Methods for controlling household appliances, computer equipment, storage media and software products

Abstract

This application discloses a control method, computer device, storage medium, and program product for household appliances. The control method includes: acquiring visual feature information captured by a camera and acquiring wireless feature information captured by a wireless communication module; wherein the visual feature information includes an RGB image of at least one household appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by a wearable device; determining a matching target household appliance based on the visual feature information and the wireless feature information; and sending a control command to the target household appliance. Through this method, precise control of household appliances can be achieved.

Description

Technical Field

[0001] This application relates to the field of smart wearable technology, and in particular to a control method for a household appliance, a computer device, a storage medium, and a program product. Background Technology

[0002] With the development of smart home appliances, the appliances in our homes are becoming increasingly diverse, and their control methods are also becoming more varied. In addition to traditional remote control and infrared control, these include WIFI control, Bluetooth control, voice control, gesture control, and so on.

[0003] While Wi-Fi and Bluetooth control allow for precise control of individual home appliances, voice and gesture control are prone to misoperation when dealing with a large number of appliances. For example, when using voice control, the command "turn on the air conditioner" cannot distinguish whether the user wants to turn on the living room air conditioner or the bedroom air conditioner. Similarly, when using voice control, the command "turn up the volume" cannot distinguish whether the user wants to turn up the volume of the TV or the smart speaker. Therefore, it is difficult to accurately operate and control home appliances. Summary of the Invention

[0004] This application provides a method for controlling household appliances, a computer device, a storage medium, and a program product, which can achieve precise control of household appliances.

[0005] To address the aforementioned technical problems, this application provides a method for controlling a household appliance, applied to a wearable device. The wearable device includes a camera and a wireless communication module. The control method includes: acquiring visual feature information collected by the camera and acquiring wireless feature information collected by the wireless communication module; wherein the visual feature information includes an RGB image of at least one household appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by the wearable device; determining a matching target household appliance based on the visual feature information and the wireless feature information; and sending a household appliance control command to the target household appliance.

[0006] The wearable device also includes an audio acquisition module, and the control method further includes: acquiring audio information acquired by the audio acquisition module; processing the audio information to obtain intent information; responding to the intent information to control the home appliance, and performing the steps of acquiring visual feature information acquired by the camera and acquiring wireless feature information acquired by the wireless communication module.

[0007] The process of processing audio information to obtain intent information includes: converting audio information into audio features; inputting the audio features into a pre-trained deep learning model to obtain output candidate text; and extracting information from the candidate text to obtain intent information.

[0008] The process of determining a matching target home appliance based on visual and wireless feature information includes: determining at least one candidate home appliance based on visual and wireless feature information; filtering the at least one candidate home appliance based on intent information to determine at least one home appliance to be evaluated; and evaluating the at least one home appliance to be evaluated based on visual and wireless feature information to determine a matching target home appliance.

[0009] The process of determining a matching target home appliance based on visual and wireless feature information includes: determining a first score for at least one home appliance to be evaluated based on visual feature information; determining a second score for at least one home appliance to be evaluated based on wireless feature information; determining a third score for at least one home appliance to be evaluated based on the matching degree of visual and wireless feature information; and determining a fourth score for at least one home appliance to be evaluated based on detection confidence. The detection confidence is determined by a target detection model used to perform target detection on RGB images. A comprehensive score for at least one home appliance to be evaluated is determined based on the first, second, third, and fourth scores. Finally, a matching target home appliance is determined based on the comprehensive score.

[0010] The first score of at least one household appliance to be evaluated is determined based on visual feature information, including: determining the position score of the household appliance to be evaluated based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image; and determining the depth score of the household appliance to be evaluated based on the ratio of the pixel size of the household appliance to the standard size in the RGB image; and determining the first score of the household appliance to be evaluated based on the position score and the depth score.

[0011] The visual feature information also includes a depth image of at least one household appliance. The first score of at least one household appliance to be evaluated is determined based on the visual feature information, including: determining the position score of the household appliance to be evaluated based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image; and determining the depth score of the household appliance to be evaluated based on the depth information of the household appliance to be evaluated in the depth image; and determining the first score of the household appliance to be evaluated based on the position score and the depth score.

[0012] The method of determining a second score for at least one household appliance to be evaluated based on wireless characteristic information includes: acquiring the signal strength of a radio signal transmitted by the household appliance to be evaluated received by a wearable device; determining the distance between the wearable device and the household appliance to be evaluated based on the signal strength; and determining the second score for the household appliance to be evaluated based on the distance between the wearable device and the household appliance to be evaluated.

[0013] The method of determining a third score for at least one household appliance to be evaluated based on the matching degree of visual feature information and wireless feature information includes: determining a first distance between the wearable device and the household appliance to be evaluated based on the RGB image, and determining a second distance between the wearable device and the household appliance to be evaluated based on the signal strength of the radio signal transmitted by the household appliance to be evaluated received by the wearable device; and determining a third score for at least one household appliance to be evaluated based on the matching degree of the first distance and the second distance.

[0014] The visual feature information also includes a depth image of at least one household appliance. A third score for at least one household appliance to be evaluated is determined based on the matching degree of the visual feature information and the wireless feature information. This includes: determining a third distance between the wearable device and the household appliance to be evaluated based on the depth image, and determining a fourth distance between the wearable device and the household appliance to be evaluated based on the signal strength of the radio signal transmitted by the household appliance to be evaluated received by the wearable device; and determining a third score for at least one household appliance to be evaluated based on the matching degree of the third distance and the fourth distance.

[0015] The process of determining a fourth score for at least one household appliance to be evaluated based on the detection confidence level includes: obtaining an initial detection confidence level; adjusting the initial detection confidence level according to the occlusion status of the household appliance to be evaluated to obtain a final detection confidence level; and determining a fourth score for the household appliance to be evaluated based on the final detection confidence level.

[0016] The process of determining a comprehensive score for at least one household appliance to be evaluated based on a first score, a second score, a third score, and a fourth score includes: determining weighting coefficients based on whether the visual feature information contains depth images and the current environmental conditions; and using the weighting coefficients to weight the first score, the second score, the third score, and the fourth score to determine a comprehensive score for at least one household appliance to be evaluated.

