Washing machine fabric identification method and device based on illumination adjustment and washing machine

Abstract

The invention provides a washing machine fabric identification method and device based on illumination adjustment and a washing machine. The washing machine is internally provided with a plurality of uniformly distributed light sources and at least one image acquisition device. The method comprises the following steps: acquiring a plurality of target brightness combinations containing different brightness combinations of all light sources; controlling light source illumination according to each target brightness combination, and controlling image acquisition equipment to acquire a first target image; and inputting the first target image into a neural network model for identification to obtain an identification result of the fabric type, and determining the fabric type included in the washing machine based on the identification result. The method solves the problem that in the prior art, a fabric identification method for multiple pieces of clothes is low in identification precision.

Description

Technical Field

[0001] The present invention relates to the technical field of washing machine control, and in particular to a method and device for identifying washing machine fabric based on light adjustment, a computer-readable storage medium and a washing machine. Background Art

[0002] In the use scenario of washing machines, there will be clothes of various fabrics with different characteristics, such as water absorption, heat resistance, wear resistance, etc. Therefore, there are huge differences between the washing programs corresponding to clothes of different fabrics. In the use of washing machines, if inappropriate washing methods are used, the washing and cleaning goals may not be achieved, a lot of time and electricity may be wasted, and even damage to clothes may be caused. Therefore, allowing washing machines to automatically identify the fabric of clothes is an important technology that can greatly improve the user experience when using washing machines.

[0003] A variety of clothing fabric identification methods have been proposed in the prior art, including:

[0004] Methods based on clothing tag recognition (such as CN110565348A): Identify fabric types by reading tag information on clothing. However, this method does not work effectively when the tag is missing, damaged or blocked, which greatly limits its scope of application and reliability.

[0005] Single-image-based neural network recognition method (such as CN115992426A): A single image of clothing is analyzed using a neural network to identify the fabric. However, under fixed lighting conditions, the fabric information in the image may be disturbed by factors such as color and texture, and the neural network is difficult to accurately distinguish the characteristics of these fabrics, resulting in a low recognition rate.

[0006] Methods based on clothing weight change identification (such as CN117051561A): The fabric is determined by measuring the weight change of the clothing in the dry and saturated states. This solution works well when dealing with clothing made of a single fabric, but when the clothing is mixed, the weight change is difficult to directly reflect the characteristics of a specific fabric, reducing the accuracy of identification.

[0007] Although the above existing technologies have achieved automatic recognition of clothing fabrics to a certain extent, it is difficult to ensure the accuracy and robustness of recognition for complex clothing mixing, lighting changes, label missing and other situations.

[0008] In summary, there is an urgent need in the prior art for a high-precision solution for fabric recognition using images in a multi-light environment. Summary of the invention

[0009] The main purpose of the present application is to provide a washing machine fabric recognition method, device, computer-readable storage medium and washing machine based on light adjustment, so as to at least solve the problem of low recognition accuracy of the fabric recognition method for multiple clothes in the prior art.

[0010] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for identifying fabrics of a washing machine based on light adjustment is provided, the washing machine comprising a plurality of evenly distributed light sources and at least one image acquisition device, the method comprising: when there are clothes in the washing machine, obtaining a plurality of target brightness combinations, the target brightness combination comprising the brightness of all light sources and the brightness of the light sources being inconsistent between different target brightness combinations; controlling each light source to illuminate according to each target brightness combination, and controlling the image acquisition device to acquire an image to obtain a first target image, the first target image corresponding one-to-one to the target brightness combination; inputting the first target image into a neural network model for identification to obtain a first identification result, and determining the target fabric type at least based on the first identification result, the neural network model being used to identify and classify according to the image, and output the fabric type.

[0011] Optionally, obtaining multiple target brightness combinations includes: obtaining multiple preset brightnesses, obtaining the number of light sources, and obtaining a target number; performing random combinations based on the target number and the multiple preset brightnesses, and deleting repeated combinations therein to obtain a target brightness combination.

[0012] Optionally, the first target image is input into the neural network model for recognition to obtain a first recognition result, including: splicing the first target image according to RGB channels to obtain a second target image; and inputting the second target image into the neural network model to obtain the first recognition result.

[0013] Optionally, the neural network model includes an input layer, a pooling layer and an output layer. The second target image is input into the neural network model to obtain a first recognition result, including: performing feature extraction on the second target image through a shared convolutional layer in the input layer to extract edges, textures and colors to obtain first target features, and performing multi-layer feature extraction on the first target features in the input layer through a residual block to obtain second target features; performing pooling operation on the second target features through a pooling layer to obtain third target features; performing classification based on the third target features through a fully connected network in the output layer to obtain a target output value, and processing the target output value through a sigmoid function in the output layer to obtain a first recognition result, and the target output value is used to characterize the probability distribution of each fabric type.

[0014] Optionally, before splicing the first target image according to RGB channels, the method also includes: cropping the first target image so that the size of the first target image is a reference size; normalizing the brightness of the first target image and mapping the brightness of the first target image to a preset interval; performing image augmentation processing on the first target image, and the image augmentation processing includes one or more of random cropping, rotation, scaling, hue adjustment, and saturation adjustment.

[0015] Optionally, after inputting the first target image into the neural network model for recognition and obtaining a first recognition result, the method also includes: controlling the washing drum of the washing machine to rotate to a preset angle, and controlling each light source to illuminate according to each target brightness combination, and controlling the image acquisition device to acquire an image to obtain a second target image; inputting the first target image into the neural network model for recognition to obtain a second recognition result.

[0016] Optionally, the target fabric type is determined at least based on the first recognition result, including: when the first recognition result is consistent with the second recognition result, the first recognition result is determined as the target fabric type; when the first recognition result is inconsistent with the second recognition result, the fabric types included in the first recognition result and the second recognition result are integrated to obtain the target fabric type.

[0017] According to another aspect of the present application, a washing machine fabric recognition device based on light adjustment is provided, the washing machine includes a plurality of evenly distributed light sources and at least one image acquisition device, the device includes: an acquisition unit, used to acquire a plurality of target brightness combinations when there are clothes in the washing machine, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; a first control unit, used to control each light source to illuminate according to each target brightness combination, and control the image acquisition device to perform image acquisition to obtain a first target image, the first target image corresponds to the target brightness combination one by one; a first classification unit, used to input the first target image into a neural network model for recognition, obtain a first recognition result, and determine the target fabric type at least according to the first recognition result, the neural network model is used to recognize and classify according to the image, and output the fabric type.

[0018] According to another aspect of the present application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein when the program is executed, a device where the computer-readable storage medium is located is controlled to execute any one of the described methods.

[0019] According to another aspect of the present application, a washing machine is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include methods for executing any one of the described methods.

