Scene control method based on visual atmosphere selection

Abstract

The invention provides a scene control method based on visual atmosphere selection, and the method comprises the steps: obtaining a plurality of environment parameters, and calculating a matching function corresponding to each environment parameter; the matching functions comprise nonlinear mapping and linear mapping, and performing weighted summation on all the matching functions to obtain the matching degree. Weighted summation can comprehensively reflect contributions of different environment variables to an environment scene, and the obtained matching degree has high confidence. According to the method, the N scene images with the highest matching degree are screened out from the image database, the M scene images are randomly selected from the image database, and the M scene images are randomly selected, so that the adaptability to different scenes can be improved. And displaying the N + M scene images, and obtaining a user selection result. A user selects one scene image from N + M scene images, and subjective expectation of the user can be combined on the basis of conforming to objectivity of the environment. The corresponding control instruction set is executed according to the user selection result, the home atmosphere can be dynamically adjusted, and the personalized experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental control, and in particular to a scene control method based on visual atmosphere selection. Background Art

[0002] Existing smart home systems usually integrate environmental perception and intelligent control, but lack personalized image recommendation. Although personalized recommendation systems can analyze user behavior, they do not involve environmental perception and smart home control. Image recognition technology can identify image styles, but it is not combined with smart home systems.

[0003] User-defined smart home scenes are usually limited to preset control instruction sets, and cannot execute control instruction sets according to the environment and user subjective feelings. After adjusting the environment according to the preset control instruction set, the environmental parameters may not meet the user's expectations.

[0004] Therefore, there is a need for a scene control method that can combine environmental parameters and user's subjective expectations to dynamically adjust the home atmosphere and enhance the personalized experience. Summary of the invention

[0005] In order to overcome the problems existing in the related art, the purpose of the present invention is to provide a scene control method based on visual atmosphere selection, which can combine environmental parameters and user's subjective expectations to dynamically adjust the home atmosphere and enhance the personalized experience.

[0006] A scene control method based on visual atmosphere selection, comprising:

[0007] Acquire multiple environmental parameters, and calculate a matching function corresponding to each of the environmental parameters;

[0008] Performing weighted summation on all the matching functions to obtain a matching degree;

[0009] Filter out N scene images with the highest matching degree from the image database;

[0010] Randomly extract M images of the scene from the image database;

[0011] Display N+M images of the scene and obtain the user's selection result;

[0012] Execute a corresponding control instruction set according to the user selection result.

[0013] In a preferred technical solution of the present invention, the step of calculating the matching function corresponding to each of the environmental parameters includes:

[0014] Matching functions using trigonometric calculation times;

[0015] The matching functions of seasons and geographic locations are calculated using a mapping method;

[0016] A nonlinear function is used to calculate the matching functions of indoor temperature, outdoor temperature and humidity.

[0017] In a preferred technical solution of the present invention, the matching function using trigonometric function to calculate time includes:

[0018] The matching function of time is calculated using the following formula:

[0019] f t (t) = sin((pi / 12)*(t-6));

[0020] Among them, f t (t) is the matching function of time, sin is the sine function, pi is the circumference of a circle, and t is time.

[0021] In a preferred technical solution of the present invention, the matching function of calculating indoor temperature, outdoor temperature and humidity using a nonlinear function includes:

[0022] The matching function of indoor temperature, outdoor temperature and humidity is calculated using the following formula:

[0023] f Tin (Tin)=1 / (1+exp(-(Tin-20)));

[0024] Among them, f Tin (Tin) is the matching function of indoor temperature, Tin is the indoor temperature, and exp is the exponential function;

[0025] The matching function for outdoor temperature is calculated using the following formula:

[0026] f Tout (Tout)=1 / (1+exp(-(Tout-15)));

[0027] Among them, f Tout (Tout) is the matching function of the outdoor temperature, Tout is the outdoor temperature;

[0028] The matching function for humidity is calculated using the following formula:

[0029] f H (H) = 1 / (1 + exp(-(H-50)));

[0030] Among them, f H (H) is the matching function of humidity, and H is the humidity.

[0031] In a preferred technical solution of the present invention, the method of calculating the matching function of season and geographical location by using a mapping method includes:

[0032] If the season is spring, the value of the season matching function is 1; if the season is summer, the value of the season matching function is 0.8; if the season is autumn, the value of the season matching function is 0.6; if the season is winter, the value of the season matching function is 0.4;

[0033] If the geographical location is a city, the value of the matching function of the geographical location is 1; if the geographical location is a rural area, the value of the matching function of the geographical location is 0.8; if the geographical location is at the seaside, the value of the matching function of the geographical location is 0.6.

