A method, system and apparatus for light adjustment

Abstract

The embodiment of the specification provides a light adjustment method, system and device, the method comprises: acquiring environment characteristics of a current space; determining a current light comfort degree of a user based on the environment characteristics; determining a light requirement based on the current light comfort degree; the light requirement is adjusted based on a user fatigue degree, and the user fatigue degree is determined based on activity data of the user and a duration of the user in the current space; and determining a light adjustment parameter based on a current light effect, the light requirement and a current light parameter.

Description

[0001] Divisional Statement

[0002] This application is a divisional application of the Chinese application with the application number 202310842302.2 and the application date of July 11, 2023, and the title of “An intelligent LED lamp, control system and method”. TECHNICAL FIELD

[0003] The present specification relates to the field of intelligent LED lamps, in particular to a light adjustment method, system and device. BACKGROUND

[0004] With the rapid development of intelligent technology, Internet technology and communication technology, the interconnection of various electrical equipment has become possible, and the intelligent demand of lighting appliance equipment has gradually emerged. At present, most of the functions of intelligent lighting products are single, and the automation level is limited.

[0005] In order to improve the automation level of LED lamps, CN111315071B proposes an LED intelligent control method and system, the LED control interface of the terminal includes selectable control modes, the control modes include: normal mode and special mode; each LED lamp is used to emit a basic color light; the server receives the LED control information sent by the terminal, obtains a plurality of basic color lights corresponding to the LED identifier, calculates the lumens value corresponding to each LED lamp, so as to obtain the LED target color light and the LED target lumens value by mixing the LED lighting lamp. The mode of adjusting the light of the LED intelligent control method and system depends on the preset table, the adjusting method is not flexible enough, and it cannot actively judge the demand of the user for the light and generate adjustment parameters accordingly.

[0006] Therefore, it is desirable to provide an intelligent LED lamp, control system and method with a more flexible control method and more adaptability to different users and their use scenarios. SUMMARY

[0007] One or more embodiments of the present specification provide a light adjustment method comprising: obtaining an environmental feature of a current space; determining a current light comfort level of a user based on the environmental feature; determining a light demand based on the current light comfort level; the light demand is adjusted based on the fatigue level of the user, and the fatigue level of the user is determined based on the activity data of the user and the duration of the user in the current space; determining a light adjustment parameter based on the current light effect, the light demand and the current light parameter.

[0008] The one or more embodiments of the specification provide a light adjustment system, which comprises a parameter determination module configured to: acquire environmental features of a current space; determine a current light comfort level of a user based on the environmental features; determine a light requirement based on the current light comfort level; the light requirement is adjusted based on a user fatigue level, the user fatigue level is determined based on activity data of the user and a time length of the user in the current space; determine a light adjustment parameter based on a current light effect, the light requirement and a current light parameter, wherein the current light effect comprises at least one of luminous flux, illuminance, light intensity, brightness, color temperature and color rendering.

[0009] The one or more embodiments of the specification provide a light adjustment device, which comprises at least one memory and at least one processor, the at least one memory is used to store computer instructions, and the at least one processor executes the computer instructions or part of the instructions to realize a light adjustment method. BRIEF DESCRIPTION OF DRAWINGS

[0010] The specification will be further illustrated in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same numbers represent the same structures, wherein:

[0011] Figure 1 is a system module schematic diagram of an intelligent LED lamp control system according to some embodiments of the specification;

[0012] Figure 2 is an exemplary flowchart of an intelligent LED lamp control method according to some embodiments of the specification;

[0013] Figure 3 is an exemplary schematic diagram of a scene analysis model according to some embodiments of the specification;

[0014] Figure 4 is an exemplary schematic diagram of a comfort level model according to some embodiments of the specification. DETAILED DESCRIPTION

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the specification, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some examples or embodiments of the specification, and for those skilled in the art, the specification can be applied to other similar scenarios without creating any creative labor. Unless it is obvious from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structures or operations.

[0016] It should be understood that the use of “system,” “apparatus,” “unit,” and / or “module” herein is merely used to differentiate different components, elements, parts, sections, or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0017] As indicated in the specification and claims herein, unless the context clearly indicates otherwise, the words “a,” “an,” “the,” and / or “this” do not necessarily refer to the singular but can include the plural, unless the context clearly indicates otherwise. Generally, the terms “comprising” and “including” merely indicate the inclusion of the elements explicitly identified, and these steps and elements do not constitute an exclusive listing of the steps or elements, and the method or device can include other steps or elements.

[0018] Flowcharts are used in the specification to illustrate the operations performed by the system according to the embodiments of the specification. It should be understood that the preceding or subsequent operations are not necessarily performed in sequence. Instead, the steps can be processed in reverse order or simultaneously. Meanwhile, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0019] The demand for intelligentization of lighting fixture devices has gradually emerged, but most of the current intelligent lighting products have relatively single functions and limited automation levels. In order to improve the automation level of LED lamps, CN111315071B proposes an LED intelligent control method and system, the LED control interface of the terminal includes selectable control modes, the control modes include: a normal mode and a special mode; each LED lamp is used to emit a basic color light; the server receives the LED control information sent by the terminal, obtains a plurality of basic color lights corresponding to the LED identifier, calculates the lumens value corresponding to each LED lamp to make the LED lighting lamp mix to obtain the LED target color light and the LED target lumens value. The mode of adjusting the light of the LED intelligent control method and system depends on the preset table, and the adjusting method is not flexible enough, and cannot actively judge the demand of the user for the light and generate adjustment parameters accordingly. In view of this, in some embodiments of the specification, it is desirable to provide an intelligent LED lamp, a control system and a method which are more flexible in control method and can better adapt to different users and their use scenarios.

[0020] In some embodiments, the smart LED luminaire can be controlled to meet the lighting needs of a user by implementing the smart LED luminaire control system and / or method disclosed in this specification. For example, in a typical application scenario, it can include a current space, a smart LED luminaire, a human sensor, a processor or a user terminal, a network. The smart LED luminaire control method can be executed by the processor of the smart LED luminaire itself or the processor of the user terminal, identifying the position information of at least one user in the current space through the human sensor; determining at least one user current scene based on at least one of the position information, the current space information and the current time information, the user current scene including the actual use type and / or the light purpose of the current space; determining the light adjustment parameter based on the current light effect and the at least one user current scene. The smart LED luminaire can obtain the light adjustment parameter from the processor executing the smart LED luminaire control method through the network, and provide lighting for the user based on the light adjustment parameter.

