Quality detection method and system for intelligent closestool production

Abstract

The invention relates to the field of intelligent production detection, and provides a quality detection method and system for intelligent closestool production, and the method comprises the steps: analyzing the infrared images of a to-be-detected intelligent closestool at a plurality of time points in a temperature control performance test, so as to know the temperature control performance of the to-be-detected intelligent closestool; the cleaning performance of the to-be-detected intelligent closestool is known by analyzing different spray diagrams of the flushing nozzle in multiple modes in the cleaning performance test of the to-be-detected intelligent closestool, and whether the to-be-detected intelligent closestool is qualified or not is judged by comprehensively considering the temperature control performance of the to-be-detected intelligent closestool and the cleaning performance of the to-be-detected intelligent closestool. In this way, more comprehensive and intelligent detection can be achieved, and the accuracy and efficiency of quality detection of the intelligent closestool can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent production inspection, and more specifically, to a quality inspection method and system for intelligent toilet production. Background Art

[0002] With the advancement of technology and consumers' pursuit of high-quality life, people are increasingly inclined to choose products that can provide more comfort and convenience, such as smart toilets. Smart toilets are sanitary ware products with intelligent functions. They not only have the basic functions of traditional toilets, such as excretion and flushing, but also integrate advanced electronics, sensors, computers and other technologies to achieve a variety of intelligent functions, such as seat heating, warm water washing, warm air drying, etc. It can provide users with a warm and comfortable toilet experience.

[0003] In traditional smart toilet production quality inspection, the focus is on performance inspection in one aspect. In many cases, certain data will be ignored, and there are problems such as poor data integration and low level of intelligent processing. This can easily lead to a decrease in the accuracy of smart toilet quality inspection and unreliable inspection results.

[0004] Therefore, an optimized quality inspection solution for smart toilet production is needed. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present application provides a quality inspection method and system for the production of smart toilets.

[0006] According to one aspect of the present application, a quality inspection method for intelligent toilet production is provided, which includes:

[0007] Obtaining the spray images of the cleaning nozzle of the intelligent toilet to be tested in multiple modes in the cleaning performance test and the infrared images of the seat ring at multiple time points in the temperature control performance test;

[0008] Performing data processing on the spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram;

[0009] The smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph are fused and feature-screened to obtain a smart toilet performance detection characteristic vector;

[0010] Based on the smart toilet performance detection feature vector, a result of whether the smart toilet to be detected is qualified is obtained.

[0011] According to another aspect of the present application, a quality inspection system for intelligent toilet production is provided, comprising:

[0012] The intelligent toilet detection related data acquisition module is used to obtain the spray images of the cleaning nozzle of the intelligent toilet to be tested in multiple modes in the cleaning performance test and the infrared images of the seat ring at multiple time points in the temperature control performance test;

[0013] A data processing module for intelligent toilet detection related data, used to respectively process the spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram;

[0014] A smart toilet detection feature generation module, used for fusing and feature screening the smart toilet cleaning performance detection feature graph and the smart toilet temperature control performance detection feature graph to obtain a smart toilet performance detection feature vector;

[0015] The smart toilet detection result generation module is used to obtain the result of whether the smart toilet to be detected is qualified based on the smart toilet performance detection feature vector.

[0016] This application has significant technical effects due to the adoption of the above technical solutions:

[0017] The quality inspection method and system for smart toilet production provided by the present application analyzes the infrared images of the seat ring of the smart toilet to be inspected at multiple time points in the temperature control performance test to understand the temperature control performance of the smart toilet to be inspected, analyzes the different spray patterns of the flushing nozzle in multiple modes of the smart toilet to be inspected in the cleaning performance test to understand the cleaning performance of the smart toilet to be inspected, and comprehensively considers the temperature control performance of the smart toilet to be inspected and the cleaning performance of the smart toilet to be inspected to judge whether the smart toilet to be inspected is qualified. In this way, a more comprehensive and intelligent inspection can be achieved, which is conducive to improving the accuracy and efficiency of smart toilet quality inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 This is a flow chart of a quality inspection method for smart toilet production according to an embodiment of the present application.

[0020] Figure 2This is a data flow diagram of a quality inspection method for smart toilet production according to an embodiment of the present application.

[0021] Figure 3 In the quality inspection method for smart toilet production according to an embodiment of the present application, data processing is performed on the spray images of the cleaning nozzle in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain a flowchart of the smart toilet cleaning performance detection characteristic diagram and the smart toilet temperature control performance detection characteristic diagram.

[0022] Figure 4 The present invention is a flow chart of a quality inspection method for smart toilet production according to an embodiment of the present application, in which feature extraction is performed on the spray images of the cleaning nozzles in the multiple modes to obtain the smart toilet cleaning performance inspection feature map.