[0020] Applying the technical solution of the present application, in the above-mentioned washing machine fabric identification method based on light adjustment, first, when there are clothes in the washing machine, multiple target brightness combinations are obtained, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; then, each light source is controlled to illuminate according to each target brightness combination, and the image acquisition device is controlled to acquire an image to obtain a first target image, and the first target image corresponds to the target brightness combination one by one; then, the first target image is input into the neural network model for recognition to obtain a first recognition result, and the target fabric type is determined at least according to the first recognition result, and the neural network model is used to identify and classify according to the image, and output the fabric type. The present application can more accurately identify the fabric type of the clothes by collecting clothing images under different light source brightness combinations inside the washing machine and using the neural network model for fabric recognition. This method not only improves the recognition accuracy, but also can adapt to the recognition needs under different lighting conditions, providing users with a more intelligent and personalized laundry experience. Through image acquisition under multiple angles and multiple lighting conditions, combined with deep learning technology, the robustness and accuracy of fabric recognition have been significantly improved, effectively avoiding the recognition errors that may occur under single lighting conditions, and ensuring that the washing machine can select the most appropriate washing program according to the actual fabric type of the clothes, thereby protecting the clothes and extending the life of the clothes. At the same time, it also improves the efficiency and effect of washing, so as to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the existing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A hardware structure block diagram of a mobile terminal according to a method for identifying washing machine fabrics based on light adjustment provided in an embodiment of the present application is shown;

[0022] Figure 2 A schematic flow chart of a washing machine fabric identification method based on light adjustment according to an embodiment of the present application is shown;

[0023] Figure 3 A structural schematic diagram of a washing machine provided according to an embodiment of the present application is shown;

[0024] Figure 4 A schematic diagram of a process of recognizing an image using a neural network model provided in an embodiment of the present application is shown;

[0025] Figure 5 A structural block diagram of a washing machine fabric identification device based on light adjustment provided according to an embodiment of the present application is shown.

[0026] The above drawings include the following reference numerals:

[0027] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] As introduced in the background technology, there is an urgent need in the prior art for a high-precision solution for fabric recognition using images in a multi-light environment. In order to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the prior art, the embodiments of the present application provide a washing machine fabric recognition method, device, computer-readable storage medium and washing machine based on light adjustment.

[0032] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0033] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 11 is a hardware structure block diagram of a mobile terminal for a washing machine fabric identification method based on light adjustment according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is for illustration only and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0034] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the washing machine fabric identification method based on light adjustment in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0035] In this embodiment, a washing machine fabric identification method based on light adjustment that runs on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0036] Figure 2 FIG. 1 is a flow chart of a washing machine fabric identification method based on light adjustment according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0037] Step S201, when there are clothes in the washing machine, a plurality of target brightness combinations are obtained, where the target brightness combination includes the brightness of all light sources and the brightness of the light sources in different target brightness combinations are inconsistent;

[0038] Specifically, to obtain multiple target brightness combinations, it is understandable that multiple light sources (such as Figure 3 As shown, three clothing lighting lamps A, B, and C may be included. In each target brightness combination, the light source can be adjusted to different brightness levels and then randomly combined to obtain the target brightness combination. As many brightness combinations as possible can provide multi-level illumination information at multiple angles and multiple brightnesses, thereby enriching the content of subsequent feature extraction.

[0039] Step S202, controlling each light source to illuminate according to each target brightness combination, and controlling an image acquisition device to acquire an image, to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one;

[0040] Specifically, the light source illumination is controlled and images are collected. Based on the acquired image combination, the brightness of each light source is adjusted to the corresponding target brightness combination one by one to provide different lighting environments for the clothes inside the washing machine. Under each target brightness combination, the image acquisition device (such as a built-in camera) will collect images of the clothes to obtain a series of first target images.

[0041] It can be understood that these images corresponding to the target brightness combinations constitute a group of first target images, which can intuitively reflect the light and dark changes, texture characteristics and color performance of clothing fabrics under different lighting.

[0042] Step S203, inputting the first target image into the neural network model for recognition, obtaining a first recognition result, and determining the target fabric type at least based on the first recognition result, the neural network model is used to recognize and classify according to the image, and output the fabric type.

[0043] Specifically, recognition is performed through a neural network model, and the acquired first target image is input into the neural network model. The model is applied to image recognition and classification, specifically recognition and classification of clothing fabrics, and then the probability corresponding to the possible fabric types of the clothes inside the washing machine is output, that is, the first recognition result, and then the fabric type contained in the washing machine is finally determined based on the first recognition result to obtain the target fabric type.

[0044] In the above embodiment, the combination of the multi-lamp lighting system and the neural network model is used to realize the intelligent recognition of the fabric type of clothes in the washing machine, which significantly improves the accuracy and efficiency of recognition and provides key technical support for the intelligent upgrade of the washing machine. Compared with the existing technology, this method shows better performance and reliability when dealing with complex situations such as mixed clothes and changes in lighting.

[0045] Through this embodiment, first, when there are clothes in the washing machine, multiple target brightness combinations are obtained, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; then, each light source is controlled to illuminate according to each target brightness combination, and the image acquisition device is controlled to acquire an image to obtain a first target image, and the first target image corresponds to the target brightness combination one by one; then, the first target image is input into the neural network model for recognition to obtain a first recognition result, and the target fabric type is determined at least according to the first recognition result, and the neural network model is used to recognize and classify according to the image, and output the fabric type. The present application can more accurately identify the fabric type of the clothes by collecting clothing images under different light source brightness combinations inside the washing machine and using the neural network model for fabric recognition. This method not only improves the recognition accuracy, but also can adapt to the recognition needs under different lighting conditions, providing users with a more intelligent and personalized laundry experience. Through image acquisition under multiple angles and multiple lighting conditions, combined with deep learning technology, the robustness and accuracy of fabric recognition have been significantly improved, effectively avoiding the recognition errors that may occur under single lighting conditions, and ensuring that the washing machine can select the most appropriate washing program according to the actual fabric type of the clothes, thereby protecting the clothes and extending the life of the clothes. At the same time, it also improves the efficiency and effect of washing, so as to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the existing technology.

[0046] In order to generate a target brightness combination, in an optional implementation, the above step S201 includes:

[0047] Step S2011, obtaining a plurality of preset brightnesses, obtaining the number of light sources, and obtaining a target number;

[0048] Specifically, first, the brightness level of the built-in light source of the washing machine can be adjusted, that is, the preset brightness, such as high brightness, medium brightness, low brightness and off. At the same time, the number of light sources is obtained. In one embodiment, Figure 3 As shown, the present application provides three light sources, namely, A, B and C clothing lighting lamps.

[0049] It is understandable that, considering that the highlight moments can provide image information dominated by lighting in a specific direction, and in order to ensure that there are no shadows on the clothes in the washing machine and that each part has sufficient brightness, the closed state is not actually applicable during the shooting process.

[0050] Step S2012: performing random combinations based on the target quantity and the plurality of preset brightnesses, and deleting repeated combinations therein to obtain a target brightness combination.