[0034] In a preferred technical solution of the present invention, performing weighted summation on all the matching functions to obtain the matching degree includes:

[0035] The matching degree is calculated using the following formula:

[0036] M i =w T *f T (T)+w S *f S (S)+w Tin *f Tin (Tin)+w Tout *f Tout (Tout)+w H *f H (H)+w L *f L (L);

[0037] Among them, M i is the matching degree of the i-th scene image, w T is the time weight, f T (T) is the matching function of time, w S is the seasonal weight, f S (S) is the seasonal matching function; w Tin is the indoor temperature weight, f Tin (Tin) is the matching function of the indoor temperature, w Tout is the outdoor temperature weight, f Tout (Tout) is the matching function of the outdoor temperature; w H Humidity weight, f H (H) is the matching function of humidity, w L is the geographical location weight, f L (L) is the matching function of the geographic location.

[0038] In a preferred technical solution of the present invention, before performing weighted summation on all the matching functions to obtain the matching degree, the method further includes:

[0039] Get the initial weight, and update the weight according to the initial weight. The formula for updating the weight is as follows:

[0040] w S' =w S +alpha*(Pw S );

[0041] Among them, w S' is the updated weight, w S is the initial weight, alpha is the learning rate, and P is the user preference strength.

[0042] In a preferred technical solution of the present invention, the step of randomly extracting M images of the scene from the image database comprises:

[0043] Extracting a style vector of each of the scene images in the image database;

[0044] Randomly select one of the scene images from the image database as a reference image;

[0045] Calculating a style difference measurement function between other scene images in the image database except the reference image and the reference image;

[0046] The scene images corresponding to the maximum M values ​​of the style difference metric function are extracted.

[0047] In a preferred technical solution of the present invention, executing a corresponding control instruction set according to the user selection result includes:

[0048] Acquire the target scene image selected by the user;

[0049] The environmental parameters are adjusted according to the target scene image.

[0050] In a preferred technical solution of the present invention, adjusting the environmental parameters according to the target scene image includes:

[0051] An environmental adjustment device is started according to the target scene image, or the environmental adjustment device is adjusted according to the target scene image; wherein the environmental adjustment device includes any one or any combination of an air conditioning system, a humidifier, a lighting system, an audio system, and security equipment.

[0052] The beneficial effects of the present invention are:

[0053] The present invention provides a scene control method based on visual atmosphere selection, including obtaining multiple environmental parameters and calculating the matching function corresponding to each environmental parameter. The matching function includes nonlinear mapping and linear mapping. The nonlinear mapping can better simulate the nonlinear change characteristics of environmental variables, and the linear mapping can reduce the amount of calculation. All matching functions are weighted and summed to obtain a matching degree. The weighted sum can comprehensively reflect the contribution of different environmental variables to the environmental scene, and the obtained matching degree has a high confidence level. N scene images with the highest matching degree are screened out from an image database, and M scene images are randomly extracted from the image database. The N scene images are obtained based on environmental parameter matching, and the M scene images are randomly selected with certain uncertainty, which can increase the adaptability to different scenes. N+M scene images are displayed to obtain user selection results. The user selects a scene image from the N+M scene images, which can be combined with the user's subjective expectations on the basis of conforming to the objectivity of the environment. The corresponding control instruction set is executed according to the user selection result, and the home atmosphere can be dynamically adjusted to enhance the personalized experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flow chart of a scene control method based on visual atmosphere selection of the present invention;

[0055] Figure 2 is a flow chart of calculating the matching function corresponding to each environmental parameter of the present invention;

[0056] Figure 3 is a flow chart of randomly extracting M scene images from an image database of the present invention;

[0057] Figure 4 It is a flow chart of executing a corresponding control instruction set according to a user selection result of the present invention. DETAILED DESCRIPTION

[0058] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0059] Example 1

[0060] like Figure 1 As shown, this embodiment provides a scene control method based on visual atmosphere selection, including:

[0061] S1: Acquire multiple environmental parameters and calculate a matching function corresponding to each of the environmental parameters.

[0062] S2: Perform weighted summation on all the matching functions to obtain a matching degree.

[0063] S3: Filter out N scene images with the highest matching degree from the image database.

[0064] S4: Randomly extract M images of the scene from the image database.

[0065] S5: Display N+M images of the scene and obtain the user selection result.

[0066] S6: Execute a corresponding control instruction set according to the user selection result.

[0067] Environmental parameters include time, season, indoor temperature, outdoor temperature, humidity and geographical location. There are two types of geographical location, and one of the geographical location parameters can be selected at will. The first geographical location parameter includes urban, rural and seaside, and the second geographical location parameter includes tropical, subtropical and cold zones. Seasons include spring, summer, autumn and winter, and different seasons correspond to different season weights. Environmental parameters are collected using an environmental parameter acquisition module, which includes a temperature sensor, a humidity sensor, a positioning device and a time correction device.

