A humanoid robot control method and system

Abstract

The present invention provides a humanoid robot control method and system, which relates to the field of robot control technology. The method includes obtaining a scanned image of a large storage tank; determining multiple first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank; determining multiple first detection points of each first corrosion degradation area based on the scanned images of the multiple first corrosion degradation areas; determining multiple second detection points of each first corrosion degradation area based on the status detection data of each first detection point; generating a second corrosion degradation degree distribution map of each first corrosion degradation area; determining multiple second corrosion degradation areas; and determining high-risk corrosion degradation areas based on the status detection data of the multiple first detection points of each second corrosion degradation area and the status detection data of the multiple second detection points of the second corrosion degradation area. The method can quickly and accurately determine the high-risk corrosion degradation areas of the large storage tank.

Description

Technical Field

[0001] The present invention relates to the field of robot control technology, and in particular to a humanoid robot control method and system. Background Art

[0002] Large storage tanks, as core infrastructure for industries such as petrochemicals and energy storage, store flammable, explosive, or toxic media for long periods of time. Their structural integrity is directly related to production safety and environmental safety. Tank corrosion can cause major safety accidents such as fires and explosions. Timely and accurate detection of tank corrosion and the identification of high-risk corrosion degradation areas are crucial to ensuring industrial production safety. Traditional tank corrosion detection relies mainly on manual visual inspections and ultrasonic thickness gauge sampling inspections. Manual inspections require point-by-point measurements, and completing comprehensive inspections of large tanks takes too long. Due to manpower and time constraints, sampling inspections can usually only be performed at certain points, making it difficult to detect potential localized corrosion. Furthermore, manual observation and inspections are susceptible to environmental influences such as lighting, as well as subjective factors such as the experience of the inspectors, resulting in a high rate of missed detections and inaccurate test results.

[0003] Therefore, how to quickly and accurately determine the high-risk corrosion degradation areas of large storage tanks is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem solved by the present invention is how to quickly and accurately determine the high-risk corrosion degradation area of ​​a large storage tank.

[0005] According to a first aspect, the present invention provides a method for controlling a humanoid robot, comprising: acquiring a scanned image of a large storage tank; determining a plurality of first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank; determining a plurality of first detection points of each first corrosion degradation area based on the scanned images of the plurality of first corrosion degradation areas; controlling a humanoid robot to detect the plurality of first detection points of each first corrosion degradation area and acquiring state detection data of each first detection point; determining a plurality of second detection points of each first corrosion degradation area based on the state detection data of each first detection point; controlling a humanoid robot to detect the plurality of first detection points of each first corrosion degradation area and acquiring state detection data of each first detection point; determining a plurality of second detection points of each first corrosion degradation area based on the state detection data of each first detection point; controlling a humanoid robot to detect the plurality of first detection points of each first corrosion degradation area and acquiring state detection data of each first detection point; controlling a humanoid robot to detect the plurality of second detection points of each first corrosion degradation area and acquiring ... The method comprises the following steps: obtaining status detection data of each second detection point of each first corrosion degraded zone and generating a second corrosion degradation degree distribution map of each first corrosion degraded zone based on the status detection data of each first detection point and the status detection data of each second detection point of each first corrosion degraded zone; determining multiple second corrosion degraded zones based on the first corrosion degradation degree distribution map of each first corrosion degraded zone and the second corrosion degradation degree distribution map of each first corrosion degraded zone; and determining a high-risk corrosion degraded zone based on the status detection data of the multiple first detection points of each second corrosion degraded zone and the status detection data of the multiple second detection points of the second corrosion degraded zone.

[0006] In a possible implementation, the state detection data includes structure state detection data and corrosion feature detection data.

[0007] In one possible implementation, the determining of a high-risk corrosion degradation zone based on the status detection data of multiple first detection points in each second corrosion degradation zone and the status detection data of multiple second detection points in the second corrosion degradation zone includes: constructing a tank map, the tank map including multiple status detection nodes and edges between the multiple status detection nodes, each status detection node being characterized by a status detection position and status detection data, and the edges being characterized by a spatial orientation relationship, a distance, and a similarity of status detection data between the status detection nodes; processing the tank map based on a graph convolutional network to determine multiple verification points in each second corrosion degradation zone; and determining a high-risk corrosion degradation zone based on the detection and verification results of the multiple verification points in each second corrosion degradation zone.

[0008] In a possible implementation, the input of the graph convolutional network is the tank map, and the output of the graph convolutional network is multiple verification points of each second corrosion degradation area.