[0021] Figure 1 is a system module schematic diagram of the smart LED luminaire control system according to some embodiments of the present specification. As shown in Figure 1 , the smart LED luminaire control system 100 can include an identification module 110, a scene determination module 120 and a parameter determination module 130.

[0022] In some embodiments, the identification module 110 can be used to identify the position information of at least one user in the current space through the human sensor.

[0023] In some embodiments, the scene determination module 120 can be used to determine at least one user current scene based on at least one of the at least one position information, the current space information and the current time information, and the user current scene can include the actual use type and / or the light purpose of the current space.

[0024] In some embodiments, the scene determination module 120 can be further used to: determine the activity data of at least one user based on the current space information and the position information of at least one user at at least one time point; determine at least one user current scene according to the activity data and the current time information.

[0025] For more information about determining the user current scene, see Figure 3 and the related description.

[0026] In some embodiments, the parameter determination module 130 can be used to determine the light adjustment parameter based on the current light effect and the at least one user current scene.

[0027] In some embodiments, the parameter determining module 130 can be further configured to determine the light demand based on the current scene of the at least one user; and determine the light adjustment parameter based on the current light effect, the light demand and the current light parameter.

[0028] For more details about determining the light adjustment parameter, see Step 1, Step 2 and the related descriptions thereof below.

[0029] In some embodiments, the parameter determining module 130 can be further configured to obtain the environmental feature of the current space; determine the current light comfort level of the at least one user based on the environmental feature; and determine the light demand based on the current light comfort level of the at least one user.

[0030] For more details about determining the light demand, see Figure 4 and the related descriptions thereof.

[0031] For more details about the human body sensor, the position information, the current scene of the user, etc., see the related descriptions of other figures (e.g. Figure 2 , etc.).

[0032] It should be noted that the above description of the intelligent LED lamp control system 100 and its modules is for the convenience of description, and cannot limit the scope of the embodiments. It can be understood that, for those skilled in the art, after understanding the principle of the system, any combination of the modules or connection of the modules to form a subsystem can be made without departing from the principle. Such variations are within the scope of the present description.

[0033] Figure 2 is an exemplary flowchart of the intelligent LED lamp control method according to some embodiments of the present description. The intelligent LED lamp control method can be executed by the processor of the intelligent LED lamp itself, or by the processor of other devices or apparatuses such as the intelligent user terminal, and the light adjustment parameter is sent to the intelligent LED lamp after the execution. As shown in Figure 2 , the flowchart 200 can include the following steps.

[0034] Step 210, the position information of at least one user in the current space can be recognized by the human body sensor.

[0035] The human body sensor can be used to sense the movement of the human body. The human body sensor can include an infrared sensor, for example, a pyroelectric infrared sensor, etc.

[0036] In some embodiments, the human body sensor can be arranged on the smart LED lamp (for example, at any position on the lamp that does not affect the light and detection range), or can be arranged at other positions in the current space where the smart LED lamp is used. In some embodiments, according to the situation in the space, the human body sensor can be installed at equal or close intervals in the room. In some embodiments, after the human body sensor is installed, the human body sensor can sense the movement of the user in the current space based on the fixed installation position.

[0037] The current space can be a space where the user needs to use the smart LED lamp. For example, the current space can be in a room, a corridor, or other spaces.

[0038] The current space can be classified by use, for example, the types can include bedroom, study, living room, office, etc.

[0039] The user can be a user who uses, debugs or functions of the smart LED lamp. In some embodiments, each person in the current space can belong to the user.

[0040] The position information can include the coordinates of the user or the relative position to the reference point, etc. The position information can include horizontal plane and / or height direction position information.

[0041] In some embodiments, the processor can detect the heat change caused by the movement of the user in the current space through the human body sensor, identify the angle information (for example, the angle relative to the human body sensor) of the user in the current space, convert the angle information into an electrical signal, and identify the position information of the human body through an analysis algorithm. The analysis algorithm can include a position estimation algorithm, a motion tracking algorithm, etc. For example, in the position estimation algorithm, the position information of the human body can be identified based on the angle information, the layout of the human body sensor and the geometric relationship.

[0042] Step 220, determining at least one user current scene based on at least one of the position information, the current space information and the current time information.

[0043] The current space information can include the type, size, internal object arrangement of the current space, etc. The current space information can be obtained by pre-setting, such as marking the type in advance, or uploading the image of the current space, pre-setting the size and internal object arrangement of the space, etc.

[0044] The current time can be the real-time time of the system. The current time information can include year, month, day, hour, minute, etc.

[0045] The user current scene can be a certain scene when the user uses the smart LED lamp, for example, office scene, learning scene, rest scene, etc.

[0046] In some embodiments, the user current scene can include an actual use type and / or a light use of the current space. For example, if the type of the current space is a study, the user actually uses it for office work, and the actual use type can be an office, etc.

[0047] The light use can be a specific subdivision of the type of the current space or the actual use type of the current space. For example, it can include single-person independent office work, office work for face-to-face service to customers, lunch break, etc.

[0048] In some embodiments, the processor can determine the at least one user current scene by querying a preset scene corresponding table based on at least one of the location information, the current space information, and the current time information. The preset scene corresponding table can include a corresponding relationship between various different location information or various different time information and various scenes.

[0049] In some embodiments, the processor can identify the location information of the user and the corresponding current time information at each time of the plurality of times, send the plurality of candidate scenes to the user for confirmation, obtain the user current scene confirmed by the user, and obtain the preset scene corresponding table.

[0050] For more details of determining the at least one user current scene based on other manners, see Figure 3 and the related descriptions.

[0051] Step 230, determining a light adjustment parameter based on the current light effect and the at least one user current scene.

[0052] The current light effect can be an effect actually produced by the light in the current space. For example, the current light effect can include at least one of luminous flux, illuminance, light intensity, brightness, color temperature, color rendering, etc. In some embodiments, the current light effect can be obtained by a light sensor arranged in the current space or on the smart LED lamp.

[0053] The light adjustment parameter can be a parameter for adjusting the light effect. For example, the light adjustment parameter can include brightness, color temperature, color, switch, etc. In some embodiments, the smart LED lamp can cause the current light effect to change based on the light adjustment parameter.

[0054] In some embodiments, the processor can determine an ideal light effect according to a scene effect parameter correspondence table based on the user current scene, and determine the light adjustment parameter based on a difference between the current light effect and the ideal light effect. The scene effect parameter correspondence table can include a corresponding relationship between various scenes and / or various different numbers of users and corresponding ideal light effects, and the ideal light effect includes corresponding ideal luminous flux, ideal illuminance, ideal light intensity, etc.