[0023] Figure 5 This is a flow chart of a quality inspection method for smart toilet production according to an embodiment of the present application, in which feature extraction is performed on the seat ring infrared images at multiple time points to obtain the smart toilet temperature control performance detection feature diagram.

[0024] Figure 6 In the quality inspection method for smart toilet production according to an embodiment of the present application, the smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph are fused and feature-screened to obtain a smart toilet performance detection feature vector.

[0025] Figure 7 This is a system block diagram of a quality inspection system for smart toilet production according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0027] With the continuous advancement of science and technology, consumers are increasingly pursuing a high-quality life. As an important part of smart home, smart toilets have gradually become the new favorite of modern families. Smart toilets not only have the basic functions of traditional toilets, such as excretion and flushing, but also integrate advanced technologies such as electronics, sensors, and computers to achieve a variety of intelligent functions such as seat heating, warm water cleaning, and warm air drying, providing users with a warm and comfortable toilet experience.

[0028] In terms of quality inspection, the smart toilet industry is facing some challenges. In traditional smart toilet production quality inspection, the focus is on performance inspection in a certain aspect. In many cases, some data will be ignored, and there are problems such as poor data integration and low degree of intelligent processing. This can easily lead to a decrease in the accuracy of smart toilet quality inspection and unreliable inspection results.

[0029] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for quality inspection in the production of smart toilets.

[0030] Figure 1 This is a flow chart of a quality inspection method for smart toilet production according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the quality inspection method for smart toilet production according to the embodiment of the present application. Figure 1 and Figure 2 As shown, the quality inspection method for smart toilet production according to the embodiment of the present application includes: S110, obtaining the spray images of the cleaning nozzle in multiple modes of the smart toilet to be inspected in the cleaning performance test and the infrared images of the seat at multiple time points in the temperature control performance test; S120, respectively processing the spray images of the cleaning nozzle in the multiple modes and the infrared images of the seat at the multiple time points to obtain the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map; S130, fusing and feature screening the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map to obtain a smart toilet performance detection feature vector; S140, based on the smart toilet performance detection feature vector, obtaining the result of whether the smart toilet to be inspected is qualified.

[0031] In step S110, the spray images of the cleaning nozzles of the intelligent toilet to be tested in multiple modes in the cleaning performance test and the infrared images of the seat ring at multiple time points in the temperature control performance test are obtained. It should be understood that the infrared image of the seat ring mainly contains the temperature distribution information of the seat ring surface. These images are usually presented in the form of heat maps, where different colors represent different temperature ranges. By obtaining and analyzing the infrared images of the seat ring at multiple time points, it is possible to understand the heating conditions of the seat ring at different time points, the uniformity of the temperature distribution, and whether there are overheated or overcooled areas. The temperature control performance of the intelligent toilet directly determines the heating effect of the seat ring and the comfort of the user. If the temperature control is inaccurate, the seat ring will be heated unevenly, which will affect the user's toilet experience. The spray image of the cleaning nozzle mainly contains information such as the shape, speed and coverage of the spray water flow. These images can intuitively show the spray effect of the cleaning nozzle, including the intensity, direction and coverage of the water flow. By analyzing these images, it can be understood whether the cleaning nozzle can spray water evenly. Intelligent toilets are usually equipped with multiple cleaning modes, such as butt washing, women's washing, etc., to meet the needs of different users. The spray effect of the cleaning nozzle in each mode may be different. Therefore, in order to comprehensively evaluate the cleaning performance of the smart toilet, it is necessary to obtain spray images in multiple modes. This helps to understand whether the spray effect of the cleaning nozzle in different modes is uniform and stable, that is, to evaluate the cleaning performance of the smart toilet through the spray images of the cleaning nozzle in different modes. In one embodiment of the present application, an infrared thermal imager can be used to collect infrared image data of the seat ring, and a high-resolution camera can be used to collect the spray image data of the cleaning nozzle.

[0032] In step S120, the spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points are processed to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram. Specifically, Figure 3 The present invention is a flow chart of a quality inspection method for smart toilet production according to an embodiment of the present application, in which the spray images of the cleaning nozzle in the multiple modes and the infrared images of the seat ring at the multiple time points are processed to obtain a smart toilet cleaning performance inspection characteristic diagram and a smart toilet temperature control performance inspection characteristic diagram. Figure 3As shown, data processing is performed on the spray images of the cleaning nozzle in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain the smart toilet cleaning performance detection characteristic diagram and the smart toilet temperature control performance detection characteristic diagram, including: S121, feature extraction of the spray images of the cleaning nozzle in the multiple modes to obtain the smart toilet cleaning performance detection characteristic diagram; S122, feature extraction of the infrared images of the seat ring at the multiple time points to obtain the smart toilet temperature control performance detection characteristic diagram.