[0017] After determining the matching target home appliance based on the comprehensive score, the process also includes: verifying the detection confidence level of the target home appliance; and in response to the successful verification of the detection confidence level of the target home appliance, executing the step of sending home appliance control commands to the target home appliance.

[0018] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide a computer device, which includes a processor and a memory coupled to the processor; wherein the memory is used to store computer programs, and the processor is used to execute the computer programs to realize the control method of household appliances as described above.

[0019] The computer device in question is a wearable device.

[0020] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium that stores a computer program, which, when executed by a processor, implements the control method for household appliances as described above.

[0021] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer program product, wherein the computer program product includes computer instructions, which are executed by a processor to implement the control method of household appliances as described above.

[0022] Unlike existing technologies, the home appliance control method of this application includes: acquiring visual feature information captured by a camera and acquiring wireless feature information captured by a wireless communication module; wherein, the visual feature information includes an RGB image of at least one home appliance, and the wireless feature information includes the signal strength of radio signals transmitted by at least one home appliance received by a wearable device; determining a matching target home appliance based on the visual feature information and the wireless feature information; and sending a home appliance control command to the target home appliance. Through the above method, various information based on the location of the home appliance can be obtained using the visual feature information and the wireless feature information, such as whether it is located in the center of the field of vision, whether it is the closest, etc. Based on these features, further evaluation can be performed on multiple home appliances that can be controlled by the home appliance control command, selecting the home appliance that the user most likely needs to control, thereby avoiding misoperation and achieving precise control of the home appliance. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the first embodiment of the control method for household appliances provided in this application; Figure 2 This is a schematic diagram of target detection based on RGB images in one embodiment; Figure 3This is a flowchart illustrating the second embodiment of the control method for household appliances provided in this application; Figure 4 yes Figure 3 A flowchart illustrating an embodiment of step S31; Figure 5 yes Figure 3 A flowchart illustrating step S36 of the first embodiment; Figure 6 yes Figure 3 A flowchart illustrating step S36 of the second embodiment; Figure 7 yes Figure 3 A flowchart illustrating step S36 of the third embodiment; Figure 8 yes Figure 3 A flowchart illustrating step S36 of the fourth embodiment; Figure 9 yes Figure 3 A flowchart illustrating step S36 of the fifth embodiment; Figure 10 yes Figure 3 A flowchart illustrating step S36 of the sixth embodiment; Figure 11 It is a schematic diagram of the interaction between various terminals in an application scenario; Figure 12 This is a schematic diagram of the structure of an embodiment of the computer device provided in this application; Figure 13 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0025] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that any embodiment described herein can be combined with associated technical features in other embodiments.

[0026] There are various ways to control home appliances, including traditional remote control infrared control, as well as WIFI control, Bluetooth control, voice control, gesture control, and so on. This application uses voice control as an example for illustration. In some scenarios, when using voice control, the voice command "turn on the air conditioner" cannot distinguish whether the user needs to turn on the living room air conditioner or the bedroom air conditioner. Similarly, when using voice control, the voice command "turn up the volume" cannot distinguish whether the user needs to turn up the volume of the TV or the smart speaker. Therefore, it is difficult to accurately operate and control home appliances.

[0027] See Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the control method for household appliances provided in this application. The method includes: Step S11: Obtain visual feature information captured by the camera and wireless feature information captured by the wireless communication module.

[0028] Understandably, the control method for household appliances of this application is applied to wearable devices, which include a camera and a wireless communication module. Optionally, some wearable devices are equipped with a monocular camera for capturing RGB images, while others are equipped with a binocular camera for simultaneously capturing RGB and depth images. Optionally, the wireless communication module can be a Bluetooth module, such as a BLE (Bluetooth Low Energy) module, also known as Bluetooth Low Energy.

[0029] Optionally, the wearable device in this embodiment can be smart glasses, smartwatches, fitness trackers, etc. Taking smart glasses as an example, the camera can generally be set on the frame between the left and right frames, or on the frame outside the left and right frames, and the wireless communication module can be set on the temples; there are no restrictions here.

[0030] Step S12: Determine the matching target home appliance based on visual and wireless feature information.

[0031] The visual feature information includes an RGB image of at least one household appliance.

[0032] Optionally, in one embodiment, the RGB image can be further uploaded to a cloud server, and the cloud server inputs the RGB image into a target detection model for target detection, which can identify the outline, type, model, etc. of household appliances.

[0033] like Figure 2 As shown, Figure 2 This is a schematic diagram of target detection based on RGB images in one embodiment. In one embodiment, a rectangular frame can be used to represent the outline of the detected household appliance. Figure 2 In the image, the dashed box represents the RGB image, and the solid box represents the outline of the detected household appliance. Household appliance A and household appliance B are detected in the image. Taking household appliance A as an example, (x1, y1) represents the lower left coordinate, and (x2, y2) represents the upper right coordinate.

[0034] In one embodiment, the position of the household appliance in the entire RGB image is obtained from visual feature information. For example, the center point of the outline of the household appliance can be determined first, then the center point of the RGB image can be determined, and the distance between the center point of the outline of the household appliance and the center point of the RGB image can be calculated to determine whether the household appliance is located in the center of the entire RGB image, that is, whether it is located in the center of the user's field of vision (the more central it is, the more likely it is to be a household appliance that the user needs to operate and control).

[0035] In another embodiment, the pixel size of the home appliance is obtained from the visual feature information. Since the model of the home appliance can be detected in the previous target detection, the standard size of the home appliance can be further obtained. By the ratio of the pixel size to the standard size of the home appliance, the distance between the home appliance and the wearable device can be estimated (the closer the distance, the more likely it is to be the home appliance that the user needs to operate and control).

[0036] The wireless characteristic information includes the signal strength of a radio signal received by the wearable device from at least one household appliance.

[0037] In one embodiment, the distance between the wearable device and at least one household appliance can be determined based on the signal strength of the radio signal received by the wearable device from at least one household appliance. RSSI (Received Signal Strength Indicator) is a metric used to represent the strength of a received wireless signal (such as Wi-Fi, Bluetooth, LoRa, etc.), typically expressed in dBm (decibels per milliwatt). RSSI is used to roughly estimate the distance between the wearable device and the household appliance based on a signal attenuation model (such as a logarithmic distance path loss model). The following formula model is generally used:

[0038] Where n is the path loss index, A is the RSSI reference value at 1 meter, and d is the distance.