[0051] Specifically, based on the target quantity and the plurality of preset brightnesses, random combinations are performed, and different brightness combinations can be generated according to the number of light sources and the preset brightnesses.

[0052] In one embodiment, there are 27 combinations of three preset brightness and three light sources in the present application, and repeated combinations cannot provide new features, so after deleting the repeated combinations and eliminating the unreasonable combinations, there are 7 combinations. That is, effective and non-repeated combinations are screened out from the theoretically generated combinations to ensure that each combination can provide unique lighting conditions, so as to capture different clothing images.

[0053] It is understandable that in order to make the features extracted richer, more gradients can be set for the brightness or more light sources can be set, which are options that can be adjusted according to actual needs.

[0054] The above 7 combinations include: lamps A, B, and C are all high brightness (full brightness combination); lamps A and B are high brightness, and lamp C is low brightness (A / B dominant combination); lamps A and C are high brightness, and lamp B is low brightness (A / C dominant combination); lamps B and C are high brightness, and lamp A is low brightness (B / C dominant combination); lamp A is high brightness, and lamps B and C are low brightness (A dominant combination); lamp B is high brightness, and lamps A and C are low brightness (B dominant combination); lamp C is high brightness, and lamps A and B are low brightness (C dominant combination). It can be seen that the combination of all low brightness is not actually set to avoid the collected image features being unclear and affecting the results.

[0055] In the above embodiment, through the above combination, the washing machine can take images of clothes under different lighting conditions, and these images form an information-rich data set, which includes the reflection and absorption characteristics of clothes under various lighting angles. Then, this set of images is input into the neural network model as a whole, and the model recognizes the fabric type based on the rich information in these images, which effectively improves the accuracy and confidence of the recognition, which is better than the recognition effect using only a single image or a single lighting condition.

[0056] In order to identify the first target image through the neural network model and obtain the type of clothing fabric, in an optional implementation, the above step S203 includes:

[0057] Step S2031, splicing the first target image according to RGB channels to obtain a second target image;

[0058] Specifically, Figure 4As shown, the first target image under different lighting conditions is spliced ​​according to the RGB channels to obtain the second target image. The key is to splice multiple RGB images in the channel dimension rather than the spatial dimension, thereby retaining the resolution of the original image and integrating the information of the lighting change into the same image.

[0059] In one embodiment, if the washing machine collects 7 different RGB images, then through channel stitching, a second target image with the same resolution but 3×7=21 channels can be obtained. This image contains a comprehensive representation of clothing fabrics under a series of lighting conditions, providing the neural network with richer and more comprehensive fabric feature information.

[0060] Step S2032, input the second target image into the neural network model to obtain a first recognition result.

[0061] Specifically, Figure 4 As shown, the second target image is input into a pre-trained artificial neural network model for feature extraction. The input of the traditional convolutional neural network is improved to adapt to the input of a high number of channels. The high-level features of clothing fabrics are extracted through operations such as convolution pooling, and the features are output through the classification head to obtain the probability that the washing machine contains various types of fabrics.

[0062] It is understandable that a probability threshold may be set, and when the probability exceeds the probability threshold, it is determined that the corresponding fabric exists in the washing machine.

[0063] Through the above embodiments, through the intelligent analysis of the neural network model, high-precision recognition of clothing fabrics is achieved, effectively overcoming the problem of interference of fabric information under fixed lighting conditions and the problem that a single image is difficult to fully reflect the characteristics of the fabric. This is a significant improvement over the existing single-image recognition method (such as CN115992426A). By integrating the information of multiple images, it not only increases the dimension of the input data, but also improves the accuracy and robustness of the neural network in identifying clothing fabrics, providing strong technical support for the intelligent control of washing machines.

[0064] In order to extract features from the second target image using the neural network model, in an optional implementation, the above step S2032 includes:

[0065] Step S20321, performing feature extraction on the second target image through the shared convolution layer in the input layer to extract edges, textures and colors to obtain first target features, and performing multi-layer feature extraction on the first target features in the input layer through the residual block to obtain second target features;

[0066] Specifically, feature extraction includes feature extraction of shared convolutional layers and multi-layer feature extraction of residual blocks, specifically including: first, the input layer of the model contains a shared convolutional layer for preliminary processing of the input multi-illumination scene images (second target images). These images show a variety of visual features of clothing under different lighting conditions. The shared convolutional layer can extract basic features such as edges, textures, and colors in the image to form the first target features, laying the foundation for subsequent deep feature extraction. Then, the model performs multi-layer feature extraction through residual blocks to obtain more abstract and advanced features (second target features). The design of residual blocks can avoid the gradient vanishing problem in the training process of deep neural networks, ensuring that the model can still maintain good training effects when processing high-resolution and high-dimensional images. After multi-layer processing of residual blocks, the model can comprehensively analyze the fabric characteristics of clothing from different perspectives and different lighting conditions to improve the accuracy of recognition.

[0067] Step S20322, performing a pooling operation on the second target feature through a pooling layer to obtain a third target feature;

[0068] Specifically, the second target feature is pooled by the pooling layer to obtain the third target feature. Pooling operations usually include maximum pooling or average pooling, which can further reduce the dimension of the feature while retaining the most critical information, such as the structural pattern and color distribution of the fabric, thereby effectively reducing the amount of calculation, accelerating the model processing speed, and improving the robustness of the model.

[0069] Step S20323, classify based on the third target feature through the fully connected network in the output layer to obtain a target output value, and process the target output value through the sigmoid function in the output layer to obtain a first recognition result, and the target output value is used to characterize the probability distribution of each fabric type.

[0070] Specifically, classification and probability prediction are first classified based on the fully connected network, and the third target feature is further processed by the fully connected network in the output layer to achieve the classification of fabric types. The fully connected layer can integrate the feature information after pooling to form a highly abstract representation of the fabric type of clothing, which is convenient for the model to make the final classification decision. Then, the output of the fully connected network is processed by the sigmoid function to obtain the first recognition result, that is, the probability distribution of each fabric type (target output value). The Sigmoid function can convert the continuous value of the output layer into a probability value between 0 and 1, clearly indicating the probability of each fabric type in the current clothing, and providing a basis for subsequent fabric type judgment.

[0071] It can be understood that, in one embodiment, assuming that the preset fabric types are 6, the network output terminal is 6 values, corresponding to fabrics ABCDEF, and then, for an empty washing machine, a result close to 000000 will be output, indicating that there is no fabric ABCDEF in the washing machine. If there is a piece of clothing made of a blend of fabrics A and C in the washing machine, the program will output a result close to 101000, indicating that there are fabrics A and C in the washing machine, but no fabrics BDEF.

[0072] Furthermore, the washing machine will give the user feedback on the detected clothing fabric type on the display screen, and then select a suitable washing program based on the detected clothing fabric type, such as "wool washing mode", "chemical fiber mode", "mixed mode", etc.