[0033] More specifically, Figure 4 This is a flow chart of extracting features of the spray images of the cleaning nozzles in the multiple modes to obtain the cleaning performance detection feature map of the smart toilet in the quality inspection method for smart toilet production according to the embodiment of the present application. Figure 4 As shown, feature extraction is performed on the spray images of the cleaning nozzles in the multiple modes to obtain the smart toilet cleaning performance detection feature map, including: S1211, the spray images of the cleaning nozzles in the multiple modes are respectively passed through the cleaning nozzle spray image encoder to obtain multiple smart toilet cleaning nozzle spray condition feature matrices; S1212, the multiple smart toilet cleaning nozzle spray condition feature matrices are aggregated along the channel dimension to obtain the smart toilet cleaning performance detection feature map.

[0034] Considering that the spray image of the cleaning nozzle contains the shape information of the spray water flow, specifically the uniformity and regularity of the shape of the spray water flow, and whether there is obvious deviation or scattering phenomenon, which helps to evaluate whether the cleaning nozzle can provide consistent and effective cleaning effects in the entire cleaning area. It also contains the coverage information of the spray water flow. In order to capture the information features directly related to the cleaning performance of the smart toilet in the spray image, it is necessary to pass the spray images of the cleaning nozzles in the multiple modes through the cleaning nozzle spray image encoder to obtain multiple smart toilet cleaning nozzle spray situation feature matrices. Through the processing of the cleaning nozzle spray image encoder, relevant features can be effectively captured, thereby obtaining a smart toilet cleaning nozzle spray situation feature matrix containing relevant associated features. The cleaning nozzle spray image encoder in this application specifically refers to a convolutional neural network model in which adjacent layers use mutually transposed convolution kernels. Since adjacent layers use mutually transposed convolution kernels, this model can maintain feature consistency during forward propagation and back propagation, thereby improving the accuracy of feature extraction, which helps to more accurately capture the detail information in the spray image. In addition, by using convolution layers with mutually transposed convolution kernels, the number of model parameters can be reduced to a certain extent, thereby reducing the complexity and computational cost of the model, which in turn helps to improve the operating efficiency and response speed of the model in practical applications.

[0035] Considering that each intelligent toilet cleaning nozzle spray situation feature matrix corresponds to the cleaning nozzle spray image in a specific mode, these images contain various feature information of the cleaning nozzle in this mode, such as the shape, speed, coverage, etc. of the spray water flow. However, analyzing each feature matrix separately may not fully reflect the overall cleaning performance of the intelligent toilet, because the cleaning performance is the result of the combined effect of multiple factors. In order to integrate the feature matrices under different modes, the multiple intelligent toilet cleaning nozzle spray situation feature matrices are aggregated along the channel dimension in the technical solution of the present application to obtain the intelligent toilet cleaning performance detection feature map. Through aggregation, a smart toilet cleaning performance detection feature map containing more comprehensive information can be formed. This feature map not only contains the feature information under each mode, but also can reflect the correlation and mutual influence between these feature information to a certain extent. It is also worth mentioning that after aggregating multiple feature matrices into one feature map, the subsequent analysis and processing process can be simplified. This helps to improve evaluation efficiency, reduce computing costs, and make the evaluation results easier to understand and apply.

[0036] More specifically, Figure 5 This is a flow chart of extracting features from the infrared images of the seat ring at multiple time points to obtain the temperature control performance detection feature map of the smart toilet in the quality inspection method for smart toilet production according to an embodiment of the present application. Figure 5 As shown, feature extraction is performed on the seat infrared images at the multiple time points to obtain the temperature control performance detection characteristic diagram of the smart toilet, including: S1221, respectively passing the seat infrared images at the multiple time points through the seat infrared image encoder to obtain multiple smart toilet seat temperature feature matrices; S1222, aggregating the multiple smart toilet seat temperature feature matrices along the channel dimension to obtain the smart toilet temperature control performance detection characteristic diagram.

[0037] Considering that the seat infrared images at different time points contain the temperature distribution information of the seat surface, specifically, each seat infrared image intuitively displays the temperature distribution of the seat surface, including high temperature areas, low temperature areas and temperature gradients, etc., and the seat infrared images at multiple time points also contain the change trend information of the seat temperature, such as the heating rate, the cooling rate and the amplitude of the temperature fluctuation, etc. In order to extract these features from the original image, in the technical solution of the present application, it is necessary to pass the seat infrared images at the multiple time points through the seat infrared image encoder to obtain multiple intelligent toilet seat temperature feature matrices. In the present application, the seat infrared image encoder specifically refers to a convolutional neural network model using a global conversion matrix as a filter. The global conversion matrix can capture the global features in the image, including the overall trend of the temperature distribution, etc. This helps to more comprehensively understand the information in the seat infrared image, so as to more accurately evaluate the temperature control performance of the intelligent toilet. At the same time, the convolutional neural network model using the global conversion matrix as a filter has strong robustness to noise and interference in the seat infrared image. This means that even if there is a certain degree of noise or interference in the seat infrared image, the encoder can still accurately extract useful feature information.