[0068] Different environmental parameters correspond to different matching functions, time corresponds to trigonometric functions, season and geographic location correspond to linear mapping, and indoor temperature, outdoor temperature and humidity correspond to nonlinear mapping.

[0069] After calculating the matching function of all environmental parameters, dynamic adjustments are made based on the preset weights to obtain updated weights. The matching function is weighted and summed according to the updated weights to obtain the matching degree.

[0070] Arrange the matching degrees in descending order to obtain a matching degree sequence. The first matching degree in the matching degree sequence has the largest value, and the last matching degree in the matching degree sequence has the smallest value. Filter out the N scene images corresponding to the 1st to Nth matching degrees in the matching degree sequence. In this embodiment, N=3 is taken as an example. All three scene images are calculated based on environmental parameters and are objective.

[0071] The image database stores multiple scene images of different styles, and each scene image is associated with a specific device control instruction set. The N scene images screened in step S3 are excluded from the image database, and M scene images are randomly selected from the remaining images in the image database. In this embodiment, M=2 is taken as an example, and the two randomly selected scene images have uncertainty, which can increase the adaptability to different scenes.

[0072] Five scene images are displayed on the interactive interface for the user to select, and the user selects one scene image from the five scene images in the interactive interface. After the system obtains the scene image selected by the user, it executes a control instruction set corresponding to the scene image, starts the environment adjustment device, or adjusts the operating parameters of the environment adjustment device to adjust the environment parameters.

[0073] Images are intuitive and easy to understand. Compared with text and voice, users are more willing to accept control settings based on scene images. Therefore, the personalized image recommendation function refers to generating several scene control instruction sets with high scene matching based on environmental parameters, and then converting these instructions into images and presenting them to users. Users select scene images. This interactive step of selecting scene images plays a role in personalized control. The above scene control method can be used in application scenarios such as home scenes, whole-house control, and device linkage.

[0074] The present embodiment provides a scene control method based on visual atmosphere selection, including obtaining multiple environmental parameters and calculating the matching function corresponding to each environmental parameter. The matching function includes nonlinear mapping and linear mapping. The nonlinear mapping can better simulate the nonlinear change characteristics of the environmental variables, and the linear mapping can reduce the amount of calculation. All matching functions are weighted and summed to obtain the matching degree. The weighted sum can comprehensively reflect the contribution of different environmental variables to the environmental scene, and the obtained matching degree has a high confidence level. N scene images with the highest matching degree are screened out from the image database, and M scene images are randomly selected from the image database. The N scene images are obtained based on the matching of environmental parameters, and the M scene images are randomly selected with a certain degree of uncertainty, which can increase the adaptability to different scenes. N+M scene images are displayed to obtain the user selection result. The user selects a scene image from the N+M scene images, which can be combined with the user's subjective expectations on the basis of conforming to the objectivity of the environment. The corresponding control instruction set is executed according to the user selection result, and the home atmosphere can be dynamically adjusted to enhance the personalized experience.

[0075] Example 2

[0076] like Figure 2 As shown, this embodiment provides a scene control method based on visual atmosphere selection, and this embodiment only describes the differences from Embodiment 1. The calculation of the matching function corresponding to each of the environmental parameters includes:

[0077] S12: Using trigonometric functions to calculate the matching function of time.

[0078] S13: The matching function between season and geographic location is calculated using the mapping method.

[0079] S14: Calculate the matching function of indoor temperature, outdoor temperature and humidity using a nonlinear function.

[0080] Before step S12, step S11 is also included: obtaining multiple environmental parameters. Using trigonometric functions such as sine functions to calculate the matching function of time can reflect the periodicity of time changes. Using linear mapping to calculate the matching function of seasons and geographical locations has the effect of simplifying operations under the premise of considering seasons and geographical locations. Using nonlinear functions such as Sigmoid functions to calculate indoor temperature, outdoor temperature and humidity can better simulate the nonlinear change characteristics of environmental parameters. The impact of environmental parameters on user comfort and preferences is nonlinear. For example, when the indoor temperature is close to 20°C, the user's comfort will be significantly improved, and the probability of selecting a preferred scene image will also increase rapidly. However, when the temperature is further increased or decreased, the degree of influence on the user's selection of scene images will gradually decrease. This means that in certain intervals, such as when the temperature is around 20°C, even a small temperature change will lead to a significant change in user experience, and when away from this interval, the impact of temperature changes on user preferences will be greatly reduced.

[0081] The matching function using trigonometric function to calculate time includes:

[0082] The matching function of time is calculated using the following formula:

[0083] f t (t) = sin((pi / 12)*(t-6));

[0084] Among them, f t (t) is the matching function of time, sin is the sine function, pi is the circumference of a circle, and t is time.