[0073] In the above embodiment, the neural network model can not only capture the subtle differences in clothing fabrics under different lighting conditions through multi-level feature extraction, pooling operations and probability prediction, but also express the degree of existence of various fabric types in the form of probability, providing a scientific basis for subsequent washing program selection.

[0074] In order to ensure the accuracy of the recognition result, in an optional implementation, before splicing the first target image according to the RGB channels, the method further includes:

[0075] Step S301, cropping the first target image so that the size of the first target image is the reference size;

[0076] Specifically, images taken under multiple lighting conditions may have different sizes due to the internal structure of the washing machine or other factors, which will affect the input consistency of the neural network model. Therefore, before stitching the images into the second target image, the first target image needs to be cropped to ensure that all images have the same base size. For example, if the base size is set to 640×480 pixels, all first target images will be cropped to this size to eliminate the impact of size differences on the recognition process.

[0077] Step S302, normalizing the brightness of the first target image, and mapping the brightness of the first target image to a preset range;

[0078] Specifically, the brightness of the first target image under different lighting conditions may be significantly different, and this difference may affect the model's accurate extraction of fabric features. Therefore, it is necessary to normalize the brightness of the image to ensure that the brightness of all images is distributed within a preset range, such as between 0 and 1. In this way, when the model receives image input, it will not produce erroneous recognition results due to brightness differences. In the embodiment, assuming that the brightness range of a pixel in the first target image is 0 to 255, after normalization, the brightness value of the pixel will be mapped to the range of 0 to 1, thereby ensuring that the brightness level of all images is consistent, which is convenient for subsequent model processing and analysis.

[0079] Step S303: performing image augmentation processing on the first target image, where the image augmentation processing includes one or more of random cropping, rotation, scaling, hue adjustment, and saturation adjustment.

[0080] Specifically, in order to enhance the robustness and generalization ability of the neural network model, the first target image needs to be augmented. This process includes but is not limited to operations such as random cropping, rotation, scaling, hue adjustment, and saturation adjustment. Image augmentation processing can simulate and increase the diversity of images, so that the model is exposed to more varying image samples during training, thereby making the recognition of fabrics more accurate and stable in practical applications. Under the same lighting combination conditions, by randomly cropping clothing images in different areas, or rotating and adjusting the hue of the images, multiple images with slightly different perspectives and appearances can be generated, which helps the neural network model learn more comprehensive fabric features.

[0081] Through image preprocessing, the washing machine can eliminate the deviation in the image acquisition process, improve the generalization ability of the recognition model, and be suitable for processing clothes of various shapes, sizes and colors, providing users with more comprehensive and accurate fabric recognition services.

[0082] In order to avoid recognition errors caused by clothing covering, in an optional implementation, after inputting the first target image into the neural network model for recognition and obtaining the first recognition result, the method further includes:

[0083] Step S401, controlling the washing tub of the washing machine to rotate to a preset angle, controlling each light source to illuminate according to each target brightness combination, and controlling the image acquisition device to acquire an image to obtain a second target image;

[0084] Specifically, in the present application, in order to avoid recognition errors caused by the stacking of clothes inside the washing machine, the washing tub is rotated after a single recognition is completed, and the recognition image is collected again to obtain a second target image.

[0085] Step S402: input the first target image into the neural network model for recognition to obtain a second recognition result.

[0086] Specifically, after acquiring the second target image, these images are also integrated and processed, and multiple RGB images are spliced ​​in the channel dimension to form high-dimensional image data, which are input into the neural network model. The model will perform secondary recognition and classification based on these second target images containing the illumination information at the rotated angle, and output the second recognition result, that is, the type of clothing fabric recognized again and its confidence.

[0087] In order to determine the type of fabric in the washing machine, in an optional implementation, the above step S203 further includes:

[0088] Step S2033, when the first recognition result is consistent with the second recognition result, determining the first recognition result as the target fabric type;

[0089] Specifically, if the two are consistent, that is, the clothing covering does not interfere with the identification of the clothing fabric type, then the clothing fabric type can be determined according to the first identification result.

[0090] Step S2034: when the first recognition result is inconsistent with the second recognition result, the fabric types included in the first recognition result and the second recognition result are integrated to obtain the target fabric type.

[0091] Specifically, if the first recognition result is inconsistent with the second recognition result, indicating that there may be a recognition error, at this time, the present invention proposes an integration mechanism, that is, fusing the fabric types contained in the two results, using probability weighting or rule reasoning methods, combining the neural network prediction probability of the first recognition result and the credibility of the second recognition result, to obtain a comprehensive target fabric type.

[0092] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0093] The embodiment of the present application also provides a washing machine fabric identification device based on light adjustment. It should be noted that the washing machine fabric identification device based on light adjustment in the embodiment of the present application can be used to execute the washing machine fabric identification method based on light adjustment provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0094] The following is an introduction to the washing machine fabric identification device based on light adjustment provided in an embodiment of the present application.

[0095] Figure 5 is a structural block diagram of a washing machine fabric identification device based on light adjustment according to an embodiment of the present application. Figure 5 As shown, the device comprises:

[0096] An acquisition unit 10 is used to acquire a plurality of target brightness combinations when there are clothes in the washing machine, wherein the target brightness combination includes the brightness of all light sources and the brightness of the light sources in different target brightness combinations is inconsistent;

[0097] Specifically, to obtain multiple target brightness combinations, it is understandable that multiple light sources (such as Figure 3 As shown, three clothing lighting lamps A, B, and C may be included. In each target brightness combination, the light source can be adjusted to different brightness levels and then randomly combined to obtain the target brightness combination. As many brightness combinations as possible can provide multi-level illumination information at multiple angles and multiple brightnesses, thereby enriching the content of subsequent feature extraction.

[0098] A first control unit 20 is used to control each light source to perform illumination according to each target brightness combination, and to control the image acquisition device to perform image acquisition to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one;

[0099] Specifically, the light source illumination is controlled and images are collected. Based on the acquired image combination, the brightness of each light source is adjusted to the corresponding target brightness combination one by one to provide different lighting environments for the clothes inside the washing machine. Under each target brightness combination, the image acquisition device (such as a built-in camera) will collect images of the clothes to obtain a series of first target images.

[0100] It can be understood that these images corresponding to the target brightness combinations constitute a group of first target images, which can intuitively reflect the light and dark changes, texture characteristics and color performance of clothing fabrics under different lighting.

[0101] The first classification unit 30 is used to input the first target image into the neural network model for recognition, obtain a first recognition result, and determine the target fabric type at least based on the first recognition result. The neural network model is used to recognize and classify according to the image and output the fabric type.

