Testing device and method for automatically detecting air tightness of bathroom towel rack product

Abstract

The invention relates to the field of bathroom towel racks, and particularly discloses a testing device and method for automatically detecting the air tightness of a bathroom towel rack product, and the method comprises the steps: firstly obtaining a bathroom towel rack air tightness testing image, then carrying out the feature extraction and correlation analysis of the image through a deep learning technology, and finally obtaining a classification result through a classifier. Therefore, manual errors caused by human eye observation are reduced, the quality consistency and the qualified rate of products are ensured, and the production efficiency and the product reliability are improved.

Description

Technical Field

[0001] The present application relates to the field of bathroom towel racks, and more specifically, to a testing device and method for automatically detecting the air tightness of bathroom towel rack products. Background Art

[0002] Bathroom towel racks are a common bathroom accessory designed to hang towels and bath towels to keep the bathroom clean and tidy. They are usually made of metal, stainless steel or plastic, and are rust-proof and water-resistant. Towel racks come in a variety of styles, such as wall-mounted, vertical and folding, to meet different space requirements. They not only make it convenient to dry and take towels, but also add beauty to the bathroom space and improve convenience.

[0003] At present, the inspection equipment and methods for bathroom towel racks on the market mainly use the air pressure test method, which is to fill the product cavity with high-pressure gas, place it in the sink, and observe whether there are bubbles by human eyes to judge whether the product has leakage. Whether the sealing meets the standard is the most important technical requirement for bathroom towel racks and radiators, but human visual observation is prone to produce tiny bubbles that are difficult to detect by human eyes, and bubbles are missed due to manual fatigue. The reliability is poor, and misjudgment often leads to unqualified products flowing into the market.

[0004] Therefore, a testing device and method for automatically detecting the air tightness of a bathroom towel rack product is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a test device and method for automatically detecting the air tightness of a bathroom towel rack product, which first obtains the air tightness test image of the bathroom towel rack, then uses deep learning technology to perform feature extraction and association analysis, and finally uses a classifier to obtain a classification result to determine whether the bathroom towel rack product air tightness test is qualified, thereby reducing the artificial error caused by human eye observation, ensuring the quality consistency and pass rate of the product, and improving production efficiency and product reliability.

[0006] According to one aspect of the present application, a testing device for automatically detecting air tightness of a bathroom towel rack product is provided, comprising:

[0007] Bathroom towel rack air tightness data acquisition module, used to obtain bathroom towel rack air tightness test images;

[0008] A bathroom towel rack air tightness data processing module, used for extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image;

[0009] The bathroom towel rack product air tightness test analysis module is used to determine whether the bathroom towel rack product air tightness test is qualified based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector.

[0010] According to another aspect of the present application, a method for automatically detecting air tightness of a bathroom towel rack product is provided, which comprises:

[0011] Get the air tightness test image of bathroom towel rack;

[0012] Extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image;

[0013] Based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector, it is judged whether the bathroom towel rack product air tightness test is qualified.

[0014] Compared with the prior art, the present application provides a testing device and method for automatically detecting the air tightness of a bathroom towel rack product. The device and method first obtain an air tightness test image of the bathroom towel rack, then use deep learning technology to perform feature extraction and association analysis, and finally use a classifier to obtain a classification result to determine whether the bathroom towel rack product has passed the air tightness test, thereby reducing the artificial errors caused by human eye observation, ensuring the quality consistency and pass rate of the product, and improving production efficiency and product reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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.

[0016] Figure 1 This is a block diagram of a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application.

[0017] Figure 2 The present invention is a block diagram of a bathroom towel rack air tightness data processing module in a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application.

[0018] Figure 3 The present invention is a block diagram of a bathroom towel rack air tightness feature coding unit in a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application.

[0019] Figure 4This is a block diagram of an air tightness test and analysis module for a bathroom towel rack product in a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application.

[0020] Figure 5 The present invention is a flow chart of a method for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] 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.

[0023] Figure 1 This is a block diagram of a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. Figure 1 As shown, a testing device 100 for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application includes: a bathroom towel rack air tightness data acquisition module 110, used to obtain an air tightness test image of the bathroom towel rack; a bathroom towel rack air tightness data processing module 120, used to extract a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image; a bathroom towel rack product air tightness test analysis module 130, used to determine whether the bathroom towel rack product air tightness test is qualified based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector.

[0024] In the above-mentioned test device 100 for automatically detecting air tightness of bathroom towel rack products, the bathroom towel rack air tightness data acquisition module is used to obtain the bathroom towel rack air tightness test image. It should be understood that the bathroom towel rack is a common accessory in the bathroom, which is used to hang towels and bath towels to keep the bathroom environment clean and tidy. It is usually made of metal, stainless steel or plastic materials and has rust-proof and water-resistant properties. According to different usage requirements, the towel rack has a variety of designs such as wall-mounted, vertical and folding to adapt to various space layouts. In addition to facilitating the hanging and taking of towels, it can also enhance the beauty and practicality of the bathroom space. In the current market, the air tightness detection of bathroom towel racks mainly relies on the air pressure test method. This method is to fill the towel rack with high-pressure gas, then immerse it in the sink, and judge whether there is a leak by observing whether there are bubbles. However, this method of relying on manual observation has certain limitations. Because it is difficult for the human eye to identify tiny bubbles, and the observation process is easily affected by manual fatigue, some bubbles may be missed, resulting in inaccurate test results. This instability sometimes causes defective products to still flow into the market, affecting the overall quality and reliability of the products. Therefore, in the technical solution of this application, the air tightness test image of the bathroom towel rack is obtained, and combined with deep learning technology, the bathroom towel rack is tested for air tightness to determine whether it meets the standards, so as to effectively reduce the errors caused by manual observation. This method ensures the consistency and pass rate of product quality, while improving production efficiency and product reliability.