[0085] The method of using a nonlinear function to calculate the matching function of indoor temperature, outdoor temperature and humidity includes:

[0086] The matching function of indoor temperature, outdoor temperature and humidity is calculated using the following formula:

[0087] f Tin (Tin)=1 / (1+exp(-(Tin-20)));

[0088] Among them, f Tin (Tin) is the matching function of the indoor temperature, Tin is the indoor temperature, and exp is the exponential function;

[0089] The matching function for outdoor temperature is calculated using the following formula:

[0090] f Tout (Tout)=1 / (1+exp(-(Tout-15)));

[0091] Among them, fTout (Tout) is the matching function of the outdoor temperature, Tout is the outdoor temperature;

[0092] The matching function for humidity is calculated using the following formula:

[0093] f H (H) = 1 / (1 + exp(-(H-50)));

[0094] Among them, f H (H) is the matching function of humidity, and H is the humidity.

[0095] The indoor temperature matching function, outdoor temperature matching function, and humidity matching function all use the Sigmoid function, which can smoothly describe the impact of the current environmental variables on the image matching degree, avoiding sudden changes and discontinuities. The process of users feeling environmental changes is gradual, not instantaneous. The smoothness of the Sigmoid function ensures that the system is not too abrupt when adjusting environmental parameters, enhancing the consistency of the user experience.

[0096] By using the Sigmoid function, the system can better adapt to the actual needs of users. For example, the performance of users under different environmental conditions can be modeled through machine learning, and the parameters of the matching function can be continuously optimized using real-time collected data, so that the system can meet the personalized preferences of different users.

[0097] The Sigmoid function is sensitive to environmental variables, allowing the system to respond highly to changes in temperature and humidity. Near the critical point of comfort, the matching degree generated by the model can increase or decrease rapidly, reflecting the immediate impact of environmental variables on user emotions and choices. This sensitivity allows the control instruction set to be implemented quickly and accurately.

[0098] By taking the nonlinear response of environmental factors into account, the system can better meet the user's expectations when executing the control instruction set. For example, if the humidity exceeds the user's comfort range, the system can adjust the air conditioning equipment in time to improve the indoor environment, thereby achieving efficient and reasonable environmental management.

[0099] The method of using a mapping method to calculate the matching function of season and geographical location includes:

[0100] If the season is spring, the value of the season matching function is 1; if the season is summer, the value of the season matching function is 0.8; if the season is autumn, the value of the season matching function is 0.6; if the season is winter, the value of the season matching function is 0.4;

[0101] If the geographical location is a city, the value of the matching function of the geographical location is 1; if the geographical location is a rural area, the value of the matching function of the geographical location is 0.8; if the geographical location is at the seaside, the value of the matching function of the geographical location is 0.6.

[0102] The seasonal matching function and the geographic location matching function are calculated by linear mapping. Under the premise of adding the contribution of season and geographic location to the matching degree, the amount of calculation can be reduced and the calculation efficiency can be improved.

[0103] The time matching function of this embodiment adopts a sine function to reflect the periodic change of time. The indoor temperature matching function, the outdoor temperature matching function and the humidity matching function all adopt a Sigmoid function, and the shape of the Sigmoid function is S-shaped. Since the influence of environmental parameters on the user's comfort and preference is nonlinear, the use of the Sigmoid function can simulate the nonlinear change characteristics of environmental variables. The Sigmoid function can smoothly describe the influence of the current environmental variables on the image matching degree, avoiding mutations and discontinuities. The process of users feeling environmental changes is gradual, not instantaneous. This smoothness of the Sigmoid function ensures that the system is not too abrupt when adjusting environmental parameters, and enhances the consistency of user experience.

[0104] Example 3

[0105] This embodiment provides a scene control method based on visual atmosphere selection, and this embodiment only describes the differences from Embodiment 1. The weighted sum of all the matching functions to obtain the matching degree includes:

[0106] The matching degree is calculated using the following formula:

[0107] M i =w T *f T (T)+w S *f S (S)+w Tin *f Tin (Tin)+w Tout *f Tout (Tout)+w H *f H (H)+w L *f L (L);

[0108] Among them, M i is the matching degree of the i-th scene image, w T is the time weight, f T (T) is the matching function of time, w S is the seasonal weight, f S(S) is the seasonal matching function; w Tin is the indoor temperature weight, f Tin (Tin) is the matching function of the indoor temperature, w Tout is the outdoor temperature weight, f Tout (Tout) is the matching function of the outdoor temperature; w H Humidity weight, f H (H) is the matching function of humidity, w L is the geographical location weight, f L (L) is the matching function of the geographic location.