[0102] Specifically, recognition is performed through a neural network model, and the acquired first target image is input into the neural network model. The model is applied to image recognition and classification, specifically recognition and classification of clothing fabrics, and then the probability corresponding to the possible fabric types of the clothes inside the washing machine is output, that is, the first recognition result, and then the fabric type contained in the washing machine is finally determined based on the first recognition result to obtain the target fabric type.

[0103] In the above embodiment, the combination of the multi-lamp lighting system and the neural network model is used to realize the intelligent recognition of the fabric type of clothes in the washing machine, which significantly improves the accuracy and efficiency of recognition and provides key technical support for the intelligent upgrade of the washing machine. Compared with the existing technology, this method shows better performance and reliability when dealing with complex situations such as mixed clothes and changes in lighting.

[0104] Through this embodiment, when there are clothes in the washing machine, the acquisition unit acquires multiple target brightness combinations, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; the first control unit controls each light source to illuminate according to each target brightness combination, and controls the image acquisition device to acquire an image, and obtains a first target image, and the first target image corresponds to the target brightness combination one by one; the first classification unit inputs the first target image into the neural network model for recognition, obtains a first recognition result, and determines the target fabric type at least according to the first recognition result, and the neural network model is used to recognize and classify according to the image, and output the fabric type. This application can more accurately identify the fabric type of the clothes by collecting clothing images under different light source brightness combinations inside the washing machine and using the neural network model for fabric recognition. This method not only improves the recognition accuracy, but also can adapt to the recognition needs under different lighting conditions, providing users with a more intelligent and personalized laundry experience. Through image acquisition under multiple angles and multiple lighting conditions, combined with deep learning technology, the robustness and accuracy of fabric recognition have been significantly improved, effectively avoiding the recognition errors that may occur under single lighting conditions, and ensuring that the washing machine can select the most appropriate washing program according to the actual fabric type of the clothes, thereby protecting the clothes and extending the life of the clothes. At the same time, it also improves the efficiency and effect of washing, so as to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the existing technology.

[0105] In order to generate a target brightness combination, in an optional implementation, the acquisition unit includes:

[0106] A first acquisition module is used to acquire a plurality of preset brightnesses, acquire the number of light sources, and obtain a target number;

[0107] Specifically, first, the brightness level of the built-in light source of the washing machine can be adjusted, that is, the preset brightness, such as high brightness, medium brightness, low brightness and off. At the same time, the number of light sources is obtained. In one embodiment, Figure 3 As shown, the present application provides three light sources, namely, A, B and C clothing lighting lamps.

[0108] It is understandable that, considering that the highlight moments can provide image information dominated by lighting in a specific direction, and in order to ensure that there are no shadows on the clothes in the washing machine and that each part has sufficient brightness, the closed state is not actually applicable during the shooting process.

[0109] The first processing module is used to perform random combinations based on the target quantity and the plurality of preset brightnesses, and delete repeated combinations therein to obtain a target brightness combination.

[0110] Specifically, based on the target quantity and the plurality of preset brightnesses, random combinations are performed, and different brightness combinations can be generated according to the number of light sources and the preset brightnesses.

[0111] In one embodiment, there are 27 combinations of three preset brightness and three light sources in the present application, and repeated combinations cannot provide new features, so after deleting the repeated combinations and eliminating the unreasonable combinations, there are 7 combinations. That is, effective and non-repeated combinations are screened out from the theoretically generated combinations to ensure that each combination can provide unique lighting conditions, so as to capture different clothing images.

[0112] It is understandable that in order to make the features extracted richer, more gradients can be set for the brightness or more light sources can be set, which are options that can be adjusted according to actual needs.

[0113] The above 7 combinations include: lamps A, B, and C are all high brightness (full brightness combination); lamps A and B are high brightness, and lamp C is low brightness (A / B dominant combination); lamps A and C are high brightness, and lamp B is low brightness (A / C dominant combination); lamps B and C are high brightness, and lamp A is low brightness (B / C dominant combination); lamp A is high brightness, and lamps B and C are low brightness (A dominant combination); lamp B is high brightness, and lamps A and C are low brightness (B dominant combination); lamp C is high brightness, and lamps A and B are low brightness (C dominant combination). It can be seen that the combination of all low brightness is not actually set to avoid the collected image features being unclear and affecting the results.

[0114] In the above embodiment, through the above combination, the washing machine can take images of clothes under different lighting conditions, and these images form an information-rich data set, which includes the reflection and absorption characteristics of clothes under various lighting angles. Then, this set of images is input into the neural network model as a whole, and the model recognizes the fabric type based on the rich information in these images, which effectively improves the accuracy and confidence of the recognition, which is better than the recognition effect using only a single image or a single lighting condition.

[0115] In order to identify the first target image through the neural network model and obtain the type of clothing fabric, in an optional implementation, the first classification unit includes:

[0116] A second processing module is used to splice the first target image according to RGB channels to obtain a second target image;

[0117] Specifically, Figure 4 As shown, the first target image under different lighting conditions is spliced ​​according to the RGB channels to obtain the second target image. The key is to splice multiple RGB images in the channel dimension rather than the spatial dimension, thereby retaining the resolution of the original image and integrating the information of the lighting change into the same image.

[0118] In one embodiment, if the washing machine collects 7 different RGB images, then through channel stitching, a second target image with the same resolution but 3×7=21 channels can be obtained. This image contains a comprehensive representation of clothing fabrics under a series of lighting conditions, providing the neural network with richer and more comprehensive fabric feature information.

[0119] The first classification module is used to input the second target image into the neural network model to obtain a first recognition result.

[0120] Specifically, Figure 4 As shown, the second target image is input into a pre-trained artificial neural network model for feature extraction. The input of the traditional convolutional neural network is improved to adapt to the input of a high number of channels. The high-level features of clothing fabrics are extracted through operations such as convolution pooling, and the features are output through the classification head to obtain the probability that the washing machine contains various types of fabrics.

[0121] It is understandable that a probability threshold may be set, and when the probability exceeds the probability threshold, it is determined that the corresponding fabric exists in the washing machine.

[0122] Through the above embodiments, through the intelligent analysis of the neural network model, high-precision recognition of clothing fabrics is achieved, effectively overcoming the problem of interference of fabric information under fixed lighting conditions and the problem that a single image is difficult to fully reflect the characteristics of the fabric. This is a significant improvement over the existing single-image recognition method (such as CN115992426A). By integrating the information of multiple images, it not only increases the dimension of the input data, but also improves the accuracy and robustness of the neural network in identifying clothing fabrics, providing strong technical support for the intelligent control of washing machines.