[0009] According to a second aspect, the present invention provides a humanoid robot control system, comprising: an acquisition module for acquiring a scanned image of a large storage tank; a degradation analysis module for determining a plurality of first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank; a point planning module for determining a plurality of first detection points of each first corrosion degradation area based on the scanned images of the plurality of first corrosion degradation areas; a first detection module for controlling the humanoid robot to detect the plurality of first detection points of each first corrosion degradation area and obtain status detection data of each first detection point; a second planning module for determining a plurality of second detection points of each first corrosion degradation area based on the status detection data of each first detection point; and a second detection module for determining a plurality of second detection points of each first corrosion degradation area based on the status detection data of each first detection point. Used to control the humanoid robot to the multiple second detection points of each first corrosion degraded area to obtain the status detection data of each second detection point; a generation module is used to generate a second corrosion degradation degree distribution map of each first corrosion degraded area based on the status detection data of each first detection point and the status detection data of each second detection point of each first corrosion degraded area; an area judgment module is used to determine the multiple second corrosion degraded areas based on the first corrosion degradation degree distribution map of each first corrosion degraded area and the second corrosion degradation degree distribution map of each first corrosion degraded area; a high-risk identification module is used to determine the high-risk corrosion degraded area based on the status detection data of the multiple first detection points of each second corrosion degraded area and the status detection data of the multiple second detection points of the second corrosion degraded area.

[0010] In a possible implementation, the state detection data includes structure state detection data and corrosion feature detection data.

[0011] In one possible implementation, the high-risk identification module is further used to: construct a tank map, the tank map including multiple state detection nodes and edges between the multiple state detection nodes, the characteristics of each state detection node are the position of the state detection and the state detection data, and the characteristics of the edges are the spatial orientation relationship, distance and similarity of the state detection data between the state detection nodes; process the tank map based on a graph convolutional network to determine multiple verification points for each second corrosion degradation zone; and determine a high-risk corrosion degradation zone based on the detection and verification results of the multiple verification points of each second corrosion degradation zone.

[0012] In a possible implementation, the input of the graph convolutional network is the tank map, and the output of the graph convolutional network is multiple verification points of each second corrosion degradation area.

[0013] The present invention provides a humanoid robot control method and system, which includes obtaining a scanned image of a large storage tank; determining a plurality of first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank; determining a plurality of first detection points of each first corrosion degradation area based on the scanned images of the plurality of first corrosion degradation areas; controlling the humanoid robot to detect the plurality of first detection points of each first corrosion degradation area and obtain state detection data of each first detection point; determining a plurality of second detection points of each first corrosion degradation area based on the state detection data of each first detection point; controlling the humanoid robot to detect the plurality of second detection points of each first corrosion degradation area The method comprises the following steps: detecting points, obtaining status detection data of each second detection point; generating a second corrosion degradation degree distribution map of each first corrosion degradation zone based on the status detection data of each first detection point and the status detection data of each second detection point of each first corrosion degradation zone; determining multiple second corrosion degradation zones based on the first corrosion degradation degree distribution map of each first corrosion degradation zone and the second corrosion degradation degree distribution map of each first corrosion degradation zone; determining a high-risk corrosion degradation zone based on the status detection data of multiple first detection points in each second corrosion degradation zone and the status detection data of multiple second detection points in the second corrosion degradation zone. The method can quickly and accurately determine the high-risk corrosion degradation zone of a large storage tank. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic flow chart of a humanoid robot control method provided by an embodiment of the present invention;

[0015] Figure 2 A schematic diagram of a large storage tank provided by an embodiment of the present invention;

[0016] Figure 3 A schematic diagram of a humanoid robot provided by an embodiment of the present invention;

[0017] Figure 4 A schematic diagram of a process for determining high-risk corrosion degradation areas provided by an embodiment of the present invention;

[0018] Figure 5 A schematic diagram of constructing a tank map provided by an embodiment of the present invention;

[0019] Figure 6 A schematic diagram of a humanoid robot control system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In an embodiment of the present invention, there is provided Figure 1 A humanoid robot control method is shown, and the humanoid robot control method includes steps S1 to S9:

[0021] Step S1: Acquire a scanned image of a large storage tank.

[0022] A large storage tank is a large industrial container that can be used to store liquid or gas raw materials. Figure 2 A schematic diagram of a large storage tank according to an embodiment of the present invention. Made of metal, the large storage tank features strong sealing and large capacity. Large storage tanks are suitable for applications in petrochemicals, energy storage, food processing, and other fields.