[0039] In another embodiment, since the locations of most household appliances are fixed, a household appliance map can be pre-built for the entire household. Then, based on the signal strength of the radio signal received by the wearable device from at least one household appliance, the distance between the wearable device and at least one household appliance can be determined, thereby determining the location (coordinates) of the wearable device in the entire household.

[0040] Using the methods described above, visual feature information can determine the location of home appliances and the distance between the wearable device and the home appliances. Wireless feature information can also determine the distance between the wearable device and the home appliances. By comprehensively judging these data, the target home appliance that the user is most likely to operate and control can be identified from among multiple home appliances. For example, the home appliance located in the center of the RGB image, or the home appliance closest to the wearable device, or a combination of the two methods.

[0041] Step S13: Send appliance control commands to the target home appliance.

[0042] Understandably, the target household appliance is identified in step S12, and the appliance control command can be determined based on the specific control method.

[0043] In one embodiment, the appliance control command is generated by a voice command. For example, the wearable device includes an audio acquisition module that acquires the voice command "turn on the air conditioner" and generates an appliance control command to turn on the air conditioner based on this command. The appliance control command is then sent to the target home appliance (air conditioner) determined in step S12.

[0044] In another embodiment, the appliance control command is generated by gesture command. For example, the camera of the wearable device captures the user's gesture and obtains the voice command "turn up the volume". Based on this, the appliance control command to turn up the volume is generated and then sent to the target home appliance (television) determined in step S12.

[0045] Unlike existing technologies, the home appliance control method provided in this embodiment includes: acquiring visual feature information collected by a camera and acquiring wireless feature information collected by a wireless communication module; wherein, the visual feature information includes an RGB image of at least one home appliance, and the wireless feature information includes the signal strength of radio signals transmitted by at least one home appliance received by a wearable device; determining a matching target home appliance based on the visual feature information and the wireless feature information; and sending a home appliance control command to the target home appliance. Through the above method, various information based on the location of the home appliance can be obtained using the visual feature information and the wireless feature information, such as whether it is located in the center of the field of vision, whether it is the closest, etc. Based on these features, further evaluation can be performed on multiple home appliances that can be controlled by the home appliance control command, selecting the home appliance that the user most likely needs to control, thereby avoiding misoperation and achieving precise control of the home appliance.

[0046] See Figure 3 , Figure 3 This is a flowchart illustrating a second embodiment of the control method for household appliances provided in this application. The method includes: Step S31: Obtain the audio information acquired by the audio acquisition module.

[0047] Wearable devices also include audio acquisition modules, such as microphone arrays. For example, in smart glasses, the audio acquisition module can be mounted on the temples of the glasses.

[0048] Step S32: Process the audio information to obtain intent information.

[0049] Among them, intent information is used to represent the user's intent for home control.

[0050] Optionally, such as Figure 4 As shown, Figure 4 yes Figure 3 A flowchart illustrating an embodiment of step S32, which may include: Step S321: Convert audio information into audio features.

[0051] In one embodiment, the audio acquisition module acquires multi-channel audio data, executes a beamforming algorithm to enhance the speech signal from the user's direction, then uses AEC (Acoustic Echo Cancellation) to remove the sound played by the device itself, then performs NS (Noise Suppression) to filter out ambient noise, and finally uses VAD (Voice Activity Detection) to determine the start and end times of the effective speech segment, outputting a clear single-channel speech signal for speech recognition.

[0052] Then, the audio information is transformed into audio features. Converting the raw audio waveform into a feature representation is a key step in speech processing, and different features are suitable for different tasks (recognition, synthesis, enhancement, etc.). Audio features can include time-domain features, frequency-domain features, and deep learning features. Time-domain features can include short-time energy, zero-crossing rate, amplitude envelope, etc. Frequency-domain features can include short-time Fourier transform, Melton spectrogram, Mel frequency cepstral coefficients, filter bank features, etc. Deep learning features can include spectrogram, constant Q-transform, chroma features, etc.

[0053] Step S322: Input the audio features into a pre-trained deep learning model to obtain the output candidate text.

[0054] Taking the ASR (Automatic Speech Recognition) model as an example, ASR is a technology that converts speech signals into text. The processing process of traditional ASR models includes: audio → feature extraction (MFCC) → acoustic model (GMM) → HMM state decoding → language model (N-gram) → text. The processing process of deep learning ASR models includes: audio → feature extraction (FBank / Mel) → acoustic model (DNN / CNN / RNN) → decoder (CTC / Attention) → text.

[0055] Among them, the acoustic model is the core of the entire ASR model, and it can adopt RNN models, CNN models, RNN-T (RNN-Transducer) models, Transformer models, hybrid models, and so on.

[0056] Step S323: Extract information from the candidate text to obtain intent information.

[0057] Optionally, identify whether it belongs to home appliance control according to the intent information. If it belongs to home appliance control, extract relevant information, including home appliance type (if there is no type, confirm through subsequent image recognition), adjustment action, and spatial position (if there is no position information, default to near reference).

[0058] For example, home appliance types can include: TV sets, air conditioners, lighting fixtures, water heaters, projectors, air purifiers, electric fans, etc.

[0059] For example, adjustment actions can include: Turn on, turn off, turn up, turn down, fast forward, pause, etc.

[0060] For example, spatial positions can include: Spatial position dictionary = { "Near reference": ["this", "this one", "in front", "ahead"], "Far reference": ["that", "that one", "behind", "at the back"], "Directional reference - left": ["left side", "left hand side", "left face", "left hand side"], "Directional reference - right": ["right side", "right hand side", "right face", "right hand side"], "Directional reference - up": ["above", "upward", "upper"], "Directional reference - down": ["below", "downward", "lower"].

[0061] Step S33: In response to the intent information indicating control of the home appliance, acquire the visual feature information collected by the camera and the wireless feature information collected by the wireless communication module.

[0062] The visual feature information includes an RGB image of at least one household appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by the wearable device.

[0063] Step S34: Based on visual feature information and wireless feature information, identify at least one candidate home appliance.