[0025] Specifically, the air tightness test images can intuitively show how the towel rack performs during the test. Through the images, you can clearly see the status of each part of the towel rack during the sealing test, and understand whether there are air leaks, water seepage or other poor sealing problems. By comparing test images from different batches or different time periods, manufacturers can observe any quality fluctuations in the production process. The accumulation and analysis of this data helps to optimize the production process and improve product design, thereby improving the quality and consistency of the overall product. Through these images, both manufacturers and consumers can understand and evaluate the products more comprehensively, thereby promoting continuous improvement of product quality.

[0026] In the above-mentioned test device 100 for automatically detecting air tightness of bathroom towel rack products, the bathroom towel rack air tightness data processing module is used to extract bathroom towel rack air tightness test tracking feature vectors and bathroom towel rack air tightness test global feature vectors from the bathroom towel rack air tightness test image. It should be understood that the air tightness performance of the towel rack can be more comprehensively evaluated by combining local and global information. Among them, the local feature vector provides insight at the detail level to help identify specific problems. The global feature vector provides a macro perspective to ensure the stability and consistency of the overall design and production process. Such multi-level analysis can greatly improve the accuracy and reliability of detection, thereby improving the overall quality and market competitiveness of the product.

[0027] Figure 2 FIG. 1 is a block diagram of a bathroom towel rack air tightness data processing module in a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, the bathroom towel rack air tightness data processing module 120 includes: a bathroom towel rack air tightness target feature extraction unit 121, which is used to perform target feature extraction on the bathroom towel rack air tightness test image to obtain a bathroom towel rack air tightness test interest region; a bathroom towel rack air tightness feature encoding unit 122, which is used to perform feature encoding on the bathroom towel rack air tightness test interest region to obtain the bathroom towel rack air tightness test tracking feature vector; a bathroom towel rack air tightness attention feature extraction unit 123, which is used to perform attention feature extraction on the bathroom towel rack air tightness test image to obtain the bathroom towel rack air tightness test global feature vector.

[0028] It should be understood that in the air tightness test, specific parts of the towel rack (such as joints, interfaces and surfaces) are usually potential leak points. Through feature extraction, the system is able to identify these key areas and focus the inspection on possible defect locations, thereby reducing unnecessary image processing workload. This targeted analysis helps to locate and evaluate problems faster and improve inspection efficiency.

[0029] Furthermore, feature encoding can provide useful data support for subsequent analysis and processing by converting visual information in the region of interest into computable feature vectors. These feature vectors usually contain information about shape, texture, color and other image features, allowing the system to accurately quantify and analyze key areas in the image.

[0030] Furthermore, attention feature extraction technology provides an efficient and effective way to obtain and utilize global features by focusing on the most important parts of the image, thereby ensuring the airtightness quality of the product. Specifically, attention feature extraction can highlight the most relevant parts of the image and ignore irrelevant background noise and interference. By applying the attention mechanism, the system can automatically focus on those areas that are closely related to airtightness, such as seams, connection parts, etc. This selective attention can reduce the waste of computing resources while improving the efficiency and effectiveness of feature extraction.

[0031] In a specific embodiment of the present application, the bathroom towel rack air tightness target feature extraction unit 121 includes: passing the bathroom towel rack air tightness test image through a bathroom towel rack air tightness test convolutional neural network model based on a mixed convolutional layer to obtain a bathroom towel rack air tightness test feature map; passing the bathroom towel rack air tightness test feature map through a bathroom towel rack air tightness test target detection network model to obtain the bathroom towel rack air tightness test area of ​​interest.