[0123] In order to extract features of the second target image through the neural network model, in an optional implementation, the first classification module includes:

[0124] The convolution submodule is used to perform feature extraction on the second target image through the shared convolution layer in the input layer to extract edges, textures and colors to obtain first target features, and perform multi-layer feature extraction on the first target features in the input layer through the residual block to obtain second target features;

[0125] Specifically, feature extraction includes feature extraction of shared convolutional layers and multi-layer feature extraction of residual blocks, specifically including: first, the input layer of the model contains a shared convolutional layer for preliminary processing of the input multi-illumination scene images (second target images). These images show a variety of visual features of clothing under different lighting conditions. The shared convolutional layer can extract basic features such as edges, textures, and colors in the image to form the first target features, laying the foundation for subsequent deep feature extraction. Then, the model performs multi-layer feature extraction through residual blocks to obtain more abstract and advanced features (second target features). The design of residual blocks can avoid the gradient vanishing problem in the training process of deep neural networks, ensuring that the model can still maintain good training effects when processing high-resolution and high-dimensional images. After multi-layer processing of residual blocks, the model can comprehensively analyze the fabric characteristics of clothing from different perspectives and different lighting conditions to improve the accuracy of recognition.

[0126] A pooling submodule, used for performing a pooling operation on the second target feature through a pooling layer to obtain a third target feature;

[0127] Specifically, the second target feature is pooled by the pooling layer to obtain the third target feature. Pooling operations usually include maximum pooling or average pooling, which can further reduce the dimension of the feature while retaining the most critical information, such as the structural pattern and color distribution of the fabric, thereby effectively reducing the amount of calculation, accelerating the model processing speed, and improving the robustness of the model.

[0128] The classification submodule is used to classify based on the third target feature through the fully connected network in the output layer to obtain a target output value, and process the target output value through the sigmoid function in the output layer to obtain a first recognition result. The target output value is used to characterize the probability distribution of each fabric type.

[0129] Specifically, classification and probability prediction are first classified based on the fully connected network, and the third target feature is further processed by the fully connected network in the output layer to achieve the classification of fabric types. The fully connected layer can integrate the feature information after pooling to form a highly abstract representation of the fabric type of clothing, which is convenient for the model to make the final classification decision. Then, the output of the fully connected network is processed by the sigmoid function to obtain the first recognition result, that is, the probability distribution of each fabric type (target output value). The Sigmoid function can convert the continuous value of the output layer into a probability value between 0 and 1, clearly indicating the probability of each fabric type in the current clothing, and providing a basis for subsequent fabric type judgment.

[0130] It can be understood that, in one embodiment, assuming that the preset fabric types are 6, the network output terminal is 6 values, corresponding to fabrics ABCDEF, and then, for an empty washing machine, a result close to 000000 will be output, indicating that there is no fabric ABCDEF in the washing machine. If there is a piece of clothing made of a blend of fabrics A and C in the washing machine, the program will output a result close to 101000, indicating that there are fabrics A and C in the washing machine, but no fabrics BDEF.

[0131] Furthermore, the washing machine will give the user feedback on the detected clothing fabric type on the display screen, and then select a suitable washing program based on the detected clothing fabric type, such as "wool washing mode", "chemical fiber mode", "mixed mode", etc.

[0132] In the above embodiment, the neural network model can not only capture the subtle differences in clothing fabrics under different lighting conditions through multi-level feature extraction, pooling operations and probability prediction, but also express the degree of existence of various fabric types in the form of probability, providing a scientific basis for subsequent washing program selection.

[0133] In order to ensure the accuracy of the recognition result, in an optional implementation manner, the above-mentioned device further includes:

[0134] A first processing unit is used to crop the first target image before splicing the first target image according to RGB channels, so that the size of the first target image is the reference size;

[0135] Specifically, images taken under multiple lighting conditions may have different sizes due to the internal structure of the washing machine or other factors, which will affect the input consistency of the neural network model. Therefore, before stitching the images into the second target image, the first target image needs to be cropped to ensure that all images have the same base size. For example, if the base size is set to 640×480 pixels, all first target images will be cropped to this size to eliminate the impact of size differences on the recognition process.

[0136] A second processing unit, configured to normalize the brightness of the first target image and map the brightness of the first target image to a preset interval;

[0137] Specifically, the brightness of the first target image under different lighting conditions may be significantly different, and this difference may affect the model's accurate extraction of fabric features. Therefore, it is necessary to normalize the brightness of the image to ensure that the brightness of all images is distributed within a preset range, such as between 0 and 1. In this way, when the model receives image input, it will not produce erroneous recognition results due to brightness differences. In the embodiment, assuming that the brightness range of a pixel in the first target image is 0 to 255, after normalization, the brightness value of the pixel will be mapped to the range of 0 to 1, thereby ensuring that the brightness level of all images is consistent, which is convenient for subsequent model processing and analysis.

[0138] The third processing unit is used to perform image augmentation processing on the first target image, where the image augmentation processing includes one or more of random cropping, rotation, scaling, hue adjustment, and saturation adjustment.

[0139] Specifically, in order to enhance the robustness and generalization ability of the neural network model, the first target image needs to be augmented. This process includes but is not limited to operations such as random cropping, rotation, scaling, hue adjustment, and saturation adjustment. Image augmentation processing can simulate and increase the diversity of images, so that the model is exposed to more varying image samples during training, thereby making the recognition of fabrics more accurate and stable in practical applications. Under the same lighting combination conditions, by randomly cropping clothing images in different areas, or rotating and adjusting the hue of the images, multiple images with slightly different perspectives and appearances can be generated, which helps the neural network model learn more comprehensive fabric features.

[0140] Through image preprocessing, the washing machine can eliminate the deviation in the image acquisition process, improve the generalization ability of the recognition model, and be suitable for processing clothes of various shapes, sizes and colors, providing users with more comprehensive and accurate fabric recognition services.

[0141] In order to avoid recognition errors due to clothing covering, in an optional embodiment, the above device further includes:

[0142] a second control unit, for inputting the first target image into the neural network model for recognition and obtaining a first recognition result, controlling the washing tub of the washing machine to rotate by a preset angle, controlling each light source to illuminate according to each target brightness combination, and controlling the image acquisition device to acquire an image to obtain a second target image;

[0143] Specifically, in the present application, in order to avoid recognition errors caused by the stacking of clothes inside the washing machine, the washing tub is rotated after a single recognition is completed, and the recognition image is collected again to obtain a second target image.

[0144] The second classification unit is used to input the first target image into the neural network model for recognition to obtain a second recognition result.

[0145] Specifically, after acquiring the second target image, these images are also integrated and processed, and multiple RGB images are spliced ​​in the channel dimension to form high-dimensional image data, which are input into the neural network model. The model will perform secondary recognition and classification based on these second target images containing the illumination information at the rotated angle, and output the second recognition result, that is, the type of clothing fabric recognized again and its confidence.

[0146] In order to determine the type of fabric in the washing machine, in an optional embodiment, the first classification unit further includes:

[0147] A first determination module, configured to determine the first recognition result as the target fabric type when the first recognition result is consistent with the second recognition result;

[0148] Specifically, if the two are consistent, that is, the clothing covering does not interfere with the identification of the clothing fabric type, then the clothing fabric type can be determined according to the first identification result.