[0064] Understandably, the candidate household appliances here refer to all household appliances covered by visual and wireless feature information. For example, all household appliances covered by target detection through RGB images and by radio signals transmitted by all household appliances obtained through wireless feature information. It is worth noting that since target detection through RGB images can determine the type and model of a household appliance, and radio signals contain identification information of the household appliance transmitting the signal, it is possible that the same household appliance may appear in both the visual and wireless feature information detection ranges simultaneously, or a household appliance may appear only in the visual feature information detection range, or a household appliance may appear only in the wireless feature information detection range. Regardless of the specific case, all such household appliances fall under the candidate household appliances category in this step.

[0065] Step S35: Filter at least one candidate home appliance based on intent information to identify at least one home appliance to be evaluated.

[0066] Understandably, the candidate home appliances identified in step S33 using visual and wireless feature information can include many, such as air conditioner A, air conditioner B, television, smart speaker, and water heater. If the user needs to "turn on the air conditioner" based on the intent information, then television, smart speaker, and water heater can be filtered out, and air conditioner A and air conditioner B can be used as the home appliances to be evaluated.

[0067] Step S36: Evaluate at least one household appliance to be evaluated based on visual feature information and wireless feature information to determine the matching target household appliance.

[0068] The process involves determining a first score for at least one household appliance to be evaluated based on visual feature information, a second score based on wireless feature information, a third score based on the matching degree between visual and wireless feature information, and a fourth score based on detection confidence. The detection confidence is determined by a target detection model used to perform target detection on RGB images. A comprehensive score for at least one household appliance to be evaluated is then determined based on the first, second, third, and fourth scores.

[0069] The first score, second score, third score, and fourth score will be introduced below through several examples.

[0070] like Figure 5 As shown, Figure 5 yes Figure 3 The flowchart of the first embodiment of step S36 is shown in this embodiment. This embodiment takes a monocular camera as an example. The monocular camera is used to acquire RGB images. Step S35 may include: Step S361a: Determine the position score of the household appliance to be evaluated based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image, and determine the depth score of the household appliance to be evaluated based on the ratio of the pixel size of the household appliance to the standard size in the RGB image.

[0071] The location score indicates whether the household appliance being evaluated is located in the center of the RGB image; the closer it is to the center, the higher the location score. Figure 2 Taking a household appliance A as an example, (x1, y1) represents the lower left coordinate, and (x2, y2) represents the upper right coordinate. The details are as follows: First, calculate the center of household appliance A: Center_x = (x1 + x2) / 2 Center_y = (y1 + y2) / 2 Next, calculate the distance between the center of household appliance A and the center of the RGB image: Distance_from_center = √[(Center_x - Image_center_x)² + (Center_y -Image_center_y)²] Where Image_center_x and Image_center_y are the coordinates of the center of the RGB image.

[0072] Finally, the above results are reverse normalized: Center_position_score = 1 - (Distance_from_center / Max_distance) Where Max_distance = √[(width / 2)² + (height / 2)²] Where width is the width of the RGB image and height is the height of the RGB image.

[0073] The depth score is used to represent the distance between the wearable device and the home appliance being evaluated. Figure 2 Taking a household appliance A as an example, (x1, y1) represents the lower left coordinate, and (x2, y2) represents the upper right coordinate. The details are as follows: First, calculate the pixel size of household appliance A: height_pixels = y2 - y1 area_pixels = (x2 - x1) × (y2 - y1) Then, based on the type or model of the identified home appliance, the system queries the equipment information database for the H_standard (standard height, unit: meters) of that type of equipment to calculate the relative depth index: Visual_depth = H_standard / height_pixels Understandably, the smaller the Visual_depth value mentioned above, the larger the device appears in the image, and the closer the distance.

[0074] Finally, normalization was performed on all the household appliances to be evaluated: depth_min = min(Visual_depth of all household appliances to be evaluated) depth_max = max(Visual_depth of all household appliances to be evaluated) Visual_size_score = 1 - (Visual_depth - depth_min) / (depth_max -depth_min) Results: The device with the highest visual score is 1.0, and the device with the lowest visual score is 0.0.

[0075] Step S361b: Determine the first rating of the household appliance to be evaluated based on the location rating and depth rating.

[0076] Optionally, in one embodiment, the location score and depth score are weighted to obtain a first score F1: F1 = w_center × Center_position_score + w_size × Visual_size_score By default, w_center = 0.40, w_size = 0.60, and w_center + w_size = 1. The reason for this setting is that users will face the target device, and the center position reflects the user's attention (weight 0.40), but the closer the device, the stronger the user's control intention, and the size depth is more important (weight 0.60).

[0077] For example, air conditioner A (leaning to the left but larger): Center_score = 0.75 (Off-center) Visual_size_score = 1.0 (Visual maximum, closest) F1 = 0.40×0.75 + 0.60×1.0 = 0.90 Air conditioner B (centered but smaller): Center_score = 0.95 (close to the center) Visual_size_score = 0.0 (Minimum visual size, farthest visual range) F1 = 0.40×0.95 + 0.60×0.0 = 0.38 As shown in the first score F1 above, although air conditioner B is more centrally located, air conditioner A scores higher due to its closer proximity. Combining the device's position and visual size within the field of vision, and comprehensively judging the user's intention and the device's distance, air conditioner A can ultimately be selected.

[0078] like Figure 6 As shown, Figure 6 yes Figure 3 This is a flowchart illustrating step S36 of the second embodiment. This embodiment uses a binocular camera as an example. The binocular camera is used to acquire RGB images and depth images. Step S36 may include: Step S362a: Determine the position score of the household appliance to be evaluated based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image, and determine the depth score of the household appliance to be evaluated based on the depth information of the household appliance to be evaluated in the depth image.

[0079] The location score indicates whether the household appliance being evaluated is located in the center of the RGB image; the closer it is to the center, the higher the location score. Figure 2Taking a household appliance A as an example, (x1, y1) represents the lower left coordinate, and (x2, y2) represents the upper right coordinate. The details are as follows: First, calculate the center of household appliance A: Center_x = (x1 + x2) / 2 Center_y = (y1 + y2) / 2 Next, calculate the distance between the center of household appliance A and the center of the RGB image: Distance_from_center = √[(Center_x - Image_center_x)² + (Center_y -Image_center_y)²] Where Image_center_x and Image_center_y are the coordinates of the center of the RGB image.