[0023] The scanning image of a large storage tank is an image that reflects the surface morphological characteristics of the large storage tank obtained by scanning the surface and internal structure of the large storage tank with a high-resolution industrial camera.

[0024] Step S2: determining a plurality of first corrosion-degraded areas and a corrosion-degraded degree distribution map of each first corrosion-degraded area based on the scanned image of the large storage tank.

[0025] In some embodiments, a first degradation analysis model can be used to determine multiple first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank. The first degradation analysis model is a convolutional neural network model. The input of the first degradation analysis model is the scanned image of the large storage tank, and the output of the first degradation analysis model is multiple first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area.

[0026] Convolutional neural network models include convolutional neural networks (CNNs), a deep learning model that can process grid-structured data. CNNs automatically extract local features from images through multiple layers of convolution operations, then combine them layer by layer to form high-level semantic representations. The CNN architecture consists of convolutional layers, pooling layers, and fully connected layers. CNNs are able to effectively capture spatial hierarchical features in images and can handle recognition tasks with distinct visual patterns.

[0027] The first corrosion degradation zone is a potential corrosion area on a large storage tank output by the first degradation analysis model. The first corrosion degradation zone shows abnormal texture, color or geometric features in the scanned image of the large storage tank. The first corrosion degradation zone can indicate that the material may have corrosion damage of varying degrees.

[0028] The corrosion degradation distribution map for the first corrosion degradation zone is a visualization of the corrosion characteristics of the first corrosion degradation zone, generated by extracting features from a scanned image of a large storage tank using the first degradation analysis model. This map, presented as a pseudo-color image, indicates the intensity of the corrosion degradation characteristics at each location, along with the confidence level of the assessment.

[0029] In the corrosion degradation distribution map for the first corrosion degradation zone, each pixel contains two precisely quantified optical characteristics: the hue value (H value) of the color channel and the lightness value (L value) of the luminance channel. The hue value (H value) of the pixel represents the intensity of the corrosion degradation characteristic, with a hue range of 0-360 degrees corresponding to a corrosion degradation characteristic intensity of 0-1. The lightness value (L value) of the pixel represents the confidence level of the assessment, with a lightness range of 0%-100% corresponding to a confidence level of 0-1.

[0030] Scanned images of large storage tanks record specific visual changes in the material surface caused by metal corrosion. When large storage tank metal corrodes, the surface develops unique texture patterns, color anomalies, or geometric deformations, such as the brown-red color change of rust spots, the irregular edges of peeling coatings, or localized shadow differences caused by pits. These visual characteristics manifest in the image as changes in pixel-level brightness, chromaticity, and texture features. Convolutional neural networks can cluster continuous pixel regions with similar corrosion characteristics into independent first-level corrosion degradation zones through image segmentation algorithms, with each region corresponding to a spatially continuous image block with similar corrosion characteristics. During training, convolutional neural networks can learn the characteristic differences between corroded and non-corroded areas, enabling pixel-level analysis of new scanned images of large storage tanks during the inference phase and outputting a corrosion degradation distribution map for each first-level corrosion degradation zone.

[0031] Step S3: determining a plurality of first detection points of each first corroded and degraded area based on the plurality of scanned images of the first corroded and degraded areas.

[0032] The first corrosion-degraded area scanning image is a scanning image corresponding to the first corrosion-degraded area portion extracted from the scanning image of the large storage tank.

[0033] In some embodiments, a first point determination model can be used to determine multiple first detection points of each first corrosion degraded area based on multiple first corrosion degraded area scanning images. The first point determination model is a Transformer model. The input of the first point determination model is multiple first corrosion degraded area scanning images, and the output of the first point determination model is multiple first detection points of each first corrosion degraded area.

[0034] The Transformer model is a deep learning model based on the self-attention mechanism. It consists of an encoder and a decoder. The encoder learns representations of the input sequence and incorporates a self-attention mechanism and a feedforward neural network. The decoder, based on the encoder, also incorporates a multi-head attention mechanism. The decoder decodes the encoder output and generates the target sequence.

[0035] The multiple first detection points of each first corrosion degradation area are a group of key detection positions planned for each first corrosion degradation area according to the output of the first point determination model.

[0036] Scanning images of the first corrosion degradation zone can extract visual features of corrosion on the metal surfaces inside and outside large storage tanks, such as texture anomalies, color changes, or geometric deformations. These visual features have recognizable spatial distribution patterns. By analyzing the distribution patterns of these features, the Transformer model can determine which areas require more intensive detection coverage and which locations may represent critical transition zones for corrosion development. For example, the edges of corroded areas, the intersections of different texture features, and core areas where abnormal features are concentrated may all be selected as first detection points.