[0149] The second determination module is used to obtain a target fabric type by integrating the fabric types included in the first recognition result and the second recognition result when the first recognition result is inconsistent with the second recognition result.

[0150] Specifically, if the first recognition result is inconsistent with the second recognition result, indicating that there may be a recognition error, at this time, the present invention proposes an integration mechanism, that is, fusing the fabric types contained in the two results, using probability weighting or rule reasoning methods, combining the neural network prediction probability of the first recognition result and the credibility of the second recognition result, to obtain a comprehensive target fabric type.

[0151] The above-mentioned washing machine fabric identification device based on light adjustment includes a processor and a memory, and the above-mentioned acquisition unit, the first control unit and the first classification unit are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in any combination.

[0152] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the accuracy of the washing machine's clothing fabric recognition can be improved by adjusting the kernel parameters.

[0153] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0154] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program. When the program is executed, the device where the computer-readable storage medium is located is controlled to execute the washing machine fabric identification method based on light adjustment.

[0155] Specifically, the washing machine fabric recognition method based on light adjustment includes:

[0156] Step S201, when there are clothes in the washing machine, a plurality of target brightness combinations are obtained, where the target brightness combination includes the brightness of all light sources and the brightness of the light sources in different target brightness combinations are inconsistent;

[0157] Specifically, to obtain multiple target brightness combinations, it is understandable that multiple light sources (such as Figure 3 As shown, three clothing lighting lamps A, B, and C may be included. In each target brightness combination, the light source can be adjusted to different brightness levels and then randomly combined to obtain the target brightness combination. As many brightness combinations as possible can provide multi-level illumination information at multiple angles and multiple brightnesses, thereby enriching the content of subsequent feature extraction.

[0158] Step S202, controlling each light source to illuminate according to each target brightness combination, and controlling an image acquisition device to acquire an image, to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one;

[0159] Specifically, the light source illumination is controlled and images are collected. Based on the acquired image combination, the brightness of each light source is adjusted to the corresponding target brightness combination one by one to provide different lighting environments for the clothes inside the washing machine. Under each target brightness combination, the image acquisition device (such as a built-in camera) will collect images of the clothes to obtain a series of first target images.

[0160] It can be understood that these images corresponding to the target brightness combinations constitute a group of first target images, which can intuitively reflect the light and dark changes, texture characteristics and color performance of clothing fabrics under different lighting.

[0161] Step S203, inputting the first target image into the neural network model for recognition, obtaining a first recognition result, and determining the target fabric type at least based on the first recognition result, the neural network model is used to recognize and classify according to the image, and output the fabric type.

[0162] Specifically, recognition is performed through a neural network model, and the acquired first target image is input into the neural network model. The model is applied to image recognition and classification, specifically recognition and classification of clothing fabrics, and then the probability corresponding to the possible fabric types of the clothes inside the washing machine is output, that is, the first recognition result, and then the fabric type contained in the washing machine is finally determined based on the first recognition result to obtain the target fabric type.

[0163] An embodiment of the present invention provides a washing machine, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, at least the following steps are implemented:

[0164] Step S201, when there are clothes in the washing machine, a plurality of target brightness combinations are obtained, where the target brightness combination includes the brightness of all light sources and the brightness of the light sources in different target brightness combinations are inconsistent;

[0165] Specifically, to obtain multiple target brightness combinations, it is understandable that multiple light sources (such as Figure 3 As shown, three clothing lighting lamps A, B, and C may be included. In each target brightness combination, the light source can be adjusted to different brightness levels and then randomly combined to obtain the target brightness combination. As many brightness combinations as possible can provide multi-level illumination information at multiple angles and multiple brightnesses, thereby enriching the content of subsequent feature extraction.

[0166] Step S202, controlling each light source to illuminate according to each target brightness combination, and controlling an image acquisition device to acquire an image, to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one;

[0167] Specifically, the light source illumination is controlled and images are collected. Based on the acquired image combination, the brightness of each light source is adjusted to the corresponding target brightness combination one by one to provide different lighting environments for the clothes inside the washing machine. Under each target brightness combination, the image acquisition device (such as a built-in camera) will collect images of the clothes to obtain a series of first target images.

[0168] It can be understood that these images corresponding to the target brightness combinations constitute a group of first target images, which can intuitively reflect the light and dark changes, texture characteristics and color performance of clothing fabrics under different lighting.

[0169] Step S203, inputting the first target image into the neural network model for recognition, obtaining a first recognition result, and determining the target fabric type at least based on the first recognition result, the neural network model is used to recognize and classify according to the image, and output the fabric type.

[0170] Specifically, recognition is performed through a neural network model, and the acquired first target image is input into the neural network model. The model is applied to image recognition and classification, specifically recognition and classification of clothing fabrics, and then the probability corresponding to the possible fabric types of the clothes inside the washing machine is output, that is, the first recognition result, and then the fabric type contained in the washing machine is finally determined based on the first recognition result to obtain the target fabric type.

[0171] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initializing at least the following method steps:

[0172] Step S201, when there are clothes in the washing machine, a plurality of target brightness combinations are obtained, where the target brightness combination includes the brightness of all light sources and the brightness of the light sources in different target brightness combinations are inconsistent;

[0173] Specifically, to obtain multiple target brightness combinations, it is understandable that multiple light sources (such as Figure 3 As shown, three clothing lighting lamps A, B, and C may be included. In each target brightness combination, the light source can be adjusted to different brightness levels and then randomly combined to obtain the target brightness combination. As many brightness combinations as possible can provide multi-level illumination information at multiple angles and multiple brightnesses, thereby enriching the content of subsequent feature extraction.

[0174] Step S202, controlling each light source to illuminate according to each target brightness combination, and controlling an image acquisition device to acquire an image, to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one;

[0175] Specifically, the light source illumination is controlled and images are collected. Based on the acquired image combination, the brightness of each light source is adjusted to the corresponding target brightness combination one by one to provide different lighting environments for the clothes inside the washing machine. Under each target brightness combination, the image acquisition device (such as a built-in camera) will collect images of the clothes to obtain a series of first target images.

[0176] It can be understood that these images corresponding to the target brightness combinations constitute a group of first target images, which can intuitively reflect the light and dark changes, texture characteristics and color performance of clothing fabrics under different lighting.

[0177] Step S203, inputting the first target image into the neural network model for recognition, obtaining a first recognition result, and determining the target fabric type at least based on the first recognition result, the neural network model is used to recognize and classify according to the image, and output the fabric type.

[0178] Specifically, recognition is performed through a neural network model, and the acquired first target image is input into the neural network model. The model is applied to image recognition and classification, specifically recognition and classification of clothing fabrics, and then the probability corresponding to the possible fabric types of the clothes inside the washing machine is output, that is, the first recognition result, and then the fabric type contained in the washing machine is finally determined based on the first recognition result to obtain the target fabric type.