[0080] Finally, the above results are reverse normalized: Center_position_score = 1 - (Distance_from_center / Max_distance) Where Max_distance = √[(width / 2)² + (height / 2)²] Where width is the width of the RGB image and height is the height of the RGB image.

[0081] The depth score is used to represent the distance between the wearable device and the home appliance being evaluated. Figure 2 Taking a household appliance A as an example, (x1, y1) represents the lower left coordinate, and (x2, y2) represents the upper right coordinate. The details are as follows: Based on the position of household appliance A in the RGB image, extract the corresponding depth region at the corresponding position in the depth image: depth_region = depth_map[y1:y2, x1:x2] Extract the corresponding depth region: Then, for the depth value statistical processing, invalid values ​​(0 or negative values) are first removed: valid_depths = depth_region[depth_region>0] Then calculate the median: device_depth = median(valid_depths) Normalize the distance score: effective_range = 10 (effective distance is controlled within 10 meters) Depth_score = 1 - (device_depth / effective_range) Depth_score = max(0, min(1, Depth_score)) For example, if the equipment depth is 6.0 meters, then Depth_score = 1 - 6.0 / 10 = 0.40 Step S362b: Determine the first rating of the household appliance to be evaluated based on the location rating and depth rating.

[0082] Optionally, in one embodiment, the location score and depth score are weighted to obtain a first score F1: F1 = w_center × Center_position_score + w_depth × Depth_score By default, w_center = 0.25, w_depth = 0.75, and w_center + w_depth = 1. The reason for this setting is that the depth sensor provides high-precision distance information, which is the main basis for judging the user's intention (weight 0.70). The center position still reflects the user's attention direction, but its importance is relatively reduced (weight 0.30). Compared with the implementation of a monocular camera, the monocular depth weight is 0.60 (estimated), and the binocular depth weight is 0.70 (direct measurement is more reliable).

[0083] For example, air conditioner A (slightly to the left but very close): Center_score = 0.75 (Off-center) Depth_score = 0.85 (2.5 meters, very close) F1 = 0.25×0.75 + 0.75×0.85 = 0.825 Air conditioner B (centrally located but relatively far away): Center_score = 0.95 (close to the center) Depth_score = 0.40 (6.0 meters, relatively far) F1 = 0.25×0.95 + 0.75×0.40 = 0.5375 As can be seen from the first score F1 obtained above, air conditioner A scores higher because it is closer.

[0084] like Figure 7 As shown, Figure 7 yes Figure 3 A flowchart illustrating step S36 of the third embodiment is provided. Step S36 may include: Step S363a: Obtain the signal strength of the radio signal transmitted by the home appliance to be evaluated, received by the wearable device.

[0085] Optionally, the signal strength here is taken as RSSI.

[0086] First, RSSI values ​​are collected 5 times within 500ms, and outliers (maximum and minimum) are removed. The remaining RSSI values ​​are then weighted and averaged. RSSI_filtered = Σ(RSSI_i × w_i) Step S363b: Determine the distance between the wearable device and the household appliance to be evaluated based on the signal strength.

[0087] Alternatively, the path loss model is still used here to calculate the distance between the wearable device and the home appliance to be evaluated: d_rssi = 10^[(RSSI0- RSSI_filtered) / (10×n)] Where RSSI0 = -59 dBm (reference value at 1 meter), n = 2.0~3.0 (environmental path loss index).

[0088] Step S363c: Determine a second rating for the household appliance to be evaluated based on the distance between the wearable device and the appliance.

[0089] The distance calculated above is normalized to obtain the second score: RSSI_distance_score = 1 - (d_rssi / d_max) F2 = RSSI_distance_score Where RSSI_distance_score = max(0, min(1, 1 - d_rssi / 10)), for example: distance 1 meter → Score = 0.90, distance 3 meters → Score = 0.70, distance 5 meters → Score = 0.50, distance 10 meters → Score = 0.00.

[0090] like Figure 8 As shown, Figure 8 yes Figure 3 This is a flowchart illustrating step S36 of the fourth embodiment. This embodiment uses a monocular camera as an example. The monocular camera is used to acquire RGB images. Step S36 may include: Step S364a: Determine a first distance between the wearable device and the home appliance to be evaluated based on the RGB image, and determine a second distance between the wearable device and the home appliance to be evaluated based on the signal strength of the radio signal transmitted by the home appliance to be evaluated received by the wearable device.

[0091] The first distance is based on the camera imaging principle: d_visual = (H_standard × f_calibrated) / height_pixels Where H_standard: standard device height (e.g., 60cm for an air conditioner), f_calibrated: calibrated focal length (in pixels), and height_pixels: the device's pixel height in an RGB image.

[0092] The second distance is calculated based on the RSSI value: d_rssi = 10^[(RSSI0- RSSI_filtered) / (10×n)] Where RSSI0 = -59 dBm (reference value at 1 meter), n = 2.0~3.0 (environmental path loss index).

[0093] Step S364b: Determine a third rating for at least one household appliance to be evaluated based on the matching degree of the first distance and the second distance.

[0094] Optionally, in one embodiment, the matching degree of the first distance and the second distance can be determined by calculating the relative error between the first distance and the second distance, thereby obtaining a third score.

[0095] The relative error between the first distance and the second distance is: consistency = 1 - |d_visual - d_rssi| / max(d_visual, d_rssi) F3 = consistency For example, d_visual=2.5 meters, d_rssi=2.8 meters → error 0.3 meters → Score = 0.88 (high consistency); as another example, d_visual=2.0 meters, d_rssi=5.0 meters → error 3.0 meters → Score = 0.40 (low consistency).

[0096] Understandably, a higher third score indicates that the assessments of the two distances are similar. For example, if the first distance is 2.3 meters, the second distance is 2.6 meters, and the third score F3 = 0.90, then both the first and second distances are considered relatively reliable. Conversely, if the first distance is 2.0 meters, the second distance is 6.0 meters, and the third score F3 = 0.33, then it is likely due to RRSI interference or visual false detection.