[0109] Time weight w T is 0.15, and the seasonal weight w S is 0.2. Indoor temperature weight w Tin is 0.15, and the outdoor temperature weight w Tout is 0.1, humidity weight w H is 0.2, and the geographical location weight w L is 0.1. Since seasonal changes have a greater impact on user emotions, the seasonal weight w S Set to a larger value. Since the indoor temperature has a moderate impact on the user's choice, the indoor temperature weight w Tin Set it to a moderate value. Since the outdoor temperature has little influence on the user's choice, the outdoor temperature weight w Tout Set it to a smaller value. Since humidity has a significant impact on indoor comfort, the humidity weight w H Set to a larger value. Since the geographical location has little influence on image selection, the geographical location weight w L Set to a smaller value. The matching degree of the scene image combines the influence of time, season, indoor temperature, outdoor temperature, humidity and geographical location, and can more accurately reflect the characteristics of the current environment.

[0110] The weights corresponding to the above environmental parameters are all initial weights. When the requirements for environmental control are high, the initial weights of different environmental parameters need to be updated. Before performing weighted summation on all the matching functions to obtain the matching degree, the following steps are also included:

[0111] Get the initial weight, and update the weight according to the initial weight. The formula for updating the weight is as follows:

[0112] w S' =w S +alpha*(Pw S );

[0113] Among them, w S' is the updated weight, w Sis the initial weight, alpha is the learning rate, and P is the user preference strength.

[0114] The above formula can be used to dynamically adjust the weights based on the user's behavior data. For example, if the data shows that users frequently select certain images in a specific season, the weight of that season will be increased. Alpha is the learning rate, which is used to control the adjustment range of the environmental parameters. The larger the learning rate, the larger the adjustment range of the environmental parameters.

[0115] For different environmental parameters, the user preference strength P can be a fixed value, that is, different environmental parameters correspond to the same user preference strength P; or it can be a value related to the environmental parameters, that is, different environmental parameters correspond to different user preference strengths P.

[0116] It is reasonable to set different weights to reflect the preference differences between different places of residence and geographical locations, especially in the personalized recommendation and decision-making of the system. The weights of cities and tropical areas are set larger mainly based on the following physical meanings:

[0117] (1) Diversity and complexity of the urban environment

[0118] Multiple stimuli: Cities are usually full of diverse visual and sensory stimuli, including architectural styles, colors, and lighting, so users may have a stronger preference for modern and fashionable images in urban environments. The pace of life and cultural atmosphere in cities are more diverse, and the interaction between users and the environment is richer.

[0119] Modern Lifestyle: City dwellers are often more open to new social trends and technological advancements, so this modern feel resonates more with them in terms of image recommendations and style choices.

[0120] (2) Natural beauty of tropical areas

[0121] Richness of natural landscapes: Tropical regions have beautiful natural environments, lush vegetation, and rich colors and forms. This ecological environment is usually easier to attract users' attention, so users in such regions may prefer images showing natural scenery and tropical colors. Tropical regions give people a sense of relaxation and pleasure, which may lead to users' higher preference for images that match this environment.

[0122] (3) Style Preference

[0123] Geographic type and lifestyle: The difference between urban life and rural, seaside or cold lifestyles makes users have significantly different preferences for images in different environments. Urban and tropical residents may prefer active and vivid images, and therefore give them higher weight.

[0124] (4) Environmental adaptation and comfort

[0125] Climate adaptation: Users in urban and tropical areas may be more familiar with and adapt to local climate characteristics, such as high humidity and high temperature. Such adaptability makes users more likely to choose images that match their climate conditions.

[0126] Living conditions: Urban living conditions may require users to be flexible and adaptable to different environmental factors, while residents in tropical areas have more access to sunlight and nature. Such living conditions affect their preferences in image selection.

[0127] This embodiment uses a weighted summation method to calculate the matching degree. For environmental parameters with greater influence, such as season and humidity, the season weight and humidity weight are set to larger values. For environmental parameters with moderate influence, such as indoor temperature, the indoor temperature weight is set to a moderate value. For environmental parameters with smaller influence, such as outdoor temperature and geographical location, the outdoor temperature weight and geographical location weight are set to smaller values. The matching degree of the scene image combines the influence of time, season, indoor temperature, outdoor temperature, humidity and geographical location, and can more accurately reflect the characteristics of the current environment. The weight is dynamically adjusted according to the user's behavior data. For example, if the data shows that the user frequently selects certain images in a specific season, the weight of the season is increased. Alpha is the learning rate. The learning rate is used to control the adjustment range of the environmental parameters. The larger the learning rate, the larger the adjustment range of the environmental parameters.