[0055] In some embodiments, the processor can determine the light demand based on the current scene of the at least one user; determine the light adjustment parameter based on the current light effect, the light demand and the current light parameter, for more details, see step one, step two and the related description below.

[0056] In some embodiments of the present specification, the position information of the at least one user is identified by the human body sensor; the current scene of the at least one user is determined based on at least one of the position information, the current space information and the current time information; the light adjustment parameter is determined based on the current light effect and the current scene of the at least one user. This can realize the pre-judgment of the current scene of the user, and can combine the current scene of the user with the current light effect to actively adjust the current light automatically, and the adjustment is more reasonable and more in line with the light demand of the user in different scenes.

[0057] It should be noted that the above description of the flow 200 of the intelligent LED lamp control method is only for example and illustration, and does not limit the scope of the present specification. Those skilled in the art can make various modifications and changes to the flow 200 under the guidance of the present specification. However, these modifications and changes are still within the scope of the present specification.

[0058] Figure 3 is an exemplary schematic diagram of a scene analysis model according to some embodiments of the present specification.

[0059] In some embodiments, the processor can determine the activity data of the at least one user based on the current space information and the position information of the at least one user at at least one time point; determine the current scene of the at least one user according to the activity data and the current time information.

[0060] The activity data can be data related to the activity of the user in the current space. For example, it can include activity frequency, activity trajectory, trajectory point, trajectory point stay duration, etc.

[0061] The activity data can also include current time information. For example, if the user has corresponding activity data in the current space at three time points of 9:00 in the morning, 12:00 in the afternoon and 5:00 in the afternoon, the activity data corresponds to the first activity data, the second activity data and the third activity data respectively.

[0062] The plurality of activity data belonging to the same user and arranged in time sequence belongs to the activity data sequence corresponding to the user.

[0063] In some embodiments, the processor can determine the activity frequency by counting the number of times the user appears in the current space or the number of times the user's location changes, based on the user's location information at at least one time point; take the location points corresponding to the location information as trajectory points based on the user's location information at at least one time point; connect the trajectory points corresponding to consecutive time points to form an activity trajectory; and determine the duration of the user's stay at a trajectory point by counting the start and end times of the user's stay at a certain trajectory point, based on the user's location information at at least one time point.

[0064] In some embodiments, the processor can determine at least one user's current scenario based on activity data and current time information, according to a preset scenario correspondence table. The preset scenario correspondence table may include various activity trajectories, activity frequencies, and / or durations of stay at trajectory points, as well as the corresponding scenario relationships. The preset scenario correspondence table can be set empirically; for example, the scenario corresponding to an activity trajectory that frequently moves around a table can be preset as an office scenario.

[0065] In some embodiments, such as Figure 3 As shown, the current user scenario 360 can be determined using scenario analysis model 340. In some embodiments, scenario analysis model 340 may include a feature extraction layer 341 and a scenario analysis layer 342. Feature extraction layer 341 may be used to determine activity features 350 based on activity data sequence 310. Scenario analysis layer 342 may be used to determine the current user scenario 360 based on activity features 350, current spatial information 320, and current time information 330.

[0066] In some embodiments, the scene analysis model 340 can take in the activity data sequence 310 and output the user's current scene 360.

[0067] In some embodiments, the scene analysis model 340 can be a machine learning model or a neural network model, such as a convolutional neural network (CNN) model.

[0068] In some embodiments, the feature extraction layer 341 may take an active data sequence 310 as input and output active features 350.

[0069] For more information on activity data sequence 310, please refer to the relevant description above.

[0070] In some embodiments, the processor may generate an activity data sequence 310 based on multiple activity data belonging to the same user and arranged in chronological order.

[0071] The activity features 350 can be features describing the user's activities. In some embodiments, the activity features 350 can include frequency features, duration features, trajectory features, activity time distribution features, activity space distribution features, etc. For example, the frequency features can include the number of activities per day, per week, etc., the trajectory features can include the trajectory shape, the total length, etc., the activity time distribution features can include the activity peak period, etc., and the activity space distribution features can include the activity density at each trajectory point, the activity range size, etc.

[0072] In some embodiments, the scene analysis layer 342 can input the activity features 350, the current spatial information 320, and the current temporal information 330, and output the user's current scene 360.

[0073] In some embodiments, the scene analysis model 340 can be obtained by jointly training the feature extraction layer 341 and the scene analysis layer 342.

[0074] In some embodiments, the initial feature extraction layer and the initial scene analysis layer can be trained based on a large number of first training samples with first labels. The first training samples can include sample activity data sequences, sample current spatial information, and sample current temporal information, and the first labels can be the actual user's current scene corresponding to the first training samples. The first training samples and the first labels can be determined based on historical data. In some embodiments, the actual user's current scene can be labeled according to the actual use situation in an artificial or automatic manner, such as a learning scene, an independent office scene, a meeting scene, etc.

[0075] An exemplary training process includes: inputting the sample activity data sequence into the initial feature extraction layer to obtain the activity features output by the initial feature extraction layer; inputting the activity features, the sample current spatial information, and the sample current temporal information as training data into the initial scene analysis layer to obtain the user's current scene output by the initial scene analysis layer; constructing a loss function based on the first label and the user's current scene output by the initial scene analysis layer, and synchronously updating the parameters of the initial feature extraction layer and the initial scene analysis layer. Through parameter updating, the trained feature extraction layer 341 and the scene analysis layer 342 are obtained.

[0076] In some embodiments, the input of the scene analysis model 340 can also include user features 370, etc. In some embodiments, when the input of the scene analysis model includes the user features 370, the user features 370, the activity features 350 output by the feature extraction layer, the current spatial information 320, and the current temporal information 330 can be input into the scene analysis layer 342, and the scene analysis layer 342 outputs the user's current scene 360. In some embodiments, when the input of the scene analysis model 340 includes the user features 370, the first training samples can also include sample user features.

[0077] For more information about the user features, see the relevant description below.

[0078] In some embodiments of the present disclosure, by inputting the activity data sequence, the current time information and the current space information into the scene analysis model to determine the current scene of the user, the relationship between the activity data, the current time information, the current space information of the user and the current scene of the user can be mined, the accuracy of determining the current scene of the user can be improved, and the automation level and the prediction efficiency of predicting the current scene of the user can be improved.