[0032] It should be understood that the hybrid convolution layer combines the characteristics of multiple convolution kernels in the convolutional neural network, such as standard convolution, deep convolution, and separable convolution. This hybrid convolution strategy can capture features of different scales and levels in the image. For example, standard convolution can extract basic texture and edge information, while deep convolution can better handle local details and complex patterns. By combining these different convolution operations, the network can fully understand the important information in the image and improve the accuracy and meticulousness of feature extraction. Specifically, each hybrid convolutional layer of the bathroom towel rack air tightness test convolutional neural network model based on the hybrid convolutional layer includes a parallel first convolution branch structure, a second convolution branch structure, a third convolution branch structure and a fourth convolution branch structure, and a multi-scale fusion structure connected to the first to fourth convolution branch structures, wherein the first convolution branch uses a first convolution kernel having a first size, the second convolution branch uses a second convolution kernel having a first size and a first void ratio, the third convolution branch uses a third convolution kernel having a first size and a second void ratio, and the fourth convolution branch uses a fourth convolution kernel having a first size and a third void ratio. More specifically, each hybrid convolutional layer of the bathroom towel rack air tightness test convolutional neural network model based on the hybrid convolutional layer performs the following operations on the input data in the forward pass of the layer: using a first convolutional kernel with a first size to perform convolution encoding on the bathroom towel rack air tightness test image to obtain a first scale feature map; using a second convolutional kernel with a first void rate to perform convolution encoding on the bathroom towel rack air tightness test image to obtain a second scale feature map; using a third convolutional kernel with a second void rate to perform convolution encoding on the bathroom towel rack air tightness test image to obtain a third scale feature map; using a fourth convolutional kernel with a third void rate to perform convolution encoding on the bathroom towel rack air tightness test image to obtain a fourth scale feature map. Feature map, wherein the first convolution kernel, the second convolution kernel, the third convolution kernel and the fourth convolution kernel have the same size, and the second convolution kernel, the third convolution kernel and the fourth convolution kernel have different void rates; the first scale feature map, the second scale feature map, the third scale feature map and the fourth scale feature map are aggregated along the channel dimension to obtain an aggregated feature map; the aggregated feature map is pooled to generate a pooled feature map; and the pooled feature map is activated to generate an activated feature map; wherein the output of the last layer of the bathroom towel rack air tightness test convolutional neural network model based on the hybrid convolution layer is the bathroom towel rack air tightness test feature map.

[0033] Furthermore, processing the feature map of the air tightness test of the bathroom towel rack through the target detection network model to obtain the region of interest in the air tightness test is a key step to achieve accurate and efficient detection. Specifically, the feature map provides deep information extracted from the original image, including texture, shape, and structure. These features are essential for identifying and analyzing key parts in the air tightness test. By using the target detection network model, possible defective areas in the image, such as poor sealing at the joint or bubbles on the surface, can be automatically identified from these feature maps. The target detection network model is able to analyze these complex features and identify target areas related to air tightness, improving the accuracy and comprehensiveness of the detection. Here, the target detection network model usually includes multiple stages, such as region proposal, target classification, and bounding box regression. These stages work together to extract the most representative area from the feature map and determine its specific location and size. In this way, the system can accurately locate the areas that need to be focused on in the air tightness test. Specifically, the bathroom towel rack air tightness test target detection network model is used to process the bathroom towel rack air tightness test feature map with the following target detection formula to obtain the bathroom towel rack air tightness test interest region;

[0034] Among them, the target detection formula is:

[0035] ROI=(cls(ψ det ,B),Regr(ψ det ,B))

[0036] Among them, ψ det is the feature map of the air tightness test of the bathroom towel rack, B is the anchor frame, ROI is the region of interest for the air tightness test of the bathroom towel rack, cls(ψ det ,B) indicates classification, Regr(ψ det ,B) indicates regression.

[0037] Figure 3 FIG. 1 is a block diagram of an air tightness feature coding unit of a bathroom towel rack in a testing device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the bathroom towel rack air tightness feature encoding unit 122 includes: a bathroom towel rack air tightness test significant target encoding subunit 1221, which is used to pass the bathroom towel rack air tightness test interest area through a bathroom towel rack air tightness test significant target detector to obtain a bathroom towel rack air tightness test tracking feature map; a bathroom towel rack air tightness test pooling subunit 1222, which is used to perform global mean pooling on the bathroom towel rack air tightness test tracking feature map to obtain the bathroom towel rack air tightness test tracking feature vector.

[0038] It should be understood that the salient object detector is designed to extract and enhance the most eye-catching areas from the image, which usually contain the most informative features. In the airtightness test, the areas of interest may include the connection parts of the towel rack, the sealing surface, and other areas that may affect the airtightness. Through the processing of the salient object detector, these key areas will be automatically identified and highlighted, so that subsequent analysis can be more focused on these important parts, thereby improving the accuracy of the detection. Specifically, each layer of the bathroom towel rack air tightness test salient target detector performs the following operations on the input data in the forward pass of the layer: convolution processing is performed on the input data using a first convolution kernel to obtain a first convolution feature map; convolution processing is performed on the first convolution feature map using a second convolution kernel to obtain a second convolution feature map, wherein the size of the first convolution kernel is larger than the size of the second convolution kernel; pooling processing is performed on the second convolution feature map to obtain a pooling feature map; and activation processing is performed on the pooling feature map to obtain an activation feature map; wherein the output of the last layer of the bathroom towel rack air tightness test salient target detector is the bathroom towel rack air tightness test tracking feature map, and the input of the first layer of the bathroom towel rack air tightness test salient target detector is the bathroom towel rack air tightness test region of interest.

[0039] Furthermore, global mean pooling can effectively summarize the global features of the image and improve the performance and efficiency of the detection system by compressing the information in the feature map into a vector of fixed length. Specifically, global mean pooling is an operation that averages all pixel values ​​in the feature map. This process maps the feature map from a high-dimensional space to a low-dimensional space and generates a fixed-length feature vector. This method not only reduces the data dimension, but also retains the global information of the image.