[0128] Example 4

[0129] like Figure 3 As shown, this embodiment provides a scene control method based on visual atmosphere selection, and this embodiment only describes the differences from Embodiment 1. The randomly extracting M images of the scene from the image database includes:

[0130] S41: Extracting the style vector of each scene image in the image database.

[0131] S42: Randomly select one of the scene images from the image database as a reference image.

[0132] S43: Calculate a style difference measurement function between the other scene images in the image database except the reference image and the reference image.

[0133] S44: Extracting the scene images corresponding to the maximum M values ​​of the style difference metric function.

[0134] The image database is a user-defined database, which allows users to upload customized images and assign one or more style tags to each scene image, such as style tags of modern, natural and / or warm. The user-defined scene images are included in the image library and marked as user-defined.

[0135] The style features of each scene image in the image database are extracted. The style features include color histogram, texture analysis and possible object recognition. All style features are stored as the style vector of the corresponding scene image.

[0136] Define a style difference metric function D(vi,vj) to calculate the style difference between two scene images. The style difference metric function is based on the Euclidean distance between style vectors. The formula of the style difference metric function is as follows:

[0137]

[0138] Among them, D(v i ,v j ) is the style difference measurement function between the i-th scene image and the j-th scene image, v i is the style vector of the i-th scene image, v j is the style vector of the jth scene image. i,k is the kth dimension of the i-th scene image, v j,k is the kth dimension of the jth scene image, and n is the total number of dimensions of the style vector.

[0139] The N scene images selected in step S3 are excluded from the image database. In this embodiment, N=3 is used as an example. A reference image I is randomly selected from the remaining scene images in the image database. r As a comparison benchmark, calculate each candidate image I in the image database c With reference image I r The style difference measurement function D(v Ic ,v Ir ).

[0140] A random selection model is used to randomly select images. The selection probability is calculated as follows:

[0141]

[0142] Among them, P(I c ) is the candidate image I c The corresponding selection probability, v Ic is the candidate image I c The style vector, v Ir is the reference image I r The style vector D(vIc ,v Ir ) is the candidate image I c and reference image I r The style difference metric function between Iq is the qth scene image in the remaining scene image set after the N scene images selected in step S3 are excluded from the image database, D(v Ic ,v Iq ) is the reference image I r and the style difference measurement function between the qth scene image in the remaining scene image set, where Q is the total number of scene images in the remaining scene image set.

[0143] P(I c ) is larger, indicating that the candidate image I c and reference image I r The greater the difference in style between them. c ) are arranged in descending order, and M scene images corresponding to the first M values ​​are extracted, and the style difference between the selected M scene images and the reference image Ir is the largest. Preferably, if there is a user-defined image, the user-defined image is preferentially selected.

[0144] First, a scene image is randomly selected from the image database as a reference image to ensure the randomness of the reference image. Then, M scene images with the largest style difference from the reference image are screened out from the remaining scene image set, so that the M scene images are also random, thereby making the recommended images diverse, which can meet the preferences of different users and enhance the satisfaction of different users.

[0145] The dimensions of the style vector include color frequency, texture analysis results, object recognition results, edge detection results, color contrast, brightness and saturation.

[0146] (1) Color frequency: The color histogram is used to describe the distribution of colors in an image. The frequency of each color in the scene image is extracted based on the color histogram, and the frequency of one of the colors is used as a dimension of the vector.

[0147] (2) Texture analysis: Describe the texture characteristics of an image by analyzing the texture patterns in the image, such as roughness, directionality, etc.

[0148] (3) Describe the content of scene images by identifying specific objects or patterns in the image, such as faces and buildings.

[0149] (4) Describe the structural information of the image by detecting the edges in the image.

[0150] (5) Color contrast: describes the intensity of color contrast in an image.

[0151] (6) Brightness and Saturation: Describes the overall brightness and color saturation of the image.

[0152] As an example, suppose there is an image of a scene and the following features are extracted:

[0153] Color histogram: red (10%), green (20%), blue (30%), yellow (20%), black (20%).

[0154] Texture analysis: roughness (0.7), directionality (0.3).

[0155] Object recognition: If a face is detected, the value is 1.

[0156] Edge detection: The edge strength is 0.5.

[0157] Color contrast: The color contrast is high and the value is 1.

[0158] Brightness and saturation: Brightness is 0.8 and Saturation is 0.6.

[0159] Then, the style vector of the scene image is in the form of: [red, green, blue, yellow, black, roughness, directionality, face, building, edge strength, high, brightness, saturation], and the style vector of the scene image is expressed as: [0.1, 0.2, 0.3, 0.2, 0.2, 0.7, 0.3, 1, 0, 0.5, 1, 0.8, 0.6].