[0079] In some embodiments, the input of the scene analysis model 340 can further include the user features 370; determining the user features 370 can include: based on the activity data sequence of the user, searching in the user feature library according to a preset search condition to determine a matched reference activity data sequence; and based on the matched reference activity data sequence, determining the user features 370.

[0080] The user features can be information related to the user. For example, it can include the identification (such as number, name, etc.), age, gender, occupation, post, etc. of the user. The processor can read the user features preset by the user through the read-write interface, or obtain the user features or confirmation information of the user features input by the administrator user through the interactive interface, etc.

[0081] The user feature library can include a plurality of reference activity data sequences and corresponding reference user features.

[0082] The user features of the user who has used the current space can be manually annotated by the user or the administrator user as the reference user features. The reference user features of the user and the reference activity data sequence capable of reflecting the reference user features can be recorded in the user feature library of the storage device.

[0083] The reference activity data sequence can be a representative activity data sequence of the user, and thus can reflect the user features of the user.

[0084] The reference activity data sequence can be used for comparison with the activity data sequence of the user and provide reference of the user features. In some embodiments, each reference activity data sequence corresponds to a reference user feature.

[0085] In some embodiments, the processor can group the historical activity data sequences based on a plurality of historical activity data sequences by combining a pattern recognition algorithm of machine learning (such as support vector machine, etc.) or a neural network model (such as convolutional neural network, etc.), sort a set of historical activity data sequences corresponding to each user according to a preset sorting rule, and determine the historical activity data sequence with high ranking as the reference activity data sequence corresponding to the user. The preset sorting rule can include activity frequency from high to low, etc.

[0086] For more information about the historical activity data sequence, see the relevant description below.

[0087] In some embodiments, determining the reference activity data sequence can include: clustering a plurality of historical activity data sequences in the historical activity database; wherein the clustering distance includes the distance between the plurality of historical activity data sequences, and the clustering number is determined based on a user preset and / or the number of new users in the historical data; determining the reference activity data sequence based on at least one historical activity data sequence in each cluster obtained by clustering.

[0088] The historical activity data sequence can be a current space-related historical generated activity data and its arrangement. The processor can read the historical stored historical activity data sequence from the storage device.

[0089] The historical activity database can include a plurality of historical activity data sequences.

[0090] In some embodiments, clustering a plurality of historical activity data sequences in the historical activity database can include: measuring the distance, similarity or difference between the historical activity data sequences by a clustering distance measurement algorithm; clustering the plurality of historical activity data sequences according to the clustering distance and the clustering number by a clustering algorithm such as K-means (k-means clustering algorithm), DBSCAN (density-based clustering algorithm), etc., to obtain a plurality of clusters, each cluster including at least one historical activity data sequence.

[0091] The clustering distance can be used to measure the distance or similarity between the historical activity data sequences. In some embodiments, by comparing the clustering distance, the processor can divide similar historical activity data sequences into the same cluster. In some embodiments, the clustering distance includes the distance between the plurality of historical activity data sequences.

[0092] In some embodiments, the processor can calculate the distance between the historical activity data sequences by a clustering distance measurement algorithm, for example, the clustering distance measurement algorithm can include Euclidean distance, Manhattan distance, etc.

[0093] The clustering number can be the number of clusters that need to be divided in the clustering process.

[0094] In some embodiments, the processor can determine the clustering number based on a user preset number and / or the number of new users in the historical data. Wherein the user preset number can be the number of commonly used users in the current space.

[0095] The number of new users can be determined in the following manner: when a user's activity data sequence is searched in the user feature library according to a preset search condition, and no matching reference activity data sequence is searched, the user can be determined as a new user, and the processor can accumulate a flag in the system for recording the number of new users.

[0096] In some embodiments, the processor can also determine whether the user is a new user according to the user's activity data in the current space. For example, if the activity frequency is low, the user can be ignored and not recorded as a new user.

[0097] In some embodiments, the processor can calculate the average of at least one historical activity data sequence in each cluster based on the at least one historical activity data sequence in each cluster obtained by clustering, and determine the reference activity data sequence.

[0098] In some embodiments, the processor can also issue a reminder to label user features to the user or an administrator user when detecting the appearance of a new user, and obtain the labeled user features. The processor can record the activity data sequence of the user as a historical activity data sequence in the historical activity database together with the labeled user features. The method of detecting whether a new user appears can refer to the method of determining the number of new users described above.

[0099] In some embodiments, the processor can find a historical activity data sequence in the cluster that has been labeled with user features based on at least one historical activity data sequence in each cluster obtained by clustering, and take the user-labeled user features as the reference user features corresponding to the reference activity data sequence. In some embodiments, if the historical activity data sequences in the cluster have not been labeled with user features, the cluster can be ignored.

[0100] In some embodiments of the present specification, clustering a plurality of historical activity data sequences in the historical activity database, and determining a reference activity data sequence based on at least one historical activity data sequence in each cluster obtained by clustering, can group similar historical activity data sequences into the same cluster, thereby realizing the integration and induction of data, and being beneficial to extracting common features of data, and making the corresponding relationship between the reference activity data sequence and the user features more accurate.

[0101] In some embodiments, the preset search condition can include that the similarity between the user's activity data sequence and the reference activity data sequence in the user feature library is higher than a similarity threshold. In some embodiments, if the reference activity data sequence higher than the similarity threshold includes multiple reference activity data sequences, the preset search condition can also include the maximum similarity.

[0102] In some embodiments, the processor can calculate the similarity between the activity data sequence of the user and each reference activity data sequence based on the activity data sequence of the user, and determine the reference activity data sequence as a matched reference activity data sequence in response to the similarity satisfying a preset retrieval condition.

[0103] In some embodiments, the processor can determine the reference user feature corresponding to the matched reference activity data sequence as the user feature based on the matched reference activity data sequence.

[0104] In some embodiments of the present specification, by retrieving the matched reference activity data sequence from the reference user feature library based on the activity data sequence of the user, determining the user feature based on the matched reference activity data sequence, and taking the user feature as one of the inputs of the scene analysis model, the user feature can be quickly explored based on the current activity of the user, the current scene of the user can be analyzed based on the user feature, the user's light demand can be more accurately predicted, and the efficiency of determining and adjusting the light parameter can be improved.

[0105] For more information about the location information, the current space information, the current time information, and the current scene of the user, see Figure 2 and the related description.

[0106] In some embodiments of the present specification, by determining the activity data based on the current space information and the location information, and determining the current scene of the user according to the activity data and the current time information, the static information such as the current space information and the location information can be converted into dynamic information (activity data) related to the user's activity, and the activity data and the current time information when the activity occurs can be combined to provide a strong basis for determining the current scene of the user.