[0038] Accordingly, considering that each temperature characteristic matrix of the smart toilet seat is obtained through an encoder based on the infrared image of the seat at a specific time point, it contains the temperature distribution information of the seat at that time point. However, only analyzing the temperature characteristic matrix at a single time point may not be able to fully reflect the temperature control performance of the smart toilet, because the temperature control performance is a process that changes over time and requires consideration of data at multiple time points. Based on this, in the technical solution of the present application, it is necessary to aggregate the multiple smart toilet seat temperature characteristic matrices along the channel dimension to obtain the smart toilet temperature control performance detection characteristic diagram. By aggregating along the channel dimension, the temperature characteristic matrices of multiple time points can be integrated together to form a smart toilet temperature control performance detection characteristic diagram containing more complete information. This characteristic diagram not only contains the temperature distribution information at each time point, but also can reflect the temperature change trend and stability over time to a certain extent.

[0039] In step S130, the smart toilet cleaning performance detection feature graph and the smart toilet temperature control performance detection feature graph are fused and feature-screened to obtain a smart toilet performance detection feature vector. Specifically, Figure 6 This is a flow chart of fusing and feature screening the smart toilet cleaning performance detection feature graph and the smart toilet temperature control performance detection feature graph to obtain a smart toilet performance detection feature vector in a quality detection method for smart toilet production according to an embodiment of the present application. Figure 6As shown, the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map are fused and feature screened to obtain a smart toilet performance detection feature vector, including: S131, fusing the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map to obtain a smart toilet performance detection fusion feature map; S132, performing mean pooling of each feature matrix along the channel dimension on the smart toilet performance detection fusion feature map to obtain a smart toilet performance detection fusion feature vector; S133, passing the smart toilet performance detection fusion feature map through a performance detection key feature capturer to obtain a smart toilet performance detection fusion enhancement feature map; S134, performing mean pooling of each feature matrix along the channel dimension on the smart toilet performance detection fusion enhancement feature map to obtain a smart toilet performance detection fusion enhancement feature vector; S135, performing feature association coupling based on low-dimensional modulated perspective collaborative projection on the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhancement feature vector to obtain the smart toilet performance detection feature vector.

[0040] The performance evaluation of smart toilets is a complex process that requires comprehensive consideration of performance indicators from multiple aspects. Cleaning performance and temperature control performance are two important performance indicators of smart toilets, which reflect the cleaning ability of the toilet and the comfort of the user, respectively. In order to achieve a comprehensive evaluation of the performance of smart toilets and avoid ignoring other important aspects due to focusing only on a single performance indicator, the technical solution of the present application needs to fuse the smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph to obtain a smart toilet performance detection fusion characteristic graph. By fusing the smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph, multiple performance indicators of the smart toilet can be comprehensively considered, thereby improving the accuracy of the evaluation. Compared with a single performance indicator, the fused feature graph can more comprehensively reflect the actual performance of the smart toilet and reduce the possibility of misjudgment.

[0041] Accordingly, considering that the fused feature map of the smart toilet performance detection obtained after fusion usually contains a large amount of feature information, but not all feature information is effective for the smart toilet performance detection, that is, the fused feature map is redundant, in order to remove these redundant feature information to reduce the subsequent unnecessary calculation amount, the technical solution of the present application needs to perform mean pooling of each feature matrix along the channel dimension of the fused feature map of the smart toilet performance detection to obtain the fused feature vector of the smart toilet performance detection. Through the mean pooling operation, each feature matrix along the channel dimension of the fused feature map can be downsampled, thereby effectively reducing the dimension of the feature and thus reducing the amount of calculation.

[0042] Although a relatively comprehensive fusion feature map of smart toilet performance detection has been obtained through the above-mentioned fusion steps, this feature map still contains some redundant information or noise. These redundant information have little effect on subsequent classification judgment and even its existence may cause interference. In order to effectively reduce the interference of these redundant information, the technical solution of this application needs to pass the smart toilet performance detection fusion feature map through a performance detection key feature capturer to obtain a smart toilet performance detection fusion enhanced feature map. Through the processing of the performance detection key feature capturer, more critical and useful features can be extracted, thereby improving the effective feature utilization and the accuracy of performance evaluation, and thus providing strong support for the quality inspection and production optimization of smart toilets. The performance detection key feature capturer in this application specifically refers to a convolutional neural network model using a spatial attention mechanism and a channel attention mechanism. By introducing the spatial attention mechanism, the convolutional neural network model can pay more attention to the more important areas in the fusion feature map of the smart toilet performance detection. Specifically, by calculating the important weights of different areas in the feature map, the features are weighted, thereby highlighting the information of the key areas; and by introducing the channel attention mechanism, the network can pay attention to the feature channels that are more useful for the current smart toilet production quality inspection task. Specifically, by calculating the importance weights of different channels, the features are weighted, thereby highlighting the information of the key channels. That is, by introducing the attention mechanism, the model can pay attention to more important areas and channels, thereby reducing the impact of redundant information and improving the accuracy and reliability of the results.