[0160] The method of randomly extracting M scene images from the image database in this embodiment includes extracting the style vector of each scene image in the image database, and randomly selecting one scene image from the image database as a reference image. The style difference measurement function between the other scene images in the image database except the reference image and the reference images is calculated, and the scene images corresponding to the largest M values ​​of the style difference measurement function are extracted. First, a scene image is randomly selected from the image database as a reference image to ensure the randomness of the reference image, and then the M scene images with the largest style difference from the reference image are screened out from the remaining scene images, so that the M scene images are also random, so that the recommended images are diverse, which can meet the preferences of different users and enhance the satisfaction of different users.

[0161] Example 5

[0162] like Figure 4 As shown, this embodiment provides a scene control method based on visual atmosphere selection, and this embodiment only describes the differences from Embodiment 1. The corresponding control instruction set is executed according to the user selection result, including:

[0163] S61: Acquire the target scene image selected by the user.

[0164] S62: Adjust the environmental parameters according to the target scene image.

[0165] The step of adjusting the environmental parameters according to the target scene image includes:

[0166] An environmental adjustment device is started according to the target scene image, or the environmental adjustment device is adjusted according to the target scene image; wherein the environmental adjustment device includes any one or any combination of an air conditioning system, a humidifier, a lighting system, an audio system, and security equipment.

[0167] After N+M scene images are displayed in the interactive interface of the display screen, the user selects a scene image from the displayed N+M scene images. The system executes the corresponding control instruction set according to the scene image selected by the user. If the user is not satisfied, he can click the refresh button in the interactive interface, and the system will reselect N+M scene images and display the updated N+M scene images in the interactive interface.

[0168] The system executes the corresponding control instruction set and adjusts the scene devices according to the target scene image selected by the user. Scene device control is one of the core functions of the intelligent environment system, which usually covers a variety of devices to improve user experience and comfort.

[0169] (1) Main types of scene devices: air conditioning systems such as air conditioners and thermostats, humidifiers or dehumidifiers, lighting systems such as smart lamps and curtain controllers, audio systems, smart home appliances such as refrigerators and coffee machines, security equipment such as surveillance cameras and door locks, and air purifiers.

[0170] (2) Controlling the scene device according to the image selected by the user and its corresponding control instruction set. For example, if the scene image selected by the user represents a warm and sunny scene, the control instruction is to set the temperature of the air conditioner to 27°C to keep the room warm.

[0171] (3) The control instruction set includes multiple parameters. For example, the control instruction set of an air conditioner is {air conditioner switch: on, temperature setting: 27°C, wind speed: medium, mode: cooling}.

[0172] The entire control process can be designed so that the system automatically identifies the user's actual environmental needs and dynamically adjusts the operating parameters of the air conditioner and humidifier according to the target scene image selected by the user. For example, if the user selects a scene image with warm sunshine, the system turns on the air conditioner and sets it to warm air. The humidifier will start when the humidity is low to ensure that the humidity of the environment remains at the level preferred by the user.

[0173] This embodiment executes the corresponding control instruction set according to the user selection result, including obtaining the target scene image selected by the user and adjusting the environmental parameters according to the target scene image. The adjustment of the environmental parameters according to the target scene image includes starting the environmental adjustment device according to the target scene image, or adjusting the environmental adjustment device according to the target scene image; wherein the environmental adjustment device includes any one or any combination of an air conditioning system, a humidifier, a lighting system, an audio system, and a security device. The entire control process can be designed so that the system automatically identifies the actual environmental needs of the user and dynamically adjusts the operating parameters of the air conditioner and humidifier according to the target scene image selected by the user. For example, if the user selects a scene image with warm sunshine, the system turns on the air conditioner and sets it to the warm wind state. The humidifier will start when the humidity is low to ensure that the humidity of the environment remains at the level preferred by the user.

[0174] Unless otherwise specifically stated, the relative arrangement, numerical expressions and numerical values ​​of the parts and steps set forth in these embodiments do not limit the scope of the application. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a restriction. Therefore, other examples of exemplary embodiments can have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, and therefore, once a certain item is defined in an accompanying drawing, it does not need to be further discussed in the subsequent accompanying drawings.

[0175] It should be understood that spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figures is inverted, a device described as "above" or "above" other devices or structures would then be positioned "below" or "below" the other devices or structures. Thus, the exemplary term "above" can include both the orientations of "above" and "below." The device may also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used herein should be interpreted accordingly.

[0176] In addition, it should be noted that the use of terms such as "first" and "second" to limit components is only for the convenience of distinguishing the corresponding components. Unless otherwise stated, the above terms have no special meaning and therefore cannot be understood as limiting the scope of protection of this application.