[0107] The processor can determine the light demand based on the current light effect and the at least one current scene of the user by various feasible methods, for example, and determine the light adjustment parameter based on the light demand. In some embodiments of the present specification, an exemplary process of determining the light adjustment parameter can include the following steps.

[0108] Step one, determining the light demand based on the at least one current scene of the user.

[0109] The light demand can be the demand of the user for the actual effect of the light in the environment. The content included in the light demand can be similar or the same as the content included in the current light effect (for example, including luminous flux, illuminance, etc.), and can be different in specific numerical value.

[0110] For more information about the current light effect, see Figure 2 and the related description.

[0111] In some embodiments, the processor can determine the light demand based on the current scene of the at least one user by various feasible manners, for example, by looking up a preset demand table. The preset demand table can include various scenes, current space information, current time information, number of users, etc. and the corresponding relationship with the corresponding light demand.

[0112] The number of users can be the total number of all users in the current space. For example, the processor can analyze the case of multiple trajectory points appearing at the same time point based on the at least one position data at the at least one time point, and determine the number of multiple trajectory points appearing at the same time point as the number of users.

[0113] For more information about the current space information and the current time information, see Figure 2 and the related description.

[0114] The preset demand table can be based on experience or preference settings. For example, the larger the size of the current space, the more users, the brightness of the light demand can be higher. For example, when the current time is in the afternoon, in order to alleviate user fatigue, the color temperature of the light demand can be lower (warm color).

[0115] In some embodiments, the processor can obtain the environmental characteristics of the current space; determine the current light comfort of the at least one user based on the environmental characteristics; and determine the light demand based on the current light comfort of the at least one user. For more information, see Figure 4 and the related description.

[0116] Step two, determine the light adjustment parameter based on the current light effect, the light demand and the current light parameter.

[0117] For more information about the current light effect and the light adjustment parameter, see Figure 2 and the related description.

[0118] The current light parameter can be the parameter that produces the current light effect. The content included in the current light parameter can be similar or the same as the content included in the light adjustment parameter, and the specific value can be different.

[0119] The processor can determine the light adjustment parameter based on the current light effect, the light demand and the current light parameter by various feasible algorithms. For example, the algorithm can include a vector matching algorithm.

[0120] The processor can construct a to-be-matched vector based on the current light effect, the light demand, and the current light parameter. The processor can perform a search in the vector database based on the to-be-matched vector, obtain a reference vector whose vector distance from the to-be-matched vector is less than a distance threshold, and determine the historical light adjustment parameter corresponding to the reference vector as the current required light adjustment parameter. The vector database is configured to store a plurality of historical vectors and historical light adjustment parameters corresponding to the historical vectors. The historical vectors are constructed based on historical current light effects, historical light demands, and historical current light parameters.

[0121] In some embodiments, the processor can determine an adjustment amplitude based on a difference between the current light effect and the light demand, and determine the light adjustment parameter based on the adjustment amplitude and the current light parameter.

[0122] In some embodiments, the processor can directly subtract the current light effect from the light demand to obtain the difference between the current light effect and the light demand. The difference can be a positive number or a negative number.

[0123] The adjustment amplitude can be a degree or a range of adjustment. The adjustment amplitude can be represented by a percentage, a level, a numerical value, or the like.

[0124] The processor can determine the adjustment amplitude based on the difference between the current light effect and the light demand by various feasible methods. For example, the processor can calculate a ratio of the difference to the current light effect, and determine the ratio as the adjustment amplitude. According to the positive or negative of the difference, the adjustment amplitude can be positive or negative accordingly.

[0125] In some embodiments, the processor can multiply the current light parameter by the adjustment amplitude (e.g., the ratio of the difference to the current light effect) to obtain an adjustment amount, add the current light parameter and the adjustment amount to determine the light adjustment parameter.

[0126] In some embodiments, the light adjustment parameter can also be related to a shadow condition. The shadow condition can be determined based on a luminaire position, a candidate light adjustment parameter, and a user station.

[0127] The shadow condition is information related to a shadow of a user or an internal object in the current space under the light. For example, the shadow condition can include a size, a shape, a position, a direction, a color, a brightness, or the like of the shadow.

[0128] The luminaire position is a position of the smart LED luminaire. The luminaire position can be represented as a coordinate or a relative displacement, or the like. The processor can obtain the pre-set luminaire position through a read-write instruction.

[0129] The candidate light adjustment parameter can be a parameter to be determined as the final light adjustment parameter for selection. In some embodiments, the processor can obtain the candidate light adjustment parameter in a random generation manner.

[0130] The user station can include a position where the user stays the longest in a time period and a posture. The user station can be obtained through the activity data sequence of the user.

[0131] In some embodiments, the processor can determine the shadow condition through a simulation algorithm based on the lamp position, the candidate light adjustment parameter, and the user station. For example, if the position of a certain smart LED lamp, the light adjustment parameter, and the user station are known, according to the principle of light propagation, the direction and size of the shadow of the user under the illumination of the lamp can be obtained through a simulation algorithm. In some embodiments, the processor can also obtain the elimination degree of other lights on the shadow (for example, the brightness of the shadow is increased by illuminating on the shadow) through a simulation algorithm in combination with the positions of other smart LED lamps and the light adjustment parameters.

[0132] In some embodiments, the simulation algorithm can include Real-time Ray Tracing, Volumetric Lighting, Real-time Shadow Mapping, etc.

[0133] In some embodiments, the light adjustment parameter can be positively correlated with the size of the shadow condition, or negatively correlated with the brightness of the shadow condition. For example, if the size of the user's shadow is larger and the brightness is lower, the adjustment amplitude can be increased, and then the light adjustment parameter is determined based on the increased adjustment amplitude and the current light parameter.

[0134] In some embodiments of the present specification, by associating the light adjustment parameter with the shadow condition, which is determined based on the lamp position, the candidate light adjustment parameter, and the user station, the spatial relationship and mutual influence between objects in the current space and between the user and the objects can be considered, which improves the rationality of the determination of the light adjustment parameter and is beneficial to further optimize the lighting effect.

[0135] In some embodiments of the present specification, by determining the adjustment amplitude based on the difference between the current light effect and the light demand, and determining the light adjustment parameter based on the adjustment amplitude and the current light parameter, the light effect can be closer to the user demand faster, and the efficiency of the light adjustment is improved.