[0037] The Transformer model's self-attention mechanism can simultaneously consider both the local manifestations and global distribution of corrosion features. The Transformer model assesses the correlation between each region in multiple scanned images of the first corroded deterioration zone and its surrounding areas, prioritizing those locations that are both representative of the corrosion characteristics and comprehensive in the detection range as the first detection points. This approach to detection point selection based on global feature correlation ensures efficient and targeted detection.

[0038] In some embodiments, the first point determination model includes a visual extraction layer, an anomaly assessment layer, and a point analysis layer. The input of the visual extraction layer is a plurality of scanned images of the first corrosion degradation area, and the output of the visual extraction layer is the abnormal position of the texture of the corrosion area, the color change distribution sequence, and the geometric deformation area. The input of the anomaly assessment layer is the abnormal position of the texture of the corrosion area, the color change distribution sequence, and the geometric deformation area. The output of the anomaly assessment layer is a plurality of abnormal areas, the corrosion risk score of each abnormal area, the feature correlation index between abnormal areas, and the importance ranking of abnormal areas. The input of the point analysis layer is a plurality of abnormal areas, the corrosion risk score of each abnormal area, the feature correlation index between abnormal areas, and the importance ranking of abnormal areas. The output of the point analysis layer is a plurality of first detection points of each first corrosion degradation area.

[0039] Different layers are responsible for different levels of information abstraction and processing. For example, the visual extraction layer extracts corrosion-related visual features from the scanned image of the first corroded and degraded area. The anomaly assessment layer uses these features to assess the risk level and correlation of each abnormal area. The point analysis layer determines multiple specific first detection points based on the assessment results. This layered processing allows for more efficient processing of complex image information. Building multiple layers increases the system's modularity and enables more efficient information processing, thereby improving both accuracy and efficiency.

[0040] Step S4: Control the humanoid robot to perform inspections at multiple first inspection points in each first corrosion degradation zone, and obtain status inspection data of each first inspection point.

[0041] A humanoid robot is a mobile robot with a humanoid form. Figure 3 This is a schematic diagram of a humanoid robot provided by an embodiment of the present invention. The robot is equipped with a multi-degree-of-freedom robotic arm, high-precision sensors, and an autonomous navigation system. The robot is capable of performing close-range inspection tasks in complex, high-risk industrial environments.

[0042] The status detection data includes structural status detection data and corrosion characteristic detection data.

[0043] Structural status detection data is obtained by using eddy current detectors to detect changes in the electrical conductivity and magnetic permeability of the metal surface of the storage tank. Structural status detection data can be used to detect corrosion or structural defects.

[0044] The corrosion characteristic detection data is the corrosion product composition data obtained by scanning each first detection point through an X-ray fluorescence spectrometer equipped with a humanoid robot.

[0045] As an example, corrosion feature detection data can be obtained by emitting X-rays to the corrosion area through an X-ray fluorescence spectrometer and then analyzing the energy spectrum of the characteristic X-rays.

[0046] Corrosion characteristic detection data includes the percentage of elements such as sulfur and chlorine in corrosion products.

[0047] Step S5: determining a plurality of second detection points of each first corrosion degradation area based on the state detection data of each first detection point.

[0048] In some embodiments, based on the status detection data of each first detection point, a second point determination model can be used to determine multiple second detection points in each first corrosion degradation zone. The second point determination model is a deep neural network model. The input of the second point determination model is the status detection data of each first detection point, and the output of the second point determination model is multiple second detection points in each first corrosion degradation zone.

[0049] Deep neural network models include deep neural networks (DNNs). DNNs are machine learning models composed of multiple hidden layers. They can automatically learn complex features and patterns in input data through layered processing. DNNs can be trained using forward and backpropagation algorithms and establish a nonlinear mapping relationship between input data and output results.

[0050] The multiple second detection points of each first corrosion degradation area are multiple supplementary detection positions calculated based on the state detection data of each first detection point through the second point determination model.

[0051] The second detection points are more targeted and can be distributed in the surrounding areas of each first detection point where the status detection data of the first detection point is abnormal, and in the transition areas where the status detection data of the first detection point mutates.

[0052] The status detection data of the first detection point includes structural status detection data and corrosion characteristic detection data, which can reflect the actual status of corrosion development. For example, if the status detection data of a certain first detection point is abnormal, it indicates that the area may have more serious corrosion problems, and it is necessary to add a second detection point in the surrounding area for verification. If the status detection data of the current first detection point is significantly different from that of the adjacent first detection point, it indicates that the transition area also needs to add a second detection point for focused inspection.

