatial function optimization model into a system of linear equations, which describes the wear value of each node in the graph and the relationships between them. Specifically, by constructing the graph Laplacian matrix, the system can transform the nodes and connections in the graph structure into a mathematical model and solve for the wear values ​​of the unobserved nodes using numerical methods.

[0094] Generally, based on the spatial optimization model, the system first constructs a graph Laplacian matrix to describe the relationships between nodes. (Graph Laplacian matrix) Defined as a The matrix, where This represents the number of nodes in the graph. Each element of this matrix... Represents a node and nodes The strength of the relationship between nodes is usually determined by the weights between them. For example, if two nodes are close to each other, their connection weight is larger.

[0095] Specifically, the Thulaplacian matrix The construction can be carried out in the following ways:

[0096] ;

[0097] In the formula, The degree matrix is ​​a diagonal matrix that represents the degree of each node. This is the weight matrix, which describes the connection strength between nodes. It describes the relationships between nodes in the graph, and for any given node... Its adjacent nodes The wear value can be calculated by solving a system of linear equations.

[0098] In one possible implementation, the system constructs a linear system equation by combining the graph Laplacian matrix with the wear value vector. The goal of the system is to determine the wear values ​​of all unobserved nodes by solving this equation. Assume the wear values ​​of the known observed nodes are... (i.e., the wear value of the observed node), while the wear value of the unknown, unobserved node is... (i.e., the wear value to be predicted), then the linear equation of the system can be expressed as:

[0099] ;

[0100] In the formula, It is a vector containing the wear values ​​of all nodes; The boundary condition vector is known. The graph Laplace matrix includes the wear values ​​of the observed nodes and other known information. The system obtains the wear values ​​of all nodes by solving this equation.

[0101] Alternatively, numerical solution methods can use the conjugate gradient method or sparse matrix factorization methods (such as sparse LU decomposition). These methods are suitable for large-scale graph structures and can efficiently solve systems of linear equations. In particular, the conjugate gradient method, through its iterative solution approach, not only improves the solution speed but also reduces the consumption of storage and computational resources, making it suitable for large-scale data processing.

[0102] Specifically, in this solution process, the system will iteratively adjust the wear value of each unobserved node until the solution converges. The convergence criteria are usually that the change in the node wear value is lower than a set threshold, or the number of iterations reaches a preset maximum value.

[0103] Generally, through this numerical solution process, the wear values ​​of unobserved nodes can be reasonably predicted. These predicted wear values ​​are used to generate a complete three-dimensional heat map, further providing a basis for monitoring the health status of storage devices and making maintenance decisions.

[0104] In another possible implementation, the system can also dynamically adjust the weights in the optimization process based on the predicted wear value. For example, when the predicted wear value of a certain storage cell is large, the system may increase the connection weights between that node and its neighboring nodes, prompting the optimization process to pay more attention to these nodes, thereby improving the accuracy in subsequent calculations.

[0105] S5. Normalize all node wear values ​​in the observed node set and the unobserved node set, and generate a three-dimensional heat map based on the normalization result.

[0106] Through this normalization operation, the system can adjust the wear values ​​of different nodes to a uniform standard range, thus facilitating the generation of easily understandable 3D heatmaps. Based on the normalization, the system uses the normalization results to generate corresponding heatmaps to display the wear status of different nodes.

[0107] In this embodiment, the wear values ​​of all nodes in both the observed node set and the unobserved node set are first normalized. The purpose of normalization is to map the wear values ​​of different nodes to a uniform range, so as to more clearly display the wear status of each part of the device. Common normalization methods include linearly mapping the wear values ​​to the [0,1] interval or using Z-score normalization.

[0108] Normalization can generally be performed using the following formula:

[0109] ;

[0110] In the formula, The normalized thermal color value; Represents a node The wear value; This represents the minimum wear value. Maximum wear value.

[0111] Specifically, the system first extracts the wear values ​​of all nodes from the numerical solution results of the aforementioned space function optimization model. These values ​​include the wear states of both observed and unobserved nodes. Next, the system calculates the minimum and maximum wear values ​​for all nodes and normalizes these values ​​according to the aforementioned formula. The normalized values ​​are distributed within the interval [0,1], with nodes showing larger values ​​indicating more severe wear and nodes showing smaller values ​​indicating less severe wear.