[0097] like Figure 9 As shown, Figure 9 yes Figure 3 This is a flowchart illustrating step S36 of the fifth embodiment. This embodiment uses a binocular camera as an example. The binocular camera is used to acquire RGB images and depth images. Step S36 may include: Step S365a: Determine the third distance between the wearable device and the home appliance to be evaluated based on the depth image, and determine the fourth distance between the wearable device and the home appliance to be evaluated based on the signal strength of the radio signal transmitted by the home appliance to be evaluated received by the wearable device.

[0098] The third distance d_depth is obtained from the depth image, which can be referred to in step S362a, and will not be repeated here.

[0099] The fourth distance d_rssi is calculated based on the RSSI value, which can be referred to in step S364a, and will not be repeated here.

[0100] Step S365b: Determine a third score for at least one household appliance to be evaluated based on the matching degree between the third and fourth distances.

[0101] Optionally, in one embodiment, the matching degree of the third distance and the fourth distance can be determined by calculating the relative error between the third distance and the fourth distance, thereby obtaining the third score.

[0102] The relative errors between the third and fourth distances are as follows: consistency = 1 - |d_depth - d_rssi| / max(d_depth, d_rssi) F3 = consistency In addition, consistency can be graded based on the third rating, F3: F3 > 0.85, consistency_level = "HIGH" (highly consistent) 0.85 > F3 > 0.65, consistency_level = "MEDIUM" (good consistency) 0.65 > F3 > 0.45, consistency_level = "LOW" (disagreement exists) 0.45 > F3, consistency_level = "CONFLICTING" (Severe conflict) In comparison with the above Figure 8 and Figure 9 In the embodiments of the monocular camera, the visual distance estimation and RSSI distance estimation are compared and found to be similar in accuracy. In the embodiments of the binocular camera, the depth distance estimation is more accurate than the RSSI distance estimation. Therefore, in the embodiments of the binocular camera, when the matching degree between the third distance and the fourth distance is low (the third score F3 is low), the result of the third distance, i.e., the depth distance estimation, can be trusted more.

[0103] like Figure 10 As shown, Figure 10 yes Figure 3 A flowchart illustrating step S36 of the sixth embodiment is provided. Step S36 may include: Step S366a: Obtain the initial detection confidence.

[0104] The initial detection confidence score is output by the target detection model: confidence_raw ∈ [0, 1] The initial detection confidence level can be determined as follows: Clear detection: 0.85 - 0.95 General testing: 0.70 - 0.85 Fuzzy detection: 0.60 - 0.70 Low confidence level: <0.60 Step S366b: Adjust the initial detection confidence level according to the obstruction situation of the household appliance to be evaluated to obtain the final detection confidence level.

[0105] Optionally, in one embodiment, the obstruction status of a household appliance can be determined by the area it obstructs; the larger the obstruction area, the lower the detection confidence level. For example: If the occlusion area is less than 25%, then confidence_adjusted = confidence_raw With an occlusion area of ​​25% - 50%, confidence_adjusted = confidence_raw × 0.85 If the occlusion area is greater than 50%, then confidence_adjusted = confidence_raw × 0.60 Step S366c: Determine the fourth score of the household appliance to be evaluated based on the final test confidence level.

[0106] The fourth rating, F4, is limited to the range [0, 1]. Confidence_score = confidence_adjusted F4 = Confidence_score Understandably, the above Figures 5-10 The embodiments described the methods for obtaining the first score, the second score, the third score, and the fourth score, respectively. Here, we will further describe how to determine the comprehensive score of at least one household appliance to be evaluated based on the first score, the second score, the third score, and the fourth score.

[0107] In one embodiment, weighting coefficients are determined based on whether the visual feature information includes a depth image and the current environmental conditions. The weighting coefficients are then used to weight the first score, the second score, the third score, and the fourth score to determine the comprehensive score of at least one household appliance to be evaluated.

[0108] Understandably, the first, second, and third scores involve depth images (from a monocular or binocular camera). The weights of these scores differ depending on the application scenario, whether or not depth images are included. In other embodiments, the allocation of weighting coefficients may also consider environmental factors, such as indoor / outdoor conditions, weather, and brightness.

[0109] Overall score = w1×F1 + w2×F2 + w3×F3 + w4×F4 Among them, w1, w2, w3, and w4 are weighting coefficients, and w1 + w2 + w3 + w4 = 1.

[0110] Optionally, in an indoor application scenario, the weights of w1, w2, w3, and w4 are allocated as follows:

[0111] Step S37: Send appliance control commands to the target home appliance.

[0112] In one embodiment, the appliance control command is generated from a voice command. For example, the wearable device includes an audio acquisition module that acquires the voice command "turn on the air conditioner," generates an appliance control command to turn on the air conditioner based on this, and then sends the appliance control command to the target home appliance (air conditioner).

[0113] In another embodiment, the appliance control command is generated from a gesture command. For example, the camera of the wearable device captures the user's gesture and obtains the voice command "turn up the volume". Based on this, an appliance control command to turn up the volume is generated and then sent to the target home appliance (television).

[0114] Optionally, in one embodiment, before step S37, the detection confidence of the target household appliance may be further verified, for example, the detection confidence of the target household appliance may be verified; in response to the successful verification of the detection confidence of the target household appliance, step S37 is executed.

[0115] Specifically, a confidence threshold can be set here. When the detection confidence of the target household appliance is greater than the confidence threshold, step S37 is executed.

[0116] The following is combined Figure 11 Let me introduce an application scenario. Figure 11 This is a schematic diagram of the interactions between different terminals in an application scenario.

[0117] 1. User voice input: "It's a bit hot, turn on the air conditioner."

[0118] The microphone array collects audio information, performs beamforming algorithm to enhance the voice signal from the user's direction, then removes the sound played by the device itself through AEC, then performs NS to filter out environmental noise, and finally uses VAD to determine the start and end times of the effective voice segment, outputting a clear single-channel voice signal for speech recognition.

[0119] 2. The text "Turn on the air conditioner" is obtained through speech recognition.

[0120] The ASR model was used for processing.

[0121] 3. Intent understanding.

[0122] Type: Air conditioner; Action: Turn on; Location: Near finger.