[0040] In a specific embodiment of the present application, the bathroom towel rack air tightness test pooling subunit 1222 includes: creating a Java class of TestBathTowelRackAirTightnessFeatureExtractor and introducing a JUnit library into the project; defining a method testGlobalAveragePooling in the Java class of TestBathTowelRackAirTightnessFeatureExtractor for testing the global average pooling function; creating a Java class of GlobalAveragePooling for implementing the logic of global average pooling, that is, performing mean calculation on the input bathroom towel rack air tightness test tracking feature map to obtain a bathroom towel rack air tightness test tracking feature vector; in the TestBathTowelRackAirTightnessFeatureExtractor class, writing test code, calling the global average pooling method of the GlobalAveragePooling class, and verifying whether its output meets expectations, using JUnit to run the test class to ensure that all test cases pass; and deploying the implemented global average pooling function and test code to the production environment.

[0041] In this way, by using the JUnit library in the Java environment, the logic of global mean pooling is implemented, and its correctness is verified through test code, and finally the code is deployed to the production environment.

[0042] Part of the deployment code is shown below.

[0043] The GlobalAveragePooling deployment code is as follows.

[0044]

[0045]

[0046] It should be understood that here, the Java programming language and JUnit testing framework are used to ensure the efficiency and accuracy of feature map processing. Among them, GlobalAveragePooling.java implements the core logic of global average pooling. The input is a two-dimensional array (feature map) and the output is a one-dimensional array (feature vector). The globalAveragePooling method traverses the input feature map, calculates the mean of each column, and stores the result in the feature vector. The design of this method ensures that the mean of each column of the feature map is accurately calculated, thereby obtaining a vector representing the overall feature. TestBathTowelRackAirTightnessFeatureExtractor uses the JUnit framework to write test cases to verify the functionality of the GlobalAveragePooling class. The testGlobalAveragePooling method defines a 3x3 feature map and expects its feature vector after global average pooling to be [4.0, 5.0, 6.0]. By calling the GlobalAveragePooling.globalAveragePooling method and using the assertArrayEquals assertion, it is verified whether the feature vector actually output is consistent with the expectation.

[0047] In this way, the main effects of deploying the code are reflected in the following aspects. a. Efficiency. The global mean pooling method implemented in Java language has high computational efficiency. Java's strong typing and optimization compiler ensure the execution speed of the code, which is suitable for large-scale data processing. The use of the JUnit test framework automates the code testing and verification process, improving development efficiency. b. Accuracy. The design of the global mean pooling method ensures that the mean of each column of the feature map is accurately calculated, thereby obtaining a vector representing the overall feature. Through the assertion function of the JUnit test framework, it is possible to accurately verify whether the output of the method meets expectations and ensure the correctness of the code. c. Maintainability. The code structure is clear and divided into implementation classes and test classes, which is convenient for subsequent maintenance and expansion. The use of the JUnit test framework standardizes the code testing and verification process, which is convenient for team collaboration and code review. d. Scalability. The design of the global mean pooling method takes into account the universality of the input feature map and can be applied to feature maps of different sizes. Through the JUnit test framework, new test cases can be easily added to verify the performance of the method in different scenarios. e. Deployment convenience. The code deployment process is simple and clear. You only need to deploy the implementation class and test class to the production environment. The use of the JUnit test framework ensures comprehensive verification of the code before deployment, reducing the risk in the production environment.

[0048] Through the combination of Java and JUnit, the global mean pooling method of the tracking feature map of the bathroom towel rack air tightness test is realized. This method is not only efficient and accurate, but also has good maintainability and scalability, providing reliable technical support for the bathroom towel rack air tightness test.

[0049] In a specific embodiment of the present application, the bathroom towel rack air tightness attention feature extraction unit 123 includes: denoising the bathroom towel rack air tightness test image to obtain a bathroom towel rack air tightness test noise reduction and enhanced image; passing the bathroom towel rack air tightness test noise reduction and enhanced image through a bathroom towel rack air tightness test enhanced convolutional neural network based on a spatial attention mechanism to obtain a bathroom towel rack air tightness test noise reduction and enhanced spatial attention feature map; passing the bathroom towel rack air tightness test noise reduction and enhanced spatial attention feature map through a bathroom towel rack air tightness test feature extractor with multi-feature extraction capabilities to obtain the bathroom towel rack air tightness test global feature vector.

[0050] It should be understood that in the actual detection process, image noise may come from various factors, such as changes in the lighting of the shooting environment, interference of the camera sensor, or loss of image compression. These noises will affect the clarity of the image and the accuracy of feature recognition. Image denoising aims to remove random noise in the image, thereby improving the quality of the image. Noise usually manifests itself as irregular brightness changes or color deviations, which may mask or interfere with important details in the image. In the air tightness test, these details include key areas such as the connection parts and sealing surfaces of the towel rack. Through denoising, these interferences can be reduced, making the important features in the image clearer, thereby improving the accuracy of subsequent detection. Specifically, the bathroom towel rack air tightness test image is input into the encoder of the denoising module, wherein the encoder uses a convolution layer to perform explicit spatial encoding on the bathroom towel rack air tightness test image to obtain image features; and the image features are input into the decoder of the denoising module, wherein the decoder uses a deconvolution layer to perform deconvolution processing on the image features to obtain the bathroom towel rack air tightness test denoising enhanced image.