[0053] By analyzing the spatial correlation and numerical variation patterns of the status detection data at each primary detection point, the deep neural network automatically identifies areas requiring focused inspection. The hidden layer of the deep neural network extracts nonlinear features from the status detection data, and the output layer uses these features to calculate the optimal distribution of multiple secondary detection points for each primary corrosion degradation zone, ensuring that the secondary detection points accurately cover all risk areas requiring verification.

[0054] Step S6: Control the humanoid robot to go to a plurality of second detection points in each first corrosion degradation zone to obtain status detection data of each second detection point.

[0055] When multiple second detection points are determined for each first corrosion degradation zone, the humanoid robot is controlled to move to the multiple second detection points in each first corrosion degradation zone and detect each second detection point to obtain state detection data for each second detection point. The state detection data includes structural state detection data and corrosion characteristic detection data.

[0056] Step S7: generating a second corrosion degradation degree distribution map of each first corrosion degradation area based on the state detection data of each first detection point and the state detection data of each second detection point of each first corrosion degradation area.

[0057] In some embodiments, a second corrosion degradation degree distribution map for each first corrosion degradation zone can be generated based on the state detection data of each first detection point and the state detection data of each second detection point in each first corrosion degradation zone using a generative adversarial network. The input of the generative adversarial network is the state detection data of each first detection point and the state detection data of each second detection point in each first corrosion degradation zone, and the output of the generative adversarial network is the second corrosion degradation degree distribution map for each first corrosion degradation zone.

[0058] A generative adversarial network (GAN) is a deep learning framework consisting of a generator network and a discriminator network. The generator network learns to map input data to a target distribution, while the discriminator distinguishes between generated data and real data. Through adversarial training, the two networks are continuously optimized, ultimately enabling the generator to output new samples that match the real data distribution.

[0059] The second corrosion degradation distribution map for the first corrosion degradation zone is generated using a generative adversarial network. It provides a more accurate visualization of the corrosion characteristics of the first corrosion degradation zone. Presented as a pseudo-color image, the second corrosion degradation distribution map indicates the intensity of the corrosion degradation characteristics at each location, along with the confidence level of the assessment.

[0060] In the second corrosion degradation distribution map for the first corrosion degradation zone, each pixel contains two precisely quantified optical characteristics: the hue value (H value) of the color channel and the lightness value (L value) of the luminance channel. The hue value (H value) of the pixel represents the intensity of the corrosion degradation characteristic, with a hue range of 0-360 degrees corresponding to a corrosion degradation characteristic intensity of 0-1. The lightness value (L value) of the pixel represents the confidence level of the assessment, with a lightness range of 0%-100% corresponding to a confidence level of 0-1.

[0061] The state measurement data for each first detection point in each first corrosion degradation zone, as well as the state measurement data for each second detection point, together constitute a set of spatial sampling points for corrosion characteristics. This data includes the measured state values ​​at discrete locations on the interior and exterior surfaces of large storage tanks, as well as spatial correlations between these points, such as the tendency of corrosion to spread along welds. By analyzing the spatial autocorrelation characteristics and numerical gradient changes of these data points, a model can be used to establish the distribution pattern of corrosion characteristics in continuous space.

[0062] The generator of the generative adversarial network can automatically learn the spatial interpolation rules between detection points. The generator can convert the state detection data of discrete detection points into a distribution map of continuous features. The discriminator can compare the generated distribution map with the statistical characteristics of real corrosion cases to ensure that the generated distribution map conforms to physical laws. In addition, the gradient penalty mechanism in the adversarial training process can ensure the smooth transition of the generated results in space, thereby avoiding the occurrence of sudden changes that do not conform to the actual corrosion expansion law. In this way, the generative adversarial network can reconstruct a second corrosion degradation distribution map that not only conforms to the measured value constraints but also has physical rationality from the state detection data of a limited number of points.

[0063] Step S8 : determining a plurality of second corrosion degraded areas based on the first corrosion degraded degree distribution map of each first corrosion degraded area and the second corrosion degraded degree distribution map of each first corrosion degraded area.

[0064] In some embodiments, a second degradation determination model can be used to determine multiple second corrosion degradation zones based on the first corrosion degradation degree distribution map of each first corrosion degradation zone and the second corrosion degradation degree distribution map of each first corrosion degradation zone. The second degradation determination model is a convolutional neural network model. The input of the second degradation determination model is the first corrosion degradation degree distribution map of each first corrosion degradation zone and the second corrosion degradation degree distribution map of each first corrosion degradation zone. The output of the second degradation determination model is the second corrosion degradation zone.