[0043] Next, the smart toilet performance detection fusion enhanced feature map is subjected to mean pooling of each feature matrix along the channel dimension to obtain the smart toilet performance detection fusion enhanced feature vector. It should be understood that the fusion enhanced feature map of the smart toilet performance detection has a high dimension and a large computational complexity. In order to simplify the subsequent processing steps and reduce the amount of calculation, it is necessary to perform dimensionality reduction processing on it, that is, to perform mean pooling operations on each feature matrix along the channel dimension. It is worth mentioning that the mean pooling operation has a certain robustness to noise and slight changes in the feature map. In this way, the impact of noise and slight changes on the detection results can be reduced, and the stability and reliability of the features can be improved.

[0044] The fusion feature vector of smart toilet performance detection contains the basic and comprehensive feature information extracted from the original data, while the fusion enhanced feature vector of smart toilet performance detection extracts deeper and more representative feature information. In order to make full use of the information advantages of the two and form a more comprehensive and accurate feature representation, the two feature vectors can be fused to integrate information from different feature spaces, which helps to improve the accuracy of smart toilet quality detection. In particular, considering that the fusion feature vector of smart toilet performance detection is specifically derived from the aggregation of the cleaning nozzle spray situation and the seat ring temperature characteristics, and the fusion enhanced feature vector of smart toilet performance detection is obtained by the key feature capturer of performance detection based on the fusion feature map, the generation process of these two feature vectors involves different image encoders and feature capturers, resulting in essential differences in their feature expression. The smoothness of the feature manifold refers to the continuity and consistency between different feature points in the feature space. If the feature vector cannot maintain this continuity during the fusion process, it will lead to the non-smoothness of the feature manifold, thereby affecting the certainty of the feature expression. In the performance detection of smart toilets, due to the existence of feature volatility, the fused feature vector may form a discontinuous area in the feature space, which in turn increases the uncertainty of feature expression. Therefore, in the technical solution of the present application, the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector are subjected to feature correlation coupling based on low-dimensional modulated perspective collaborative projection to obtain the smart toilet performance detection feature vector.

[0045] Specifically, in an embodiment of the present application, a possible implementation method for performing feature correlation coupling based on low-dimensional modulated perspective collaborative projection on the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector to obtain the smart toilet performance detection feature vector may be: constructing a smart toilet performance detection distance topology matrix between the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector; constructing a smart toilet performance detection eigenvalue granularity association matrix between the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector; based on the smart toilet performance detection distance topology matrix, The smart toilet performance detection eigenvalue granularity association matrix is ​​topologically associated modulated to obtain a smart toilet performance detection topological modulation association matrix; the smart toilet performance detection topological modulation association matrix is ​​used as the perspective co-projection space, and the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhancement feature vector are projected into the perspective co-projection space to obtain a perspective modulated smart toilet performance detection fusion feature vector and a perspective modulated smart toilet performance detection fusion enhancement feature vector; the perspective modulated smart toilet performance detection fusion feature vector and the perspective modulated smart toilet performance detection fusion enhancement feature vector are fused to obtain the smart toilet performance detection feature vector.

[0046] In an embodiment of the present application, another implementation expression of performing feature correlation coupling based on low-dimensional modulated perspective collaborative projection on the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector to obtain the smart toilet performance detection feature vector can be: processing the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector using the following formula to obtain the smart toilet performance detection feature vector; wherein the formula is:

[0047] M t =D(V1,V2)

[0048]

[0049] V c =α·V1′+β·V2′

[0050] Among them, V1 represents the intelligent toilet performance detection fusion feature vector, V2 represents the intelligent toilet performance detection fusion enhanced feature vector, D(V1, V2) is the intelligent toilet performance detection distance topological matrix between the intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhanced feature vector, that is, d i,j =d(v 1i ,v 2j ), d i,j is the eigenvalue of the position (i, j) in the distance topology matrix of the smart toilet performance detection, d(v 1i ,v 2j ) represents the distance between the eigenvalue of the ith position of the intelligent toilet performance detection fusion feature vector and the eigenvalue of the jth position of the intelligent toilet performance detection fusion enhanced feature vector, and V1 and V2 are both column vectors, M t represents the intelligent toilet performance detection distance topology matrix, T represents the transpose of the vector, Represents matrix multiplication, M v represents the eigenvalue granularity correlation matrix of the smart toilet performance detection, exp represents the natural exponential function, M c represents the topological modulation association matrix of the intelligent toilet performance detection, V1' represents the fusion feature vector of the perspective modulation intelligent toilet performance detection, V2' represents the fusion enhancement feature vector of the perspective modulation intelligent toilet performance detection, α and β represent weighted hyperparameters, V c Represents the smart toilet performance detection feature vector.