[0051] Furthermore, the denoised and enhanced image of the air tightness test of the bathroom towel rack is processed by a convolutional neural network (CNN) based on the spatial attention mechanism to obtain the denoised and enhanced spatial attention feature map, which is an important method to improve the accuracy of image analysis and feature recognition ability. The denoised and enhanced image has reduced the interference noise in the image and made the image details clearer, but how to extract the most valuable information from these details remains a challenge. The spatial attention mechanism dynamically adjusts the weight of the focus area, allowing the network to focus on the most critical part of the image, thereby improving the effect of feature extraction. This mechanism can ensure that the model better captures the details related to air tightness when processing complex air tightness test images. The spatial attention feature map provides a new feature representation method. Specifically, the convolution coding part of the bathroom towel rack air tightness test enhanced convolution neural network based on the spatial attention mechanism is used to perform deep convolution coding on the bathroom towel rack air tightness test denoising enhanced image to obtain an initial convolution feature map; the initial convolution feature map is input into the spatial attention part of the bathroom towel rack air tightness test enhanced convolution neural network based on the spatial attention mechanism to obtain a spatial attention map; the spatial attention map is passed through a Softmax activation function to obtain a spatial attention feature map; the spatial attention feature map and the initial convolution feature map are calculated by point-by-point multiplication to obtain the bathroom towel rack air tightness test denoising enhanced spatial attention feature map.

[0052] Furthermore, the denoising enhanced spatial attention feature map has optimized the focus on important areas through the spatial attention mechanism, but these feature maps may still contain rich multi-dimensional information. The multi-feature extractor can capture more detailed and comprehensive features in the image by extracting features from different angles and scales. This diverse feature extraction capability enables the detection system to better identify and analyze complex details related to air tightness. Specifically, the bathroom towel rack air tightness test feature extractor with multi-feature extraction capability includes: a first convolution layer, a second convolution layer parallel to the first convolution layer, and a cascade layer connected to the first convolution layer and the second convolution layer, wherein the first convolution layer uses a one-dimensional convolution kernel with a first scale, and the second convolution layer uses a one-dimensional convolution kernel with a second scale. More specifically, the first convolutional layer of the bathroom towel rack air tightness test feature extractor with multi-feature extraction capability is used to perform convolution encoding on the bathroom towel rack air tightness test denoising enhanced spatial attention feature map to obtain a first-scale feature vector; the second convolutional layer of the bathroom towel rack air tightness test feature extractor with multi-feature extraction capability is used to perform convolution encoding on the bathroom towel rack air tightness test denoising enhanced spatial attention feature map to obtain a second-scale feature vector; the first-scale feature vector and the second-scale feature vector are cascaded to obtain the bathroom towel rack air tightness test global feature vector.

[0053] In the above-mentioned test device 100 for automatically detecting air tightness of bathroom towel rack products, the bathroom towel rack product air tightness test analysis module 130 is used to determine whether the bathroom towel rack product air tightness test is qualified based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector. It should be understood that in the air tightness test, specific parts such as the sealing strip and connection parts of the towel rack may be key areas for evaluating air tightness. The tracking feature vector can accurately capture these local details by analyzing the features of these specific areas. Using the tracking feature vector, the detection system can identify and evaluate the air tightness performance of these key areas to ensure that no important information is missed during the test. The global feature vector integrates the comprehensive information of the entire image and provides a global perspective. The air tightness test not only depends on local details, but also requires a general understanding of the air tightness status of the entire product. The global feature vector generates a comprehensive representation by integrating all local features to help the detection system understand the air tightness status of the entire product. By fusing these two feature vectors, the detection system can more comprehensively evaluate airtightness issues, ensuring that the detection results are not only accurate in local areas but also reflect the overall situation.

[0054] Figure 4 FIG. 1 is a block diagram of an air tightness test and analysis module for a bathroom towel rack product in a test device for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. Figure 4 As shown, in a specific embodiment of the present application, the bathroom towel rack product air tightness test analysis module 130 includes: an air tightness test feature fusion unit 131, which is used to fuse the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector to obtain an air tightness test result judgment feature vector; an air tightness test feature optimization unit 132, which is used to adjust the air tightness test result judgment feature vector based on the time entropy change of the entropy degradation modulation value limit standard to obtain an optimized air tightness test result judgment feature vector; an air tightness test result judgment unit 133, which is used to pass the optimized air tightness test result judgment feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the bathroom towel rack product air tightness test is qualified.

[0055] It should be understood that the fusion process combines local and global information to form a unified judgment feature vector. This fused feature vector combines detailed local analysis and comprehensive overall evaluation, allowing the detection system to make accurate judgments at both the detail and overall levels. By combining local and global information, the detection system can better adapt to different test conditions and environmental changes, reducing the risk of misjudgment and missed judgment. This approach ensures that the system can maintain high accuracy and stability even under complex or non-ideal test conditions.