[0065] The second corrosion degradation area is an area with high corrosion degradation characteristic intensity and high confidence, which is screened out after the first corrosion degradation degree distribution map and the second corrosion degradation degree distribution map are fused and analyzed by the second degradation determination model on the basis of each first corrosion degradation area.

[0066] The first and second corrosion degradation degree distribution maps for each first corrosion degradation zone together constitute a dual evidence system for corrosion risk assessment. Both the first and second corrosion degradation degree distribution maps can present the corrosion characteristic intensity and confidence level at each location within the first corrosion degradation zone, but the data sources and analysis dimensions of the two can form a complementary relationship. The generation of the first corrosion degradation degree distribution map is based on image visual feature recognition, and the first corrosion degradation degree distribution map can provide a corrosion probability distribution at a macro level. The generation of the second corrosion degradation degree distribution map is based on the measured data of the detection points and can verify and correct the corrosion characteristic intensity and confidence level at a micro level.

[0067] Through its unique hierarchical feature extraction capabilities, the convolutional neural network can effectively fuse the spatial information of the first corrosion degradation distribution map and the second corrosion degradation distribution map for the first corrosion degradation zone. Using multi-scale convolution kernels, the convolutional neural network can simultaneously process the data of the two distribution maps and automatically learn the spatial correlation patterns between them. Shallow convolution can identify the consistency of local features, such as the spatial overlap between image anomaly areas and measured high-value areas, while deep convolution can establish mapping relationships between global features, such as the matching of the corrosion expansion direction with the measured data gradient. Through this feature learning process, the convolutional neural network can accurately identify areas that meet both image feature significance and measured data reliability, thereby outputting a judgment result for the second corrosion degradation zone that meets the standards.

[0068] Step S9 : determining a high-risk corrosion degradation area based on the state detection data of the plurality of first detection points of each second corrosion degradation area and the state detection data of the plurality of second detection points of the second corrosion degradation area.

[0069] In some embodiments, the Figure 4 The process is to identify high-risk corrosion degradation areas. Figure 4 A schematic diagram of a process for determining a high-risk corrosion degradation area provided by an embodiment of the present invention. The process for determining a high-risk corrosion degradation area includes steps S41 to S43:

[0070] Step S41: construct a tank map, which includes multiple state detection nodes and edges between the multiple state detection nodes. The characteristics of each state detection node are the state detection position and state detection data, and the characteristics of the edges are the spatial orientation relationship, distance and similarity of the state detection data between the state detection nodes.

[0071] A tank graph is a networked data structure that can represent large-scale tank inspection data. It consists of state detection nodes and the edges between them. A state detection node corresponds to a specific inspection location on the interior and exterior surfaces of the tank, and each node records the spatial coordinates and state detection data for that location. The edges connecting the nodes describe the spatial correlation and data relevance between different inspection locations. Edge characteristics include the spatial orientation relationship, physical distance, and similarity of the state detection data between the state detection nodes. Figure 5 A schematic diagram of a tank map provided by an embodiment of the present invention. Figure 5 As shown, Figure 5 Including state detection node A, state detection node B, state detection node C, state detection node D, the characteristics of the edge between two state detection nodes are the spatial orientation relationship, distance and similarity of state detection data between each two state detection nodes.

[0072] The positions of the status detection of the plurality of status detection nodes include a plurality of first detection points and a plurality of second detection points in a plurality of second corrosion degradation areas.

[0073] In some embodiments, the similarity between the state detection data in each detection node can be determined by a deep neural network model.

[0074] By constructing a tank atlas, a modeling form, discrete detection data can be organized into an interconnected network system, thereby achieving a quantitative expression of the spatial distribution laws of tank corrosion characteristics and structural states.

[0075] Step S42: Process the tank map based on a graph convolutional network to determine multiple verification points for each second corrosion degradation zone.

[0076] A graph convolutional network (GCN) is a deep learning model that can process networked data. It extracts node features through a local neighborhood aggregation mechanism. A GCN is a neural network that operates directly on graph data. The input of the GCN is the tank graph, and the output is multiple verification points for each second corrosion degradation zone.

[0077] The multiple verification points of each second corrosion degradation zone are a set of decisive detection position points determined through graph convolutional network analysis for the final judgment of high-risk corrosion degradation zones.