[0112] Alternatively, other standardization methods can be used in the normalization process. For example, the Z-score standardization method subtracts the mean from the wear value of each node and divides it by the standard deviation, as shown in the formula:

[0113] ;

[0114] In the formula, This represents the average wear value of all nodes; This represents the standard deviation of wear values ​​for all nodes; This is the normalized wear value; These are the original wear values. After Z-score standardization, the wear values ​​of all nodes will be normalized with the mean as the center and the standard deviation as the scale, which has certain advantages for wear values ​​of different scales.

[0115] In one possible implementation, after normalization, the system generates a 3D heatmap based on the normalized wear values. During this process, the system maps the normalized wear values ​​to color values ​​within the heatmap to display the wear status of different areas. Common heatmap color mappings include gradients from blue (low wear) to red (high wear), which visually reflects the wear condition of the equipment.

[0116] The process of generating a 3D heatmap can be achieved through the following steps:

[0117] Data preparation: The system associates the normalized wear values ​​of all nodes with the three-dimensional spatial location of the nodes to form the spatial coordinates and wear value of each node.

[0118] Color Mapping: The system selects a suitable color mapping function based on the normalized wear value of each node. For example, the following linear mapping relationship can be used to map the normalized wear value... Mapped to RGB color space:

[0119] ;

[0120] In the formula, The color of the node; This is the normalized wear value; This represents the maximum value of the normalized wear value, used to map it to the corresponding color range. According to... The colors vary, gradually changing from blue to red, to indicate areas of low to high wear.

[0121] 3D Heatmap Generation: Based on the spatial coordinates and color values ​​of the nodes, the system generates a 3D heatmap to display the wear and tear of various parts of the equipment. The heatmap clearly identifies areas with more severe wear and areas with less wear through different colors.

[0122] As an extension, this 3D heatmap can be linked with other monitoring systems or maintenance tools to achieve real-time monitoring of equipment status. By updating normalized wear values ​​in real time and generating corresponding heatmaps, the system can help maintenance personnel promptly identify potential equipment problems and take appropriate maintenance measures.

[0123] In another possible implementation, the normalized wear values ​​can also be used as data input for training machine learning models. For example, normalized wear values ​​can be used as features to train a predictive model to predict the future wear trend of the equipment. In this way, the system can not only monitor the equipment status in real time, but also predict the future performance of the equipment, providing longer-term support for equipment management.

[0124] Through this normalization and heatmap generation process, the system can present equipment wear in an intuitive and scientific way, improving the efficiency of equipment management and maintenance.

[0125] The image data processing-based three-dimensional storage space heatmap generation system described below can be referred to in correspondence with the image data processing-based three-dimensional storage space heatmap generation method described above.

[0126] Please see the appendix Figure 2 The present invention also provides a three-dimensional storage space heatmap generation system based on image data processing, comprising:

[0127] The graph structure construction module constructs a three-dimensional graph structure based on the physical location of the storage device. Each storage unit corresponds to a node in the graph structure, and the adjacency relationship between nodes is determined based on their relative positions in physical space.

[0128] The data acquisition and node partitioning module is used to collect wear values ​​of some storage units from image data, and based on the collected wear values, divide the nodes into a set of observed nodes and a set of unobserved nodes.

[0129] The optimization model construction and calculation module establishes a harmonic energy optimization model with boundary conditions based on the wear values ​​and graph structure in the observation node set.

[0130] The heatmap generation and visualization module is used to normalize the wear values ​​of all nodes and map them into color values, thereby generating a three-dimensional heatmap.

[0131] The display and control module is used to display the generated 3D heat map and provide an interactive interface for users. Based on the set conditions, the system will generate real-time status prompts and maintenance suggestions.

[0132] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0133] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various 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 method for generating a three-dimensional storage space heatmap based on image data processing, characterized in that, Includes the following steps: S1. Construct a three-dimensional graph structure corresponding to the physical arrangement of storage devices, wherein the three-dimensional graph structure includes nodes and the adjacency relationships between nodes; S2. Obtain the wear value of the node based on the image acquisition data, and divide the observed node set and the unobserved node set based on the wear value; S3. Based on the wear values ​​in the observation node set and the node connection relationship in the three-dimensional graph structure, establish a spatial function optimization model with fixed boundary conditions. The spatial function optimization model uses the measure of the wear value difference between adjacent nodes as the objective expression. S4. Based on the space function optimization model, the predicted wear value corresponding to the set of unobserved nodes is obtained by numerical solution. S5. Normalize all node wear values ​​in the observed node set and the unobserved node set, and generate a three-dimensional heat map based on the normalization result.

2. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 1, characterized in that, In step S1, constructing the three-dimensional graph structure corresponding to the physical arrangement of the storage devices includes: Based on the spatial distribution image of the storage devices acquired by the image acquisition device, the physical location information of each storage unit in the three-dimensional coordinate system is extracted; Each storage unit is mapped to a node in the graph structure, and the adjacency relationship between nodes is determined based on the relative positional relationship between storage units in the physical space.

3. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 1, characterized in that, In step S2, dividing the set of observed nodes and the set of unobserved nodes based on the wear value includes: Extract currently available partial node wear value data from image data; the wear value includes the number of writes to the storage unit, error statistics, and percentage of lifetime usage. Nodes with clear and reliable wear value data are marked as observation nodes, forming an observation node set; The remaining nodes that failed to collect wear values ​​in the current period are marked as unobserved nodes, forming a set of unobserved nodes.

4. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 3, characterized in that, The extraction of currently available partial node wear value data from image data includes: Extract the location information of each storage unit from images acquired by image data sources or image acquisition devices, and determine the precise coordinates of the storage units through image analysis technology; Using the interface of the backend storage system, for each storage unit, the wear value of each node is calculated and obtained based on historical records or real-time collected working parameters.

5. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 1, characterized in that, In step S3, establishing a spatial function optimization model with fixed boundary conditions includes: The wear values ​​in the set of observed nodes are used as the boundary conditions of the model, keeping the values ​​unchanged during the optimization process; Based on the adjacency relationship between nodes in the graph structure, a variational energy expression of the wear function on the graph structure is defined, which is used to measure the difference in wear values ​​between adjacent nodes; The optimization objective is to minimize the total energy across the entire graph structure, where the total energy is defined as: ； In the formula, Represents a node With nodes Connection weights between them; Represents a node The wear value; The total energy of the three-dimensional graph structure; Represents a node The wear value; For the summation symbol node.

6. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 5, characterized in that, The variational energy expression of the wear function on the graph structure includes: A local energy function is defined using the wear value of each node in the three-dimensional graph structure. The local energy function is used to represent the wear difference of each node relative to its neighboring nodes. Construct a variational energy expression to quantify the energy difference between adjacent nodes in the graph; The energy expression for the entire graph is formed by accumulating the local energy differences of each node to create a global energy function, and the optimization objective is to minimize this global energy function.

7. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 1, characterized in that, In step S4, obtaining the predicted wear value corresponding to the set of unobserved nodes through numerical solution includes: Based on the constructed spatial function optimization model, the linear system expression form corresponding to the graph Laplacian matrix is ​​extracted; The linear system described above is divided into matrix blocks to separate the subsystems related to unobserved nodes; The subsystem is solved using a sparse matrix solving algorithm, which includes the conjugate gradient method and sparse LU decomposition, to efficiently obtain the predicted wear values ​​of all unobserved nodes. The numerical solution process is a linear, stable, and deterministic solution process.

8. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 7, characterized in that, The linear system representation corresponding to the extracted graph Laplacian matrix includes: In the constructed spatial function optimization model, the three-dimensional graph structure is discretized, and the corresponding Laplacian matrix is ​​extracted. The Laplacian matrix is ​​used to describe the relationship between nodes and the flow of information. Each element of the Laplace matrix represents the connection strength between nodes and their relative relationships, and reflects the energy transfer process of each node.

9. The method for generating a three-dimensional storage space heatmap based on image data processing according to claim 1, characterized in that, In step S5, the unified normalization process includes: Find the minimum wear value for all nodes. With the maximum value ; The wear value of each node Mapped to color The mapping relationship is as follows: ； In the formula, The normalized thermal color value; Represents a node The wear value; This represents the minimum wear value. Maximum wear value.

10. A three-dimensional storage space heatmap generation system based on image data processing, wherein the three-dimensional storage space heatmap generation method based on image data processing according to any one of claims 1-9 is characterized in that, include: The graph structure construction module constructs a three-dimensional graph structure based on the physical location of the storage device. Each storage unit corresponds to a node in the graph structure, and the adjacency relationship between nodes is determined based on their relative positions in physical space. The data acquisition and node partitioning module is used to collect wear values ​​of some storage units from image data, and based on the collected wear values, divide the nodes into a set of observed nodes and a set of unobserved nodes. The optimization model construction and calculation module establishes a harmonic energy optimization model with boundary conditions based on the wear values ​​and graph structure in the observation node set. The heatmap generation and visualization module is used to normalize the wear values ​​of all nodes and map them into color values, thereby generating a three-dimensional heatmap. The display and control module is used to display the generated 3D heat map and provide an interactive interface for users. Based on the set conditions, the system will generate real-time status prompts and maintenance suggestions.

Images