[0056] In particular, in the technical solution of the present application, the airtightness test process involves multiple feature extraction steps, including feature maps generated from the original image, detection features of the region of interest, and features obtained by noise reduction and spatial attention mechanisms. These features may contain overlapping information. For example, when certain areas in the image are extracted through multiple steps, similar airtightness-related information may be captured, thereby forming redundancy. In this way, there may be a large amount of duplicate data in the final fused airtightness test result judgment feature vector, making it difficult for the classifier to effectively identify the key features that are truly helpful in distinguishing between qualified and unqualified. Further, since similar features are introduced multiple times, the decision boundary of the classifier becomes blurred, which may lead to a decrease in the distinction between different categories (for example, qualified and unqualified). In this case, some samples that should have been clearly classified may be misclassified due to the overlap of features. This ambiguity not only reduces the accuracy of the classifier, but may also cause false positives or negatives in the production environment, affecting product quality control. Therefore, in the technical solution of the present application, the airtightness test result judgment feature vector is adjusted by adjusting the time entropy change based on the entropy degradation modulation value limit standard to obtain an optimized airtightness test result judgment feature vector.

[0057] Among them, the airtightness test result judgment feature vector is subjected to time entropy change adjustment based on the entropy degradation modulation value limit standard to obtain the optimized airtightness test result judgment feature vector, including: extracting each row vector in the prediction parameter space of the classifier to obtain a set of airtightness test prediction parameter row vectors; calculating the airtightness test class probability entropy degradation simulation value between the airtightness test result judgment feature vector and each airtightness test prediction parameter row vector in the set of airtightness test prediction parameter row vectors to obtain a set of airtightness test class probability entropy degradation simulation values; calculating the square root of the sum of squares of the airtightness test class probability entropy degradation simulation values ​​to obtain an airtightness test entropy degradation modulation value; based on the airtightness test entropy degradation modulation value, performing feature distribution modulation on the airtightness test result judgment feature vector to obtain the optimized airtightness test result judgment feature vector.

[0058] The optimization formula is:

[0059] Dj =Softmax(M m v c +H m p j +b m )

[0060]

[0061] Among them, P represents the set of row vectors of airtightness test prediction parameters, p1, p2, p j 、p n represents the first, second, jth and nth airtightness test prediction parameter row vectors in the set of airtightness test prediction parameter row vectors, v c represents the airtightness test result judgment feature vector, Softmax represents the normalized exponential function, M m represents the first weight matrix, H m represents the second weight matrix, b m represents the bias vector, D j represents the jth airtightness test class probability entropy degradation simulation value of the set of airtightness test class probability entropy degradation simulation values, t represents the airtightness test entropy degradation modulation value, v c ⊙t Indicates the calculation of the airtightness test result judgment feature vector v c Each eigenvalue of the t-th power, represents the i-th eigenvalue of the airtightness test result judgment eigenvector, α represents a predetermined hyperparameter, represents addition by position, and V' represents the optimized airtightness test result judgment feature vector.

[0062] Specifically, in the technical solution of the present application, the airtightness test result judgment feature vector has feature redundancy, which leads to the problem of inter-category squeeze relative to the class label of the classifier when the classifier is used for classification judgment, thus affecting the classification iteration effect.

[0063] Based on this, in the technical solution of the present application, the entropy degradation of the airtightness test result judgment feature vector based on the prediction parameter space of the classifier is maximized. First, each row vector is extracted from the prediction parameter space of the classifier to form a row vector set. Then, the airtightness test class probability entropy degradation simulation value between the airtightness test result judgment feature vector and each row vector in the above row vector set is calculated to measure the uncertainty or information entropy change of the airtightness test result judgment feature vector under given parameters. Then, all the airtightness test class probability entropy degradation simulation values ​​are squared and summed, and then the square root is taken to obtain a comprehensive airtightness test entropy degradation modulation value, which reflects the degree of entropy change of the entire system when processing the feature vector. Finally, the airtightness test result judgment feature vector is adjusted according to the airtightness test entropy degradation modulation value obtained above. Specifically, the attribute set of the airtightness test result judgment feature vector is represented in a power series based on the airtightness test entropy degradation modulation value, and a set aggregated bounded deviation represented by an aggregation radius based on each attribute value of the airtightness test result judgment feature vector is added, so as to perform attribute selection and sparse optimization on the airtightness test result judgment feature vector under a unified airtightness test entropy degradation modulation value restriction standard, so as to ensure that the airtightness test result judgment feature vector contains sufficient data but is not too cumbersome so as to introduce unnecessary interference or redundancy.