[0078] The tank graph represents condition detection nodes and the relationships between them as a graph structure, integrating various information (such as the location of condition detection and condition detection data) into a single structure. Edge features include the spatial orientation relationship, distance, and similarity of condition detection data between detection point nodes. Edges can represent the spatial position relationship and data association between detection points. This information can help the graph convolutional network better understand the propagation path and distribution patterns of corrosion characteristics on the tank surface. The tank graph can integrate feature information from multiple nodes, including the location of condition detection and condition detection data. By integrating this multi-source information into the tank graph, more comprehensive input data can be provided, enabling the graph convolutional network to more accurately learn and determine multiple verification points for each secondary corrosion degradation zone.

[0079] When processing tank graphs, the graph convolutional network can extract the spatial correlation features of each state detection node and its surrounding nodes through neighborhood aggregation operations, and then use multi-layer convolution kernels to learn the spatial propagation law of corrosion features. Finally, through the node classification task, high-risk areas that require key verification can be identified.

[0080] Step S43: determining a high-risk corrosion degradation area based on the detection and verification results of the multiple verification points of each second corrosion degradation area.

[0081] The verification results of the verification points are obtained by using a humanoid robot equipped with an ultrasonic phased array thickness gauge and a laser confocal microscope to perform two detection tasks: metal wall thickness measurement and surface morphology scanning on the verification points in the second corrosion degradation zone. Then, after data filtering, noise reduction, feature extraction and evaluation, a standardized data set including the actual wall thickness measurement value, pitting characteristic parameters, and residual strength coefficient is formed.

[0082] In some embodiments, a high-risk corrosion degradation zone can be determined based on a deep neural network model based on the detection and verification results of multiple verification points of each second corrosion degradation zone. The input of the deep neural network model is the detection and verification results of multiple verification points of each second corrosion degradation zone, and the output of the deep neural network model is the high-risk corrosion degradation zone.

[0083] High-risk corrosion degradation areas are identified through deep neural network assessment as areas of severe corrosion that pose a direct threat to the structural safety of large storage tanks. These areas are characterized by metal remaining thickness significantly below design standards, the presence of penetrating defects or cracks, and corrosion rates exceeding safety thresholds. Immediate repair or replacement measures are required to prevent safety incidents such as tank leakage and structural failure.

[0084] Verification results from verification points include multi-dimensional data such as residual metal thickness, pitting characteristic parameters, and residual strength coefficients. Deep neural networks use multiple layers of neurons to perform nonlinear transformations and feature extraction on this data, automatically mining the complex relationships and underlying patterns among the verification results. Deep neural networks can learn patterns in corrosion risk levels corresponding to different data combinations, enabling the identification of high-risk corrosion degradation areas based on verification results from multiple verification points.

[0085] Based on the same inventive concept, Figure 6 A schematic diagram of a humanoid robot control system provided in an embodiment of the present invention, wherein the humanoid robot control system includes:

[0086] An acquisition module 61 is used to acquire a scanned image of a large storage tank;

[0087] a degradation analysis module 62 for determining a plurality of first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on the scanned image of the large storage tank;

[0088] A point planning module 63 is configured to determine a plurality of first detection points of each first corroded and degraded area based on the plurality of scanned images of the first corroded and degraded areas;

[0089] A first detection module 64 is used to control the humanoid robot to detect multiple first detection points in each first corrosion degradation area and obtain status detection data of each first detection point;

[0090] A second planning module 65 is configured to determine a plurality of second detection points in each first corrosion degradation area based on the state detection data of each first detection point;

[0091] A second detection module 66 is used to control the humanoid robot to go to a plurality of second detection points in each first corrosion degradation zone to obtain status detection data of each second detection point;

[0092] A generating module 67 is configured to generate a second corrosion degradation degree distribution map of each first corrosion degradation area based on the state detection data of each first detection point and the state detection data of each second detection point of each first corrosion degradation area;

[0093] an area determination module 68 for determining a plurality of second corrosion degradation areas based on the first corrosion degradation degree distribution map of each first corrosion degradation area and the second corrosion degradation degree distribution map of each first corrosion degradation area;

[0094] The high-risk identification module 69 is configured to determine a high-risk corrosion degradation area based on the state detection data of the multiple first detection points in each second corrosion degradation area and the state detection data of the multiple second detection points in the second corrosion degradation area.

[0095] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0096] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the implementations explicitly described and illustrated in this specification.