[0179] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0180] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0181] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0182] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0183] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0184] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0185] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0186] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0187] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0188] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0189] 1) The fabric identification method of a washing machine based on light adjustment of the present application, first, when there are clothes in the washing machine, multiple target brightness combinations are obtained, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; then, each light source is controlled to illuminate according to each target brightness combination, and the image acquisition device is controlled to acquire an image to obtain a first target image, and the first target image corresponds to the target brightness combination one by one; then, the first target image is input into a neural network model for recognition to obtain a first recognition result, and the target fabric type is determined at least according to the first recognition result, and the neural network model is used to identify and classify according to the image, and output the fabric type. The present application can more accurately identify the fabric type of the clothes by collecting clothing images under different light source brightness combinations inside the washing machine and using a neural network model for fabric recognition. This method not only improves the recognition accuracy, but also can adapt to the recognition needs under different lighting conditions, providing users with a more intelligent and personalized laundry experience. Through image acquisition under multiple angles and multiple lighting conditions, combined with deep learning technology, the robustness and accuracy of fabric recognition have been significantly improved, effectively avoiding the recognition errors that may occur under single lighting conditions, and ensuring that the washing machine can select the most appropriate washing program according to the actual fabric type of the clothes, thereby protecting the clothes and extending the life of the clothes. At the same time, it also improves the efficiency and effect of washing, so as to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the existing technology.

[0190] 2) The fabric identification device for washing machines based on light adjustment of the present application, the acquisition unit acquires multiple target brightness combinations when there are clothes in the washing machine, the target brightness combination includes the brightness of all light sources and the brightness of the light sources is inconsistent between different target brightness combinations; the first control unit controls each light source to illuminate according to each target brightness combination, and controls the image acquisition device to acquire an image, and obtains a first target image, and the first target image corresponds to the target brightness combination one by one; the first classification unit inputs the first target image into the neural network model for recognition, obtains a first recognition result, and determines the target fabric type at least according to the first recognition result, and the neural network model is used to recognize and classify according to the image, and output the fabric type. The present application can more accurately identify the fabric type of the clothes by collecting clothing images under different light source brightness combinations inside the washing machine and using the neural network model for fabric recognition. This method not only improves the recognition accuracy, but also can adapt to the recognition needs under different lighting conditions, providing users with a more intelligent and personalized laundry experience. Through image acquisition under multiple angles and multiple lighting conditions, combined with deep learning technology, the robustness and accuracy of fabric recognition have been significantly improved, effectively avoiding the recognition errors that may occur under single lighting conditions, and ensuring that the washing machine can select the most appropriate washing program according to the actual fabric type of the clothes, thereby protecting the clothes and extending the life of the clothes. At the same time, it also improves the efficiency and effect of washing, so as to solve the problem of low recognition accuracy of fabric recognition methods for multiple clothes in the existing technology.

[0191] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A washing machine fabric recognition method based on light adjustment, characterized in that: The washing machine includes a plurality of uniformly distributed light sources and at least one image acquisition device, and the method includes: When there are clothes in the washing machine, a plurality of target brightness combinations are acquired, wherein the target brightness combinations include the brightness of all the light sources and the brightness of the light sources is inconsistent between different target brightness combinations; Controlling each of the light sources to perform illumination according to each of the target brightness combinations, and controlling the image acquisition device to perform image acquisition to obtain a first target image, wherein the first target image corresponds to the target brightness combination one by one; The first target image is input into a neural network model for recognition to obtain a first recognition result, and the target fabric type is determined at least based on the first recognition result. The neural network model is used to recognize and classify according to the image and output the fabric type.

2. The method according to claim 1, characterized in that Get multiple target brightness combinations, including: Obtain multiple preset brightnesses, obtain the number of the light sources, and obtain a target number; The target brightness combination is obtained by performing random combinations based on the target quantity and the plurality of preset brightnesses and deleting repeated combinations.

3. The method according to claim 1, characterized in that Inputting the first target image into a neural network model for recognition to obtain a first recognition result includes: The first target image is spliced ​​according to RGB channels to obtain a second target image; The second target image is input into the neural network model to obtain the first recognition result.

4. The method according to claim 3, characterized in that The neural network model includes an input layer, a pooling layer and an output layer. The second target image is input into the neural network model to obtain the first recognition result, including: Performing feature extraction on the second target image through the shared convolution layer in the input layer to extract edges, textures and colors to obtain first target features, and performing multi-layer feature extraction on the first target features in the input layer through a residual block to obtain second target features; Performing a pooling operation on the second target feature through the pooling layer to obtain a third target feature; The third target feature is classified by the fully connected network in the output layer to obtain a target output value, and the target output value is processed by the sigmoid function in the output layer to obtain the first recognition result, and the target output value is used to characterize the probability distribution of each fabric type.

5. The method according to claim 3, characterized in that: Before splicing the first target image according to RGB channels, the method further includes: Cropping the first target image so that the size of the first target image is a reference size; Normalizing the brightness of the first target image to map the brightness of the first target image to a preset interval; Performing image augmentation processing on the first target image, wherein the image augmentation processing includes one or more of random cropping, rotation, scaling, hue adjustment, and saturation adjustment.

6. The method according to claim 1, characterized in that After inputting the first target image into the neural network model for recognition and obtaining a first recognition result, the method further includes: Controlling the washing tub of the washing machine to rotate to a preset angle, controlling each of the light sources to illuminate according to each of the target brightness combinations, and controlling the image acquisition device to acquire an image to obtain a second target image; The first target image is input into a neural network model for recognition to obtain a second recognition result.

7. The method according to claim 6, characterized in that Determining the target fabric type at least according to the first recognition result includes: When the first recognition result is consistent with the second recognition result, determining the first recognition result as the target fabric type; In the case that the first recognition result is inconsistent with the second recognition result, the target fabric type is obtained by integrating the fabric types included in the first recognition result and the second recognition result.

8. A washing machine fabric recognition device based on light adjustment, characterized in that: The washing machine comprises a plurality of uniformly distributed light sources and at least one image acquisition device, the device comprising: an acquisition unit, configured to acquire, when there are clothes in the washing machine, a plurality of target brightness combinations, wherein the target brightness combinations include the brightness of all the light sources and the brightness of the light sources is inconsistent between different target brightness combinations; A first control unit, used to control each of the light sources to perform illumination according to each of the target brightness combinations, and to control the image acquisition device to perform image acquisition to obtain a first target image, wherein the first target image corresponds to the target brightness combination in a one-to-one manner; The first classification unit is used to input the first target image into the neural network model for recognition, obtain a first recognition result, and determine the target fabric type at least based on the first recognition result. The neural network model is used to recognize and classify according to the image and output the fabric type.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A washing machine, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include methods for executing any one of claims 1 to 7.

Images