[0051] That is, in the technical solution of the present application, the intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhanced feature vector are subjected to feature correlation coupling based on low-dimensional modulated perspective collaborative projection, which first constructs the intelligent toilet performance detection distance topological matrix between the intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhanced feature vector. This step utilizes the idea of ​​graph theory, takes the eigenvalues ​​of each position in the intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhanced feature vector as nodes, and uses the distance metric function to determine the low-dimensional embedding expression of the edges between nodes, so as to construct a feature expression that can quantify the topological correlation relationship between the intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhanced feature vector. This step is crucial to understanding the spatial distribution of data points and their mutual connections, and also provides basic structural information for subsequent steps.

[0052] Next, the smart toilet performance detection eigenvalue granularity association matrix between the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector is constructed. This step focuses on exploring the correlation between the internal attributes of the feature vector. In a specific example, the product between the transposed vectors of the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector is calculated to obtain the smart toilet performance detection eigenvalue granularity association matrix. Furthermore, the smart toilet performance detection eigenvalue granularity association matrix is ​​topologically modulated based on the smart toilet performance detection distance topological matrix to obtain the smart toilet performance detection topological modulation association matrix. That is, the eigenvalue granularity association information of the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector is driven to wander on the distance topological map to construct a low-dimensional modulation perspective co-projection space domain for mapping the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector to a continuous high-dimensional regression space attribute.

[0053] Next, the smart toilet performance detection topological modulation association matrix is ​​used as the perspective co-projection space, and the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhancement feature vector are projected into the perspective co-projection space to obtain the perspective modulation smart toilet performance detection fusion feature vector and the perspective modulation smart toilet performance detection fusion enhancement feature vector. That is, the key to this technical solution is to find an implicit third space with modulation capability, in which as much information of the original high-dimensional data as possible is retained, so that the converted new feature vector not only reduces the dimension, but also maintains the relative position relationship between the original features.

[0054] Finally, the perspective modulation intelligent toilet performance detection fusion feature vector and the perspective modulation intelligent toilet performance detection fusion enhancement feature vector are fused to obtain the intelligent toilet performance detection feature vector. Fusion can be achieved in a variety of forms, ranging from simple arithmetic average to complex ensemble learning algorithms, depending on the application scenario and the goals pursued. The final intelligent toilet performance detection feature vector combines the main features of the two sets of input data and provides a more comprehensive data representation, which is not only conducive to improving the performance of machine learning models, but also promotes the ability of cross-modal or multi-source data analysis, thereby significantly enhancing the understanding and utilization efficiency of complex data structures.

[0055] In step S140, based on the smart toilet performance detection feature vector, a result of whether the smart toilet to be detected is qualified is obtained. Specifically, in an embodiment of the present application, based on the smart toilet performance detection feature vector, a result of whether the smart toilet to be detected is qualified is obtained, including: passing the smart toilet performance detection feature vector through a quality detection classifier to obtain a quality detection classification result, and the quality detection classification result is used to indicate whether the smart toilet to be detected is qualified. It should be understood that the smart toilet performance detection feature vector is complex feature information and does not directly correspond to the quality detection result of smart toilet production. In order to be able to convert complex information into easy-to-understand and easy-to-apply classification results, thereby achieving an accurate assessment of the quality of the smart toilet, it is necessary to input the smart toilet performance detection feature vector into the quality detection classifier. The quality detection classifier is a classification model based on deep learning, which can analyze and judge the input feature vector and output the corresponding classification result. Here, the quality detection classifier is designed as a binary classification model that can distinguish whether the smart toilet is qualified. Through training, the classifier can learn the characteristic differences between qualified and unqualified smart toilets, and accurately classify new smart toilets based on these characteristic differences in actual applications. The quality inspection classification results can directly reflect the quality status of smart toilets and are an important basis for evaluating product quality. Through the classification results, manufacturers can quickly understand the quality of smart toilets, which is conducive to improving the inspection efficiency and accuracy on the production line.

[0056] In summary, the quality inspection method for the production of smart toilets based on the embodiment of the present application is explained, which analyzes the infrared images of the seat ring of the smart toilet to be inspected at multiple time points in the temperature control performance test to understand the temperature control performance of the smart toilet to be inspected, analyzes the different spray patterns of the flushing nozzle in multiple modes of the smart toilet to be inspected in the cleaning performance test to understand the cleaning performance of the smart toilet to be inspected, and comprehensively considers the temperature control performance of the smart toilet to be inspected and the cleaning performance of the smart toilet to be inspected to judge whether the smart toilet to be inspected is qualified. In this way, a more comprehensive and intelligent inspection can be achieved, which is conducive to improving the accuracy and efficiency of smart toilet quality inspection.