[0136] In some embodiments of the present specification, by determining the light demand based on the current scene of the user, and determining the light adjustment parameter based on the current light effect, the light demand, and the current light parameter, the determination of the light adjustment parameter can be based on the user demand, which makes the adjustment process more flexible and meets the user demand in different scenes.

[0137] Figure 4is an exemplary schematic diagram of a comfort model according to some embodiments of the present specification.

[0138] In some embodiments, the processor can acquire an environmental feature of the current space; determine a current light comfort level of at least one user based on the environmental feature; and determine a light demand based on the current light comfort level of the at least one user.

[0139] The environmental feature can be an environmental-related feature of the current space. For example, it can include environmental light and / or environmental sound, etc. Among them, the environmental light can include at least one of light intensity, brightness, color temperature, etc. The environmental sound can include at least one of loudness, frequency, etc.

[0140] In some embodiments, the processor can acquire the environmental light and the environmental sound respectively through the light sensor and the microphone arranged in the current space or the smart LED lamp.

[0141] The current light comfort level can be the comfort level of the user under the current light.

[0142] The processor can determine the current light comfort level of the at least one user by comparing with the light standard and / or the noise standard based on the environmental feature. For example, when the loudness of the environmental sound exceeds the loudness specified by the noise standard (e.g., 50 decibels during the day, etc.), the more decibels exceeded, the lower the current light comfort level can be. If the current light comfort level is low, the light demand of the user can be softer, more uniform, and warmer color tone light.

[0143] In some embodiments, as Figure 4 shown, the processor can determine the current light comfort level 470 of the at least one user based on the environmental feature 410 through the comfort model 460.

[0144] In some embodiments, the comfort model 460 can input the environmental feature 410, the user current scene 420, the number of users 430, the at least one user feature 440, the candidate light effect 450, etc., and output the current light comfort level 470 of the at least one user.

[0145] For more information about the user current scene and the user feature, please refer to Figure 2 or other parts and their related descriptions.

[0146] The processor can acquire the candidate light effect 450 based on the adjustment of the current light effect. For example, under the current light effect, if the current light comfort level is low, the adjustment (e.g., softer, more uniform, and warmer color tone light effect than the current light effect) based on the current light effect is made to obtain the candidate light effect 450.

[0147] In some embodiments, when the output current light comfort level 470 is low, the processor can re-input the environmental features 410, the current user scene 420, the number of users 430, the at least one user feature 440, and the adjusted candidate light effect 450, re-predict and output the current light comfort level 470. When the output current light comfort level 470 reaches the preset comfort requirement, the output current light comfort level 470 at this time can be taken as the final output of the comfort model 460. The processor can store the candidate light effect 450 at this time as the preferred light effect.

[0148] In some embodiments, the comfort model 460 can be a machine learning model or a neural network model, such as a convolutional neural network model (CNN).

[0149] In some embodiments, the processor can train the comfort model 460 based on a plurality of normal training samples and second labels by a gradient descent method or the like. The normal training samples can include a plurality of sets of sample environmental features, sample current user scenes, sample numbers of users, sample user features, and sample candidate light effects, which can be obtained from historical data. The second label is the actual current light comfort level corresponding to each set of normal training samples. The actual current light comfort level can be labeled according to the frequency and interval of manual adjustment of the light by the user (or the average frequency and interval of manual adjustment of the light by a plurality of users) in each set of normal training samples in the historical data. For example, the lower the frequency and the longer the interval of manual adjustment of the light by the user, the higher the actual current light comfort level labeled by the processor.

[0150] In some embodiments, the processor can obtain the comfort model 460 by training an initial comfort model; the types of training data include normal training samples and gold standard training samples, the gold standard training samples are determined based on feedback information of a plurality of test users on a plurality of normal training samples; and the learning rate of the initial comfort model is related to the types of training data.

[0151] For more information about the normal training samples, see the relevant description above.

[0152] The gold standard training sample can be a sample with high reliability.

[0153] The feedback information can be the actual current light comfort level actually felt by the test user under each set of normal training samples.

[0154] In some embodiments, the processor can label the actual perceived current light comfort degree contained in the feedback information as a third label corresponding to the normal training sample based on the feedback information of the plurality of test users on the plurality of normal training samples, and determine the plurality of normal training samples and the corresponding third label as the gold standard training samples.

[0155] The learning rate can be a degree of parameter update in the training process of the initial comfort degree model. In some embodiments, the learning rate can be represented as a step size of gradient descent in the training process.

[0156] In some embodiments, when the training data input into the initial comfort degree model is the normal training sample, the processor can set the learning rate to a preset conventional learning rate (for example, any value in 0.01-0.001), and when the training data input into the initial comfort degree model is the gold standard training sample, the processor can set the learning rate to be greater than the preset conventional learning rate.

[0157] In some embodiments of the present specification, by determining the gold standard training sample based on the feedback information and using the gold standard training sample for training of the initial comfort degree model, the comfort degree model can better learn the relationship between the input data and the corresponding output data, so that the comfort degree model trained has good performance and generalization ability, and can more accurately predict the current light comfort degree of the user.

[0158] In some embodiments of the present specification, by determining the current light comfort degree of the at least one user based on the environmental features through the comfort degree model, the relationship between the environmental features and the current light comfort degree of the user can be mined, and the accuracy of the light comfort degree determination is improved, which is beneficial to subsequent determination of the light demand of the user.

[0159] The processor can determine the light demand based on the various current light comfort degrees and a preset corresponding relationship between the corresponding light demands, based on the current light comfort degree of the at least one user. For example, the lower the current light comfort degree, the lower the light demand, such as lower color temperature (warm color tone) and lower brightness. The preset corresponding relationship can be set according to experience or preference. In some embodiments, if the current light comfort degree is output and determined by the comfort degree model, the light demand can also be determined based on the preferred light effect stored by the processor.

[0160] In some embodiments, the processor can adjust the light demand based on the user state. The user state can include the user fatigue degree. The user fatigue degree can be determined according to the activity data of the user and the duration of the user in the current space.

[0161] The user state can be a manifestation in the aspects of body and emotion. In some embodiments, the user state can include the user fatigue degree and the like.

[0162] The user fatigue degree can reflect the fatigue degree of the user after the user is active in the current space.

[0163] In some embodiments, the processor can determine the user fatigue degree according to the activity data of the user and the duration of the user in the current space, according to a first positive correlation between the activity frequency in the activity data and the user fatigue degree, or according to a second positive correlation between the duration of the user in the current space and the user fatigue degree. For example, the higher the activity frequency or the longer the duration of the user in the current space, the higher the user fatigue degree.