[0177] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A scene control method based on visual atmosphere selection, characterized in that: include: Acquire multiple environmental parameters, and calculate a matching function corresponding to each of the environmental parameters; Performing weighted summation on all the matching functions to obtain a matching degree; Filter out N scene images with the highest matching degree from the image database; Randomly extract M images of the scene from the image database; Display N+M images of the scene and obtain the user's selection result; Execute a corresponding control instruction set according to the user selection result.

2. The scene control method based on visual atmosphere selection according to claim 1, characterized in that: The calculating the matching function corresponding to each of the environmental parameters comprises: Matching functions using trigonometric calculations of time; The matching functions of seasons and geographic locations are calculated using a mapping method; A nonlinear function is used to calculate the matching functions of indoor temperature, outdoor temperature and humidity.

3. The scene control method based on visual atmosphere selection according to claim 2, characterized in that: The matching function using trigonometric function to calculate time includes: The matching function of time is calculated using the following formula: f t (t)=sin((pi / 12)*(t-6)); Among them, f t (t) is the matching function of time, sin is the sine function, pi is the circumference of a circle, and t is time.

4. The scene control method based on visual atmosphere selection according to claim 2, characterized in that: The method of using a nonlinear function to calculate the matching function of indoor temperature, outdoor temperature and humidity includes: The matching function of indoor temperature, outdoor temperature and humidity is calculated using the following formula: f Tin (Tin)=1 / (1+exp(-(Tin-20))); Among them, f Tin (Tin) is the matching function of the indoor temperature, Tin is the indoor temperature, and exp is the exponential function; The matching function for outdoor temperature is calculated using the following formula: f Tout (All)=1 / (1+exp(-(All-15))); Among them, f Tout (Tout) is the matching function of the outdoor temperature, Tout is the outdoor temperature; The matching function for humidity is calculated using the following formula: f H (H)=1 / (1+exp(-(H-50))); Among them, f H (H) is the matching function of humidity, and H is the humidity.

5. The scene control method based on visual atmosphere selection according to claim 2, characterized in that: The method of using a mapping method to calculate the matching function of season and geographical location includes: If the season is spring, the value of the season matching function is 1; if the season is summer, the value of the season matching function is 0.8; if the season is autumn, the value of the season matching function is 0.6; if the season is winter, the value of the season matching function is 0.4; If the geographical location is a city, the value of the matching function of the geographical location is 1; if the geographical location is a rural area, the value of the matching function of the geographical location is 0.8; if the geographical location is at the seaside, the value of the matching function of the geographical location is 0.

6.

6. The scene control method based on visual atmosphere selection according to claim 2, characterized in that: The weighted sum of all the matching functions to obtain the matching degree includes: The matching degree is calculated using the following formula: M i =w T *f T (T)+w S *f S (S)+w Tin *f Tin (Tin)+w Tout *f Tout (Tout)+w H *f H (H)+w L *f L (L); Among them, M i is the matching degree of the i-th scene image, w T is the time weight, f T (T) is the matching function of time, w S is the seasonal weight, f S (S) is the seasonal matching function; w Tin is the indoor temperature weight, f Tin (Tin) is the matching function of the indoor temperature, w Tout is the outdoor temperature weight, f Tout (Tout) is the matching function of the outdoor temperature; w H Humidity weight, f H (H) is the matching function of humidity, w L is the geographical location weight, f L (L) is the matching function of the geographic location.

7. The scene control method based on visual atmosphere selection according to claim 1, characterized in that: Before performing weighted summation on all the matching functions to obtain the matching degree, the method further includes: Get the initial weight, and update the weight according to the initial weight. The formula for updating the weight is as follows: w S' =w S +alpha*(P-w S ); Among them, w S' is the updated weight, w S is the initial weight, alpha is the learning rate, and P is the user preference strength.

8. The scene control method based on visual atmosphere selection according to claim 1, characterized in that: The randomly extracting M images of the scene from the image database comprises: Extracting a style vector of each of the scene images in the image database; Randomly select one of the scene images from the image database as a reference image; Calculating a style difference measurement function between other scene images in the image database except the reference image and the reference image; The scene images corresponding to the maximum M values ​​of the style difference metric function are extracted.

9. The scene control method based on visual atmosphere selection according to claim 1, characterized in that: The executing a corresponding control instruction set according to the user selection result includes: Acquire the target scene image selected by the user; The environmental parameters are adjusted according to the target scene image.

10. The scene control method based on visual atmosphere selection according to claim 9, characterized in that: The step of adjusting the environmental parameters according to the target scene image includes: An environmental adjustment device is started according to the target scene image, or the environmental adjustment device is adjusted according to the target scene image; wherein the environmental adjustment device includes any one or any combination of an air conditioning system, a humidifier, a lighting system, an audio system, and security equipment.

Images