[0123] 4. Trigger photo capture.

[0124] RGB image (+depth image), object detection: air conditioner A, air conditioner B, television (based on intent filtering).

[0125] 5. Simultaneous BLE scan.

[0126] Three devices were detected: Air Conditioner A, Air Conditioner B, and Television (based on intent filtering).

[0127] 6. Scoring calculation.

[0128]

[0129] 7. Adaptive weight adjustment.

[0130] w1, w2, w3, and w4 are adaptively adjusted according to the detection environment.

[0131] 8. Overall score and ranking.

[0132] Overall score = w1×F1 + w2×F2 + w3×F3 + w4×F4 Air conditioner A: 0.82 Air conditioner B: 0.58 9. Confidence verification.

[0133] Air conditioner A: 0.82 > 0.75, which meets the confidence level verification condition.

[0134] 10. Send appliance control commands.

[0135] In one embodiment, the wearable device can send the decision to the cloud server, and the cloud server can issue appliance control commands to air conditioner A. In another embodiment, the wearable device can directly issue appliance control commands to air conditioner A via a wireless communication module (such as Bluetooth).

[0136] See Figure 12 , Figure 12 This is a schematic diagram of an embodiment of a computer device provided in this application. The computer device 120 includes a processor 121 and a memory 122 coupled to the processor 121.

[0137] The memory 122 is used to store computer programs, and the processor 121 is used to execute the computer programs to implement the following control method for household appliances: The system acquires visual feature information captured by a camera and wireless feature information captured by a wireless communication module. The visual feature information includes an RGB image of at least one household appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by the wearable device. The system determines a matching target household appliance based on the visual feature information and the wireless feature information, and sends a control command to the target household appliance.

[0138] Optionally, in one embodiment, the processor 121 is configured to execute a computer program to implement a method for controlling a home appliance as follows: acquiring audio information acquired by an audio acquisition module; processing the audio information to obtain intent information; in response to the intent information indicating control of the home appliance, acquiring visual feature information acquired by a camera and wireless feature information acquired by a wireless communication module; wherein the visual feature information includes an RGB image of at least one home appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one home appliance received by a wearable device; determining at least one candidate home appliance based on the visual feature information and the wireless feature information; filtering the at least one candidate home appliance based on the intent information to determine at least one home appliance to be evaluated; evaluating the at least one home appliance to be evaluated based on the visual feature information and the wireless feature information to determine a matching target home appliance; and sending a home appliance control command to the target home appliance.

[0139] Optionally, in one embodiment, a first score for at least one household appliance to be evaluated is determined based on visual feature information; a second score for at least one household appliance to be evaluated is determined based on wireless feature information; a third score for at least one household appliance to be evaluated is determined based on the matching degree of visual feature information and wireless feature information; and a fourth score for at least one household appliance to be evaluated is determined based on detection confidence; wherein, the detection confidence is determined by a target detection model, which is used to perform target detection on RGB images; a comprehensive score for at least one household appliance to be evaluated is determined based on the first score, second score, third score, and fourth score; and a matching target household appliance is determined based on the comprehensive score.

[0140] Optionally, in one embodiment, the computer device 120 can be a wearable device, such as smart glasses, a smartwatch, or a fitness tracker. In a further embodiment, the wearable device 120 may also include a camera, a wireless communication module, and a voice acquisition module. Taking smart glasses as an example, the camera can generally be mounted on the frame between the left and right frames, or on the outer side of the frame; the wireless communication module can be mounted on the temple; and the voice acquisition module can be mounted on the temple—there are no limitations on this.

[0141] Optionally, in another embodiment, the computer device 120 can be other devices connected to the wearable device, such as a server, mobile phone, computer, etc. The wearable device may include a camera, a wireless communication module, a voice acquisition module, etc. Taking smart glasses as an example, the smart glasses' camera collects visual feature information, the wireless communication module collects wireless feature information, and then sends the visual feature information and wireless feature information to the server, which then determines the target home appliance.

[0142] See Figure 13, Figure 13 This is a schematic diagram of an embodiment of the computer-readable storage medium 130 provided in this application. The computer-readable storage medium 130 stores a computer program 131, which, when executed by a processor, implements the following method for controlling household appliances: The system acquires visual feature information captured by a camera and wireless feature information captured by a wireless communication module. The visual feature information includes an RGB image of at least one household appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by the wearable device. The system determines a matching target household appliance based on the visual feature information and the wireless feature information, and sends a control command to the target household appliance.

[0143] Optionally, in one embodiment, when the computer program 131 is executed by the processor, it implements the following method for controlling a home appliance: acquiring audio information acquired by an audio acquisition module; processing the audio information to obtain intent information; in response to the intent information indicating control of the home appliance, acquiring visual feature information acquired by a camera and wireless feature information acquired by a wireless communication module; wherein the visual feature information includes an RGB image of at least one home appliance, and the wireless feature information includes the signal strength of a radio signal transmitted by at least one home appliance received by a wearable device; determining at least one candidate home appliance based on the visual feature information and the wireless feature information; filtering the at least one candidate home appliance based on the intent information to determine at least one home appliance to be evaluated; evaluating the at least one home appliance to be evaluated based on the visual feature information and the wireless feature information to determine a matching target home appliance; and sending a home appliance control command to the target home appliance.

[0144] Optionally, in one embodiment, a first score for at least one household appliance to be evaluated is determined based on visual feature information; a second score for at least one household appliance to be evaluated is determined based on wireless feature information; a third score for at least one household appliance to be evaluated is determined based on the matching degree of visual feature information and wireless feature information; and a fourth score for at least one household appliance to be evaluated is determined based on detection confidence; wherein, the detection confidence is determined by a target detection model, which is used to perform target detection on RGB images; a comprehensive score for at least one household appliance to be evaluated is determined based on the first score, second score, third score, and fourth score; and a matching target household appliance is determined based on the comprehensive score.

[0145] This application also provides a computer program product or computer program, which includes program instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the steps described in the above method embodiments.