[0164] In some embodiments, the processor can determine the duration of the user in the current space by calculating the time period between the time point corresponding to the first trajectory point in the activity trajectory and the time point corresponding to the last trajectory point in the activity trajectory according to the activity data of the user.

[0165] In some embodiments, the processor can adjust the current light comfort degree of the user based on the user fatigue degree, and determine the light demand based on the current light comfort degree. In some embodiments, the adjustment trend of the current light comfort degree can be in an opposite relationship with the high or low of the user fatigue degree. For example, the higher the user fatigue degree, the lower the current light comfort degree can be adjusted by the processor.

[0166] For more information about determining the light demand based on the current light comfort degree, see the relevant description above.

[0167] In some embodiments of the present specification, by determining the user fatigue degree according to the activity data of the user and the duration of the user in the current space, and adjusting the light demand based on the user state including the user fatigue degree, the change of the light demand that can occur after the user is active in the current space for a period of time can be automatically determined, and the light demand can be automatically adjusted according to the user state, which is beneficial to improve the flexibility of light adjustment.

[0168] In some embodiments of the present specification, the current light comfort degree of at least one user is determined by the environmental characteristics of the current space, and the light demand is determined based on the current light comfort degree, which combines the experience of environmental light, environmental sound and the like that can easily affect the user (for example, the attention, visual comfort, emotion, sleep quality and the like of the user) and is used in the intelligent LED lamp control system. The current light comfort degree of the user is automatically determined according to the environmental characteristics, and the light demand that meets the user in the environment is determined accordingly, which is beneficial to improve the comfort of the user, and can also improve the automation level of the product.

[0169] Some embodiments of the present specification provide an intelligent LED lamp, comprising at least one memory for storing computer instructions, and at least one processor for executing the computer instructions or part of the instructions to implement the intelligent LED lamp control method according to any one of the above-mentioned embodiments.

[0170] Some embodiments of the present specification provide a computer readable storage medium storing computer instructions, when a computer reads the computer instructions in the storage medium, the computer executes the intelligent LED lamp control method according to any one of the above-mentioned embodiments.

[0171] The above has described the basic concept, and it is obvious that the above detailed disclosure is only used as an example, and does not limit the present specification. Although it is not explicitly stated here, those skilled in the art can make various modifications, improvements and corrections to the present specification. Such modifications, improvements and corrections are suggested in the present specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of the present specification.

[0172] Meanwhile, the present specification uses specific words to describe the embodiments of the present specification. As "one embodiment", "an embodiment", and / or "some embodiments" means a certain feature, structure or characteristic related to at least one embodiment of the present specification. Therefore, it should be emphasized and noted that the "an embodiment" or "one embodiment" or "one alternative embodiment" mentioned in different places in the present specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present specification can be properly combined.

[0173] In addition, unless the claim explicitly states, the order of the processing elements and sequences described in the present specification, the use of numerals and letters, or the use of other names, is not intended to limit the order of the processes and methods of the present specification. Although some currently considered useful embodiments of the application are discussed in the above disclosure through various examples, it should be understood that such details are only for the purpose of illustration, and the additional claims are not limited to the disclosed embodiments, on the contrary, the claims are intended to cover all modifications and equivalent combinations that meet the spirit and scope of the embodiments of the present specification. For example, although the system components described above can be realized by hardware devices, they can also be realized by only software solutions, such as installing the described system on existing servers or mobile devices.

[0174] For simplicity and to facilitate understanding of one or more embodiments, a description of an embodiment sometimes refers to a plurality of features in a single embodiment, drawing, or description of an embodiment. However, this method of disclosure is not to be interpreted as meaning that the claimed embodiment requires more features than are explicitly recited in the claims. In fact, claims that do not specifically claim a combination of features are intended to cover the various possible combinations of features as would be understood by a person of ordinary skill in the art.

[0175] Some embodiments use numerical values to describe components, quantities of attributes. It should be understood that such numerical values used in the description of embodiments are, in some examples, modified by the adjectives "about," "approximately," or "substantially." Unless otherwise stated, "about," "approximately," or "substantially" indicate that the described numerical value allows for a variation of ±20%. Accordingly, numerical values used in the specification and claims of some embodiments are approximations that can vary depending on the desired characteristics of the individual embodiments. In some embodiments, numerical values used in the specification and claims are approximations that can vary depending on the desired characteristics of the individual embodiments. In some embodiments, numerical values should be considered in the context of the number of significant digits used in the number and the accepted bits of precision of the number. Although the numerical ranges and parameters setting forth the broadest scope of some embodiments of the specification are approximations, the numerical values set forth in the specific examples are reported as precisely as reasonably possible. The application is not limited to the specific numerical values set forth in the examples.

[0176] Each patent, patent application, patent publication, and other material, such as articles, books, specifications, publications, documents, and the like, referenced herein are hereby incorporated by reference in their entirety for the teachings relevant to the sentence and / or paragraph in which the reference is made. Discrepancies between document file histories and the contents of the present specification, except to the extent that the present specification otherwise provides, are hereby resolved by reference to the contents of the present specification. In the event of an inconsistency between a document file history and a description, definition, and / or a term used in the present specification, the description, definition and / or term as used in the present specification controls. In the event of an inconsistency between a document file history and a description, definition, and / or a term used in the present specification, the description, definition and / or term as used in the present specification controls.

[0177] Finally, it should be understood that the embodiments described herein are merely exemplary of the principles of the application. Other variations having essentially the same structure and function are within the scope of the claimed application. Accordingly, the application is not limited to the embodiments explicitly described and illustrated herein.