[0057] Figure 7 FIG. 1 is a system block diagram of a quality inspection system for smart toilet production according to an embodiment of the present application. Figure 7 As shown, according to the embodiment of the present application, the quality inspection system 100 for the production of smart toilets includes: a smart toilet inspection-related data acquisition module 110, which is used to obtain the spray images of the cleaning nozzle in multiple modes of the smart toilet to be inspected in the cleaning performance test and the infrared images of the seat at multiple time points in the temperature control performance test; a smart toilet inspection-related data data processing module 120, which is used to respectively process the spray images of the cleaning nozzle in the multiple modes and the infrared images of the seat at the multiple time points to obtain the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map; a smart toilet detection feature generation module 130, which is used to fuse and feature filter the smart toilet cleaning performance detection feature map and the smart toilet temperature control performance detection feature map to obtain a smart toilet performance detection feature vector; a smart toilet detection result generation module 140, which is used to obtain the result of whether the smart toilet to be inspected is qualified based on the smart toilet performance detection feature vector.

[0058] Here, those skilled in the art will appreciate that the specific functions and operations of the various units and modules in the quality inspection system 100 for smart toilet production have been described in detail above. Figures 1 to 6 The quality inspection method for smart toilet production has been introduced in detail, and therefore, its repeated description will be omitted.

[0059] In summary, the quality inspection system 100 for smart toilet production based on the embodiment of the present application is explained, which analyzes the infrared images of the seat ring of the smart toilet to be inspected at multiple time points in the temperature control performance test to understand the temperature control performance of the smart toilet to be inspected, analyzes the different spray patterns of the flushing nozzle in multiple modes of the smart toilet to be inspected in the cleaning performance test to understand the cleaning performance of the smart toilet to be inspected, and comprehensively considers the temperature control performance of the smart toilet to be inspected and the cleaning performance of the smart toilet to be inspected to judge whether the smart toilet to be inspected is qualified. In this way, a more comprehensive and intelligent inspection can be achieved, which is conducive to improving the accuracy and efficiency of smart toilet quality inspection.

[0060] As described above, the quality inspection system 100 for smart toilet production according to the embodiment of the present application can be implemented in various wireless terminals, such as a server for quality inspection of smart toilet production. In one example, the quality inspection system 100 for smart toilet production according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the quality inspection system 100 for smart toilet production can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the quality inspection system 100 for smart toilet production can also be one of the many hardware modules of the wireless terminal.

[0061] Alternatively, in another example, the quality inspection system 100 for smart toilet production and the wireless terminal may also be separate devices, and the quality inspection system 100 for smart toilet production may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0062] In other embodiments of the present application, a smart toilet is also provided, wherein a seat cushion fixing pad paper is provided on the toilet seat of the smart toilet, and the seat cushion fixing pad paper is stacked and bonded in sequence. An easy-tear strip is provided on the side of each seat cushion fixing pad paper, and the tear strip head is inserted into the paginator, so that the user can press a button to separate the tear strip, and the user can tear it apart and throw it into the toilet after use. And after automatically tearing and throwing it into the toilet, it can also display in real time the number of remaining seat cushion fixing pad paper sheets and whether there are any damaged seat cushion fixing pad paper sheets. The stacked adhesive surface of the seat cushion fixing pad paper sheet has a certain degree of stickiness and is easily decomposed when exposed to water, while the other side that contacts the user has a certain degree of toughness and is waterproof.

Claims

1. A quality inspection method for intelligent toilet production, characterized in that: include: Obtaining the spray images of the cleaning nozzle of the intelligent toilet to be tested in multiple modes in the cleaning performance test and the infrared images of the seat ring at multiple time points in the temperature control performance test; Performing data processing on the spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram; The smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph are fused and feature-screened to obtain a smart toilet performance detection characteristic vector; Based on the smart toilet performance detection feature vector, a result of whether the smart toilet to be detected is qualified is obtained.

2. The quality inspection method for intelligent toilet production according to claim 1 is characterized in that: The spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points are respectively processed to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram, including: Extracting features of the spray images of the cleaning nozzles in the multiple modes to obtain a characteristic graph for detecting the cleaning performance of the smart toilet; Feature extraction is performed on the seat ring infrared images at the multiple time points to obtain the temperature control performance detection feature map of the smart toilet.

3. The quality inspection method for intelligent toilet production according to claim 2 is characterized in that: Feature extraction is performed on the spray images of the cleaning nozzles in the multiple modes to obtain the cleaning performance detection feature map of the smart toilet, including: The spray images of the cleaning nozzles in the multiple modes are respectively passed through the cleaning nozzle spray image encoder to obtain multiple intelligent toilet cleaning nozzle spray condition feature matrices; The plurality of smart toilet cleaning nozzle spray condition characteristic matrices are aggregated along the channel dimension to obtain the smart toilet cleaning performance detection characteristic graph.