Claims

1. A method for controlling a humanoid robot, characterized in that: include: Obtain scanned images of large storage tanks; Determining a plurality of first corrosion-degraded areas and a corrosion-degraded degree distribution map of each first corrosion-degraded area based on a scanned image of the large storage tank; Determining a plurality of first detection points of each first corroded and degraded area based on the plurality of scanned images of the first corroded and degraded areas; Controlling the humanoid robot to perform inspections at multiple first inspection points in each first corrosion degradation zone, and obtaining status inspection data of each first inspection point; Determining a plurality of second detection points of each first corrosion degradation area based on the state detection data of each first detection point; Controlling the humanoid robot to a plurality of second detection points in each of the first corrosion degradation zones to obtain status detection data of each second detection point; generating a second corrosion degradation degree distribution map of each first corrosion degradation area based on the state detection data of each first detection point and the state detection data of each second detection point of each first corrosion degradation area; determining a plurality of second corrosion deterioration areas based on the first corrosion deterioration degree distribution map of each first corrosion deterioration area and the second corrosion deterioration degree distribution map of each first corrosion deterioration area; The high-risk corrosion degradation area is determined based on the state detection data of the plurality of first detection points of each second corrosion degradation area and the state detection data of the plurality of second detection points of the second corrosion degradation area.

2. The humanoid robot control method according to claim 1, wherein: The state detection data includes structure state detection data and corrosion characteristic detection data.

3. The humanoid robot control method according to claim 1, wherein: The step of determining the high-risk corrosion degradation area based on the state detection data of the plurality of first detection points in each second corrosion degradation area and the state detection data of the plurality of second detection points in the second corrosion degradation area comprises: Constructing a tank graph, wherein the tank graph includes a plurality of state detection nodes and edges between the plurality of state detection nodes, wherein each state detection node is characterized by a state detection position and state detection data, and the edges are characterized by a spatial orientation relationship, a distance, and a similarity of the state detection data between the state detection nodes; Processing the tank map based on a graph convolutional network to determine multiple verification points of each second corrosion degradation zone; A high-risk corrosion degradation area is determined based on the detection and verification results of multiple verification points of each second corrosion degradation area.

4. The humanoid robot control method according to claim 3, wherein: The input of the graph convolutional network is the tank map, and the output of the graph convolutional network is multiple verification points of each second corrosion degradation area.

5. A humanoid robot control system, characterized in that: include: An acquisition module, used to acquire scanned images of large storage tanks; a degradation analysis module, configured to determine a plurality of first corrosion degradation areas and a corrosion degradation degree distribution map of each first corrosion degradation area based on a scanned image of the large storage tank; A point planning module, configured to determine a plurality of first detection points of each first corroded and degraded area based on the plurality of scanned images of the first corroded and degraded areas; A first detection module is used to control the humanoid robot to detect multiple first detection points in each first corrosion degradation area and obtain status detection data of each first detection point; A second planning module is configured to determine a plurality of second detection points in each first corrosion degradation area based on the state detection data of each first detection point; a second detection module, configured to control the humanoid robot to a plurality of second detection points in each first corrosion degradation zone, and obtain status detection data of each second detection point; a generating module, configured to generate a second corrosion degradation degree distribution map of each first corrosion degradation area based on the state detection data of each first detection point and the state detection data of each second detection point of each first corrosion degradation area; an area determination module, configured to determine a plurality of second corrosion degradation areas based on the first corrosion degradation degree distribution map of each first corrosion degradation area and the second corrosion degradation degree distribution map of each first corrosion degradation area; The high-risk identification module is used to determine the high-risk corrosion degradation area based on the state detection data of the multiple first detection points of each second corrosion degradation area and the state detection data of the multiple second detection points of the second corrosion degradation area.

6. The humanoid robot control system according to claim 5, characterized in that: The state detection data includes structure state detection data and corrosion characteristic detection data.

7. The humanoid robot control system according to claim 5, wherein: The high-risk identification module is also used to: Constructing a tank graph, wherein the tank graph includes a plurality of state detection nodes and edges between the plurality of state detection nodes, wherein each state detection node is characterized by a state detection position and state detection data, and the edges are characterized by a spatial orientation relationship, a distance, and a similarity of the state detection data between the state detection nodes; Processing the tank map based on a graph convolutional network to determine multiple verification points of each second corrosion degradation zone; A high-risk corrosion degradation area is determined based on the detection and verification results of multiple verification points of each second corrosion degradation area.

8. The humanoid robot control system according to claim 7, wherein: The input of the graph convolutional network is the tank map, and the output of the graph convolutional network is multiple verification points of each second corrosion degradation area.

Images