Claims

1. A light adjustment method characterized by, The method comprises: obtaining environmental characteristics of a current space; determining a current light comfort level of a user based on the environmental characteristics; the determination of the current light comfort level of the user based on the environmental characteristics comprises: determining the current light comfort level of the user based on the environmental characteristics, a current scene of the user, a number of users, user characteristics, and a candidate light effect through a comfort level model; the environmental characteristics comprise environmental light and environmental sound; the user characteristics comprise a number, a name, an age, a gender, an occupation, and a post of the user; the determination of the user characteristics comprises: based on a sequence of activity data of the user, searching in a user characteristic library according to a preset search condition to determine a matched reference sequence of activity data; the sequence of activity data is a sequence formed based on activity data, and the activity data comprises activity frequency, activity trajectory, trajectory points, trajectory point stay duration, and current time information; determining the user characteristics based on the matched reference sequence of activity data; the comfort level model is a machine learning model; the training data types of the comfort level model comprise ordinary training samples and gold standard training samples, and the gold standard training samples are determined based on feedback information of a plurality of test users on a plurality of ordinary training samples; in response to the training data types of the input initial comfort level model containing the gold standard training samples, the learning rate of model training is increased; the determination of the current scene of the user comprises: determining the current scene of the user based on a scene analysis model according to current time information, activity data of the user, and the user characteristics; the scene analysis model comprises a feature extraction layer and a scene analysis layer, and the scene analysis model is a machine learning model; the feature extraction layer inputs a sequence of activity data and outputs activity characteristics; the activity characteristics comprise frequency characteristics, duration characteristics, trajectory characteristics, activity time distribution characteristics, and activity space distribution characteristics; the scene analysis layer inputs the activity characteristics, current space information, and current time information and outputs a current scene of a user; the current space information comprises a type, a size, and an internal object arrangement of a current space. The scene analysis model is obtained by jointly training the feature extraction layer and the scene analysis layer, and includes training an initial feature extraction layer and an initial scene analysis layer based on first training samples with first labels; wherein the first training samples include sample activity data sequences, sample current spatial information, and sample current temporal information, and the first labels are actual user current scenes corresponding to the first training samples; the first training samples and the first labels are determined based on historical data; the sample activity data sequences are input into the initial feature extraction layer to obtain activity features output by the initial feature extraction layer; the activity features, the sample current spatial information, and the sample current temporal information are input into the initial scene analysis layer as training data to obtain user current scenes output by the initial scene analysis layer; a loss function is constructed based on the first labels and the user current scenes output by the initial scene analysis layer, and parameters of the initial feature extraction layer and the initial scene analysis layer are updated synchronously; through parameter updating, a trained feature extraction layer and a scene analysis layer are obtained; based on the current light comfort level, a light demand is determined; the light demand is adjusted based on a user fatigue level, and the user fatigue level is determined based on the activity data of the user and a duration of the user in the current space; Based on the current light effect, the light demand, and the current light parameters, a light adjustment parameter is determined, wherein the current light effect includes at least one of luminous flux, illuminance, light intensity, brightness, color temperature, and color rendering.

2. The method of claim 1, wherein, The determination of the activity data of the user includes: The position information of the user at at least one time point in the current space is recognized through a human body sensor; Based on the current spatial information and the position information of the user at the at least one time point, the activity data of the user is determined.

3. The method of claim 1, wherein, The user feature library includes a plurality of reference activity data sequences, and each reference activity data sequence corresponds to a reference user feature; The determination of the reference activity data sequence includes: A plurality of historical activity data sequences in a historical activity database are clustered; wherein the clustering distance includes the distance between the plurality of historical activity data sequences, and the number of clustering clusters is determined based on a user preset number and / or a number of new users; Based on at least one historical activity data sequence in each of the clustering clusters obtained by clustering, the reference activity data sequence is determined.

4. The method of claim 1, wherein, The determination of the light adjustment parameter based on the current light effect, the light demand, and the current light parameters includes: Based on the difference between the current light effect and the light demand, an adjustment amplitude is determined; Based on the adjustment amplitude and the current light parameters, the light adjustment parameter is determined.

5. The method of claim 1, wherein, The light adjustment parameter is also related to a shadow condition, and the shadow condition is determined based on a lamp position, a candidate light adjustment parameter, and a user position.

6. A light adjustment system, characterized by The system includes a parameter determination module, which is configured to: Obtain environmental features of a current space; Determine a current light comfort level of a user based on the environmental features; the determination of the current light comfort level of the user based on the environmental features includes: determine the current light comfort degree of the user based on the environment features, the current scene of the user, the number of users, the user features, and the candidate light effect, through a comfort degree model; the environment features include ambient light and ambient sound; the user features include the number, name, age, gender, occupation, and post of the user; the determination of the user features includes: based on the activity data sequence of the user, searching in a user feature library according to a preset search condition, and determining a matched reference activity data sequence; the activity data sequence is a sequence formed based on activity data, and the activity data includes activity frequency, activity trajectory, trajectory point, trajectory point stay duration, and current time information; based on the matched reference activity data sequence, determine the user features; the comfort degree model is a machine learning model; the training data types of the comfort degree model include ordinary training samples and gold standard training samples, and the gold standard training samples are determined based on feedback information of multiple test users on multiple ordinary training samples; in response to the training data types of the input initial comfort degree model containing the gold standard training samples, the learning rate of model training is increased; the determination of the current scene of the user includes: based on a scene analysis model, determining the current scene of the user according to the current time information, the activity data of the user, and the user features; the scene analysis model includes a feature extraction layer and a scene analysis layer, and the scene analysis model is a machine learning model; the feature extraction layer inputs the activity data sequence and outputs activity features; the activity features include frequency features, duration features, trajectory features, activity time distribution features, and activity space distribution features; the scene analysis layer inputs the activity features, current space information, and current time information, and outputs the current scene of the user; the current space information includes the type, size, and internal object placement of the current space; the scene analysis model is obtained by jointly training the feature extraction layer and the scene analysis layer, including training an initial feature extraction layer and an initial scene analysis layer based on a first training sample with a first label; wherein the first training sample includes a sample activity data sequence, sample current space information, and sample current time information, and the first label is the actual current scene of the user corresponding to the first training sample; the first training sample and the first label are determined based on historical data; the sample activity data sequence is input into the initial feature extraction layer to obtain the activity features output by the initial feature extraction layer; the activity features, the sample current space information, and the sample current time information are input into the initial scene analysis layer as training data to obtain the current scene of the user output by the initial scene analysis layer; a loss function is constructed based on the first label and the current scene of the user output by the initial scene analysis layer, and the parameters of the initial feature extraction layer and the initial scene analysis layer are updated synchronously; through parameter updating, the trained feature extraction layer and scene analysis layer are obtained; determine a light demand based on the current light comfort; the light demand is adjusted based on a user fatigue level, the user fatigue level is determined based on the activity data of the user and a duration of the user in the current space; determine a light adjustment parameter based on a current light effect, the light demand and a current light parameter, wherein the current light effect comprises at least one of luminous flux, illuminance, light intensity, brightness, color temperature and color rendering.

7. A light adjusting device, characterized by The apparatus comprises at least one memory and at least one processor, the at least one memory is configured to store computer instructions, and the at least one processor executes the computer instructions or part of the instructions to implement the light adjustment method in any one of claims 1-5.

Images