4. The quality inspection method for intelligent toilet production according to claim 3 is characterized in that: Extracting features from the seat infrared images at the multiple time points to obtain a temperature control performance detection feature map of the smart toilet includes: Passing the seat infrared images at the multiple time points through the seat infrared image encoder respectively to obtain multiple intelligent toilet seat temperature feature matrices; The multiple smart toilet seat temperature feature matrices are aggregated along the channel dimension to obtain the smart toilet temperature control performance detection feature diagram.

5. The quality inspection method for intelligent toilet production according to claim 4 is characterized in that: The intelligent toilet cleaning performance detection characteristic graph and the intelligent toilet temperature control performance detection characteristic graph are fused and feature-screened to obtain an intelligent toilet performance detection characteristic vector, including: The smart toilet cleaning performance detection characteristic graph and the smart toilet temperature control performance detection characteristic graph are integrated to obtain a smart toilet performance detection fusion characteristic graph; Performing mean pooling of each feature matrix along the channel dimension on the smart toilet performance detection fusion feature map to obtain a smart toilet performance detection fusion feature vector; The smart toilet performance detection fusion feature map is passed through a performance detection key feature capturer to obtain a smart toilet performance detection fusion enhanced feature map; Performing mean pooling of each feature matrix along the channel dimension on the smart toilet performance detection fusion enhancement feature map to obtain a smart toilet performance detection fusion enhancement feature vector; The smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector are subjected to feature association coupling based on low-dimensional modulated perspective collaborative projection to obtain the smart toilet performance detection feature vector.

6. The quality inspection method for intelligent toilet production according to claim 5, characterized in that: The cleaning nozzle spray image encoder is a convolutional neural network model in which adjacent layers use mutually transposed convolution kernels, the seat ring infrared image encoder is a convolutional neural network model that uses a global transformation matrix as a filter, and the performance detection key feature capturer is a convolutional neural network model that uses a spatial attention mechanism and a channel attention mechanism.

7. The quality inspection method for intelligent toilet production according to claim 6, characterized in that: The intelligent toilet performance detection fusion feature vector and the intelligent toilet performance detection fusion enhancement feature vector are subjected to feature correlation coupling based on low-dimensional modulated perspective collaborative projection to obtain the intelligent toilet performance detection feature vector, including: Constructing a smart toilet performance detection distance topology matrix between the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhanced feature vector; Constructing a smart toilet performance detection eigenvalue granularity association matrix between the smart toilet performance detection fusion eigenvector and the smart toilet performance detection fusion enhanced eigenvector; Based on the smart toilet performance detection distance topological matrix, the smart toilet performance detection eigenvalue granularity association matrix is ​​topologically modulated to obtain a smart toilet performance detection topological modulation association matrix; Taking the smart toilet performance detection topological modulation association matrix as the perspective co-projection space, projecting the smart toilet performance detection fusion feature vector and the smart toilet performance detection fusion enhancement feature vector into the perspective co-projection space to obtain the perspective modulation smart toilet performance detection fusion feature vector and the perspective modulation smart toilet performance detection fusion enhancement feature vector; The visual angle modulation intelligent toilet performance detection fusion feature vector and the visual angle modulation intelligent toilet performance detection fusion enhancement feature vector are fused to obtain the intelligent toilet performance detection feature vector.

8. The quality inspection method for intelligent toilet production according to claim 7 is characterized in that: Based on the smart toilet performance detection feature vector, a result of whether the smart toilet to be detected is qualified is obtained, including: passing the smart toilet performance detection feature vector through a quality detection classifier to obtain a quality detection classification result, and the quality detection classification result is used to indicate whether the smart toilet to be detected is qualified.

9. A quality inspection system for intelligent toilet production, characterized in that: include: The intelligent toilet detection related data acquisition module is used to obtain the spray images of the cleaning nozzle of the intelligent toilet to be tested in multiple modes in the cleaning performance test and the infrared images of the seat ring at multiple time points in the temperature control performance test; A data processing module for intelligent toilet detection related data, used to respectively process the spray images of the cleaning nozzles in the multiple modes and the infrared images of the seat ring at the multiple time points to obtain a smart toilet cleaning performance detection characteristic diagram and a smart toilet temperature control performance detection characteristic diagram; A smart toilet detection feature generation module, used for fusing and feature screening the smart toilet cleaning performance detection feature graph and the smart toilet temperature control performance detection feature graph to obtain a smart toilet performance detection feature vector; The smart toilet detection result generation module is used to obtain the result of whether the smart toilet to be detected is qualified based on the smart toilet performance detection feature vector.

10. The quality inspection system for intelligent toilet production according to claim 9, characterized in that: The smart toilet detection result generation module is used to: pass the smart toilet performance detection feature vector through a quality detection classifier to obtain a quality detection classification result, and the quality detection classification result is used to indicate whether the smart toilet to be detected is qualified.

Images