[0146] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0147] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0148] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0149] If the integrated units in the other embodiments described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0150] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for controlling a household appliance, characterized in that, Applied to wearable devices, the wearable devices including cameras and wireless communication modules, the control method includes: The device acquires visual feature information collected by the camera and wireless feature information collected by the wireless communication module; wherein the visual feature information includes an RGB image of at least one household appliance and the wireless feature information includes the signal strength of a radio signal transmitted by at least one household appliance received by the wearable device. Based on the visual feature information and the wireless feature information, a matching target home appliance is determined; Send appliance control commands to the target home appliance.

2. The control method for household appliances according to claim 1, characterized in that, The wearable device also includes an audio acquisition module, and the control method further includes: Obtain the audio information acquired by the audio acquisition module; The audio information is processed to obtain intent information; In response to the intent information indicating control of the home appliance, the steps of acquiring visual feature information collected by the camera and acquiring wireless feature information collected by the wireless communication module are performed.

3. The control method for household appliances according to claim 2, characterized in that, The process of processing the audio information to obtain intent information includes: The audio information is converted into audio features; The audio features are input into a pre-trained deep learning model to obtain the output candidate text; Information is extracted from the candidate text to obtain intent information.

4. The control method for household appliances according to claim 2, characterized in that, The step of determining the matching target home appliance based on the visual feature information and the wireless feature information includes: Based on the visual feature information and the wireless feature information, at least one candidate home appliance is identified; At least one candidate household appliance is filtered based on the intent information to identify at least one household appliance to be evaluated; At least one household appliance to be evaluated is evaluated based on the visual feature information and the wireless feature information to determine a matching target household appliance.

5. The control method for household appliances according to claim 1, characterized in that, The step of determining the matching target home appliance based on the visual feature information and the wireless feature information includes: A first score for at least one household appliance to be evaluated is determined based on the visual feature information; a second score for at least one household appliance to be evaluated is determined based on the wireless feature information; a third score for at least one household appliance to be evaluated is determined based on the matching degree between the visual feature information and the wireless feature information; and a fourth score for at least one household appliance to be evaluated is determined based on the detection confidence level; wherein the detection confidence level is determined by a target detection model, which is used to perform target detection on the RGB image; A comprehensive score for at least one household appliance to be evaluated is determined based on the first score, the second score, the third score, and the fourth score. The target home appliances that match the comprehensive score are determined.

6. The control method for household appliances according to claim 5, characterized in that, The step of determining a first score for at least one household appliance to be evaluated based on the visual feature information includes: The position score of the household appliance to be evaluated is determined based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image; and the depth score of the household appliance to be evaluated is determined based on the ratio of the pixel size of the household appliance to the standard size in the RGB image. Based on the location score and the depth score, a first score is determined for the household appliance to be evaluated.

7. The control method for household appliances according to claim 5, characterized in that, The visual feature information also includes a depth image of at least one household appliance, and determining a first score for at least one household appliance to be evaluated based on the visual feature information includes: The location score of the household appliance to be evaluated is determined based on the distance between the center point of the household appliance to be evaluated in the RGB image and the center point of the RGB image; and the depth score of the household appliance to be evaluated is determined based on the depth information of the household appliance to be evaluated in the depth image. Based on the location score and the depth score, a first score is determined for the household appliance to be evaluated.

8. The control method for household appliances according to claim 5, characterized in that, The step of determining a second score for at least one household appliance to be evaluated based on the wireless feature information includes: The signal strength of the radio signal transmitted by the household appliance to be evaluated, received by the wearable device, is obtained. The distance between the wearable device and the household appliance to be evaluated is determined based on the signal strength. A second score for the household appliance to be evaluated is determined based on the distance between the wearable device and the appliance.

9. The control method for household appliances according to claim 5, characterized in that, The step of determining a third rating for at least one home appliance to be evaluated based on the matching degree of the visual feature information and the wireless feature information includes: A first distance between the wearable device and the home appliance to be evaluated is determined based on the RGB image, and a second distance between the wearable device and the home appliance to be evaluated is determined based on the signal strength of the radio signal transmitted by the home appliance to be evaluated received by the wearable device. Based on the matching degree between the first distance and the second distance, a third rating is determined for at least one household appliance to be evaluated.

10. The control method for household appliances according to claim 5, characterized in that, The visual feature information also includes a depth image of at least one household appliance. The step of determining a third score for at least one household appliance to be evaluated based on the matching degree between the visual feature information and the wireless feature information includes: A third distance between the wearable device and the home appliance to be evaluated is determined based on the depth image, and a fourth distance between the wearable device and the home appliance to be evaluated is determined based on the signal strength of the radio signal transmitted by the home appliance to be evaluated received by the wearable device. Based on the matching degree between the third distance and the fourth distance, a third score is determined for at least one household appliance to be evaluated.

11. The control method for household appliances according to claim 5, characterized in that, A fourth rating is determined for at least one household appliance to be evaluated based on the detection confidence level, including: Obtain the initial detection confidence; The initial detection confidence level is adjusted based on the obstruction status of the household appliance to be evaluated to obtain the final detection confidence level; The fourth rating of the household appliance to be evaluated is determined based on the final test confidence level.

12. The control method for household appliances according to claim 5, characterized in that, The step of determining a comprehensive score for at least one household appliance to be evaluated based on the first score, the second score, the third score, and the fourth score includes: The weighting coefficients are determined based on whether the visual feature information contains a depth image and the current environmental conditions. The first score, the second score, the third score, and the fourth score are weighted using the weighting coefficients to determine a comprehensive score for at least one household appliance to be evaluated.

13. The control method for household appliances according to claim 5, characterized in that, After determining the matching target home appliance based on the comprehensive score, the process further includes: The detection confidence level of the target household appliance was verified; In response to the successful verification of the detection confidence level of the target home appliance, the step of sending the home appliance control command to the target home appliance is executed.

14. A computer device, characterized in that, The computer device includes a processor and a memory coupled to the processor; The memory is used to store a computer program, and the processor is used to execute the computer program to implement the control method for household appliances as described in any one of claims 1-13.

15. The computer device according to claim 14, characterized in that, The computer device is a wearable device.

16. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the control method for household appliances as described in any one of claims 1-13.

17. A computer program product, characterized in that, The computer program product includes computer instructions that are executed by a processor to implement the control method for a household appliance as described in any one of claims 1-13.

Images