[0064] Furthermore, the optimized air tightness test result judgment feature vector is a comprehensive feature representation that integrates local and global information. It can comprehensively reflect the air tightness status of the towel rack, including the details of key areas and the overall performance of the entire product. However, these feature vectors usually contain a lot of information, and directly judging them may be complicated and unintuitive. Therefore, these feature vectors need to be processed by a classifier to generate clear test results. A classifier is a machine learning model that can map the input optimized air tightness test result judgment feature vector to predefined categories. Specifically, in the air tightness test, the role of the classifier is to classify the test results into "qualified" or "unqualified" categories based on the value of the optimized air tightness test result judgment feature vector. For example, if the classifier classifies the feature vector as "qualified", it means that the air tightness of the bathroom towel rack meets the standard; if it is classified as "unqualified", it means that there is an air tightness problem and further processing or improvement is required.

[0065] In another embodiment of the present application, a bathroom towel rack air tightness test equipment is provided. It should be understood that the bathroom towel rack product is tested underwater, and its main function is to monitor the air tightness index of water heating products online in real time. When the HVAC product is tested under water, the bubbles caused by poor air tightness are accurately captured. This equipment is suitable for electroplating pools with good water clarity, and can monitor a depth of 600 to 1200 mm. Lasers are combined with industrial cameras for image recognition, and bubble recognition algorithms are developed for accurate capture, as well as overall control application software to complete the entire process of air tightness testing. The entire product involves multi-dimensional comprehensive technologies such as optical design, visual algorithms, electrical control, and structural design. The system composition of this equipment consists of five parts: electroplating pool, electronic control system, laser light source, visual sensor, and central control software. Among them, 2 industrial visual sensors, 4 laser light sources, 1 electronic control cabinet, 1 industrial computer, 1 application software and software instruction manual, electroplating pool and some control electrical appliances are prepared.

[0066] Specifically, this device provides a bubble monitoring function, where a single system provides two sets of sensors (dual screen) monitoring. Users can query the original image of the bubble result for retrospective review. In addition, users can set monitoring condition parameters according to actual working conditions.

[0067] More specifically, the operation logic flow of the device is as follows: (1) Place the workpiece; (2) Press the start button; (3) Turn on the vision; (4) Turn off the lighting; (5) Wait for the bubbles to disappear; (6) Open the air valve; (7) Start monitoring; (8) Continue monitoring for a specified period of time; (9) Output the results; (10) Close the air valve and take the material.

[0068] The performance parameters can be set as follows: (1) Bubbles with a diameter greater than 1 mm can be monitored (2) Bubbles with a diameter less than 1 mm but a cluster width greater than 2 mm can be monitored (3) Non-continuous bubbles can be monitored (with a longer interval, which can be set by parameters) (4) Bubbles within a depth range of 100 to 1200 mm can be monitored under conditions where clarity allows (5) Monitorable area: 1900 mm × 800 mm × 20 mm (length × width × height).

[0069] In summary, the embodiment of the present application first obtains the air tightness test image of the bathroom towel rack, then uses deep learning technology to perform feature extraction and association analysis, and finally obtains the classification result through a classifier to determine whether the air tightness test of the bathroom towel rack product is qualified, thereby reducing the artificial errors caused by human eye observation, ensuring the quality consistency and pass rate of the product, and improving production efficiency and product reliability.

[0070] As described above, the test device 100 for automatically detecting air tightness of bathroom towel rack products according to the embodiment of the present application can be implemented in various terminal devices. In one example, the test device 100 for automatically detecting air tightness of bathroom towel rack products can be integrated into the terminal device as a software module and / or a hardware module. For example, the test device 100 for automatically detecting air tightness of bathroom towel rack products can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the test device 100 for automatically detecting air tightness of bathroom towel rack products can also be one of the many hardware modules of the terminal device.

[0071] Alternatively, in another example, the testing device 100 for automatically detecting air tightness of bathroom towel rack products and the terminal device may be separate devices, and the testing device 100 for automatically detecting air tightness of bathroom towel rack products may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0072] Figure 5 Flow chart of a method for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application. Figure 5 As shown, a test method for automatically detecting air tightness of a bathroom towel rack product according to an embodiment of the present application includes: S110, obtaining an air tightness test image of the bathroom towel rack; S120, extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image; S130, judging whether the bathroom towel rack product air tightness test is qualified based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector.

[0073] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned test method for automatically detecting air tightness of bathroom towel rack products have been referred to above. Figures 1 to 4 The description of the testing device for automatically detecting air tightness of bathroom towel rack products has been introduced in detail, and therefore, its repeated description will be omitted.

[0074] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a digital video disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).

[0075] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0076] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0077] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0078] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0079] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A test device for automatically detecting the air tightness of bathroom towel rack products, characterized in that: include: Bathroom towel rack air tightness data acquisition module, used to obtain bathroom towel rack air tightness test images; A bathroom towel rack air tightness data processing module, used for extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image; The bathroom towel rack product air tightness test analysis module is used to determine whether the bathroom towel rack product air tightness test is qualified based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector.

2. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 1 is characterized in that: The bathroom towel rack air tightness data processing module comprises: A bathroom towel rack air tightness target feature extraction unit, used for performing target feature extraction on the bathroom towel rack air tightness test image to obtain a bathroom towel rack air tightness test interest region; A bathroom towel rack air tightness feature encoding unit, used for performing feature encoding on a region of interest for the bathroom towel rack air tightness test to obtain a tracking feature vector for the bathroom towel rack air tightness test; The bathroom towel rack air tightness attention feature extraction unit is used to extract attention features from the bathroom towel rack air tightness test image to obtain the bathroom towel rack air tightness test global feature vector.

3. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 2 is characterized in that: The bathroom towel rack air tightness target feature extraction unit comprises: The bathroom towel rack air tightness test image is passed through a bathroom towel rack air tightness test convolutional neural network model based on a hybrid convolutional layer to obtain a bathroom towel rack air tightness test feature map; The bathroom towel rack air tightness test feature map is passed through the bathroom towel rack air tightness test target detection network model to obtain the bathroom towel rack air tightness test interest region.

4. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 3 is characterized in that: The bathroom towel rack air tightness characteristic coding unit comprises: A bathroom towel rack air tightness test significant target encoding subunit, used for passing the bathroom towel rack air tightness test interest region through a bathroom towel rack air tightness test significant target detector to obtain a bathroom towel rack air tightness test tracking feature map; The bathroom towel rack air tightness test pooling subunit is used to perform global mean pooling on the bathroom towel rack air tightness test tracking feature map to obtain the bathroom towel rack air tightness test tracking feature vector.

5. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 4 is characterized in that: The bathroom towel rack air tightness test pool subunit includes: Create the TestBathTowelRackAirTightnessFeatureExtractor Java class and import the JUnit library into the project. Define the method testGlobalAveragePooling in the Java class of TestBathTowelRackAirTightnessFeatureExtractor to test the global average pooling function; Create a Java class called GlobalAveragePooling to implement the logic of global average pooling, that is, to calculate the average of the input bathroom towel rack air tightness test tracking feature map to obtain the bathroom towel rack air tightness test tracking feature vector; In the TestBathTowelRackAirTightnessFeatureExtractor class, write test code to call the global average pooling method of the GlobalAveragePooling class and verify whether its output meets expectations. Use JUnit to run the test class and ensure that all test cases pass. Deploy the implemented global mean pooling function and test code to the production environment.

6. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 5 is characterized in that: The bathroom towel rack air tightness attention feature extraction unit comprises: Denoising the bathroom towel rack air tightness test image to obtain a bathroom towel rack air tightness test noise reduction enhanced image; The bathroom towel rack air tightness test denoising and enhancement image is passed through a bathroom towel rack air tightness test enhancement convolutional neural network based on a spatial attention mechanism to obtain a bathroom towel rack air tightness test denoising and enhancement spatial attention feature map; The bathroom towel rack air tightness test denoising enhanced spatial attention feature map is passed through a bathroom towel rack air tightness test feature extractor with multi-feature extraction capability to obtain the bathroom towel rack air tightness test global feature vector.

7. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 6 is characterized in that: The air tightness test and analysis module for bathroom towel rack products includes: An air tightness test feature fusion unit, used for fusing the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector to obtain an air tightness test result judgment feature vector; An airtightness test characteristic optimization unit, used for adjusting the airtightness test result judgment characteristic vector based on the entropy degradation modulation value limit standard to obtain an optimized airtightness test result judgment characteristic vector; The air tightness test result judgment unit is used to pass the optimized air tightness test result judgment feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the bathroom towel rack product air tightness test is qualified.

8. The testing device for automatically detecting air tightness of bathroom towel rack products according to claim 7, characterized in that: The airtightness test feature optimization unit comprises: Extracting each row vector in the prediction parameter space of the classifier to obtain a set of airtightness test prediction parameter row vectors; Calculating the airtightness test class probability entropy degradation simulation value between the airtightness test result judgment feature vector and each airtightness test prediction parameter row vector in the set of airtightness test prediction parameter row vectors to obtain a set of airtightness test class probability entropy degradation simulation values; Calculating the square root of the sum of squares of the airtightness test class probability entropy degradation simulation values ​​to obtain an airtightness test entropy degradation modulation value; Based on the airtightness test entropy degradation modulation value, the airtightness test result judgment feature vector is subjected to feature distribution modulation to obtain the optimized airtightness test result judgment feature vector.

9. A method for automatically detecting the air tightness of bathroom towel rack products, characterized in that: include: Get the air tightness test image of bathroom towel rack; Extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image; Based on the bathroom towel rack air tightness test tracking feature vector and the bathroom towel rack air tightness test global feature vector, it is judged whether the bathroom towel rack product air tightness test is qualified.

10. The method for automatically detecting air tightness of bathroom towel rack products according to claim 9, characterized in that: Extracting a bathroom towel rack air tightness test tracking feature vector and a bathroom towel rack air tightness test global feature vector from the bathroom towel rack air tightness test image, including: Performing target feature extraction on the bathroom towel rack air tightness test image to obtain a bathroom towel rack air tightness test area of ​​interest; Performing feature encoding on the area of ​​interest of the air tightness test of the bathroom towel rack to obtain a tracking feature vector of the air tightness test of the bathroom towel rack; Attention feature extraction is performed on the bathroom towel rack air tightness test image to obtain the bathroom towel rack air tightness test global feature vector.

Images