Summary of the invention
The purpose of the embodiment of the present invention is that provide a kind of illumination control method, apparatus and system, for realizing to light into
Row intelligent control to meet users ' individualized requirement automatically.
To achieve the goals above, the embodiment of the present invention provides a kind of illumination control method, applies in server, and described
Illumination control method includes: the voice data and device operational data that designated user is obtained from lighting system;By institute's predicate
Sound data are converted to text data, then carry out sentiment analysis to the text data to obtain the emotional value of this article notebook data, and
According to data obtaining time, by the text data and corresponding emotional value with the device operational Data Integration same
In one sample set;Effective sample relevant to illumination is filtered out from all sample sets;Based on it is all filter out it is effective
Sample constructs Lighting control decision model;And it in response to the illumination request from lighting system, is determined by the Lighting control
Plan model calculates lighting parameter, and calculated lighting parameter is fed back to the lighting system.
Optionally, carrying out sentiment analysis to the text data with the emotional value for obtaining this article notebook data includes: to described
Text data carries out semantic analysis, and extracts the emotion word to match with the lexical semantic in preconfigured sentiment dictionary,
It wherein include that preset multiple emotion words and each emotion word are corresponding for showing the feelings of the emotion word in the sentiment dictionary
Feel the weight of intensity;And the corresponding weight of all emotion words extracted is weighted and averaged to obtain text number
According to corresponding emotional value.
Optionally, including lamplight brightness, user's operation before time, text, emotional value, user's operation in the sample set
Lamplight brightness, user's operation headlight light color temperature and user's operation back light light color temperature afterwards.
Optionally, effective sample relevant to lamp lighting is screened from all sample sets includes: by each sample set
Conjunction is matched with preconfigured lamps and lanterns and illumination information dictionary;If can be extracted from the sample set with the lamps and lanterns and
The vocabulary that lexical semantic in information dictionary matches, then retain the sample set, otherwise filters out the sample set;And it is logical
It crosses emotion word and analyzes the evaluation for determining user to the sample set retained, if front evaluation or neutral evaluation, then by the sample
This set record is effective sample, if unfavorable ratings, then user is carried out the record before light adjustment for unfavorable ratings and made
Retain for an effective sample, and correspond to the sample record generated after adjusting and be also used as an effective sample, corresponding evaluation lacks
It saves as front evaluation.
It optionally, include: to be pre-configured with initially based on all effective sample building Lighting control decision models filtered out
Illumination mathematical model and illumination rule model, wherein the illumination data model is directed to fuzzy scene, the illumination rule mould
Block is directed to given scenario;All effective samples are handled using multiple linear regression method or non-linear regression method, with
Correct the illumination mathematical model;And revised illumination mathematical model and the illumination rule model are integrated to be formed
Final Lighting control decision model.
It is optionally, described that calculate lighting parameter by the Lighting control decision model include: from the illumination request
Middle extraction key message;If there is the rule to match with the key message in the illumination rule model, described in selection
Illumination rule model provides the lighting parameter for given scenario;If there is no believe with the key in the illumination rule model
The matched rule of manner of breathing then selects the illumination mathematical model to calculate illumination compatible with current time and current emotional value
Parameter.
On the other hand, the embodiment of the present invention provides a kind of machine readable storage medium, deposits on the machine readable storage medium
Instruction is contained, which is used for so that machine executes the above-mentioned illumination control method of the application.
On the other hand, the embodiment of the present invention also provides a kind of illumination control method, applies in lighting system, and the illumination
Control method includes: the voice data and device operational data for acquiring designated user;By the voice data of acquisition and
The device operational data are transmitted to server;And in response to user's operation, Xiang Suoshu server sends illumination request,
And lighting parameter that the server is fed back in response to illumination request is received to control each light source in lighting system by described
Lighting parameter shines.Wherein, the server is configured for: the voice data being converted to text data, then to described
Text data carries out sentiment analysis to obtain the emotional value of this article notebook data, and according to data obtaining time, by the textual data
Accordingly and corresponding emotional value and the device operational Data Integration are in same sample set;From all sample sets
Filter out effective sample relevant to illumination;Lighting control decision model is constructed based on all effective samples filtered out;And
In response to the illumination request, the lighting parameter is calculated by the Lighting control decision model, and the illumination is joined
Number feeds back to the lighting system.
On the other hand, the embodiment of the present invention also provides a kind of illumination control apparatus, applies in server, and the illumination is controlled
Device processed includes: data collection module, for obtaining the voice data and device operational number of designated user from lighting system
According to;Data processing module for the voice data to be converted to text data, then carries out sentiment analysis to the text data
To obtain the emotional value of this article notebook data, and according to data obtaining time, by the text data and corresponding emotional value with
The device operational Data Integration is in same sample set;Data screening module, for being sieved from all sample sets
Select effective sample relevant to illumination;Model construction module, for based on all effective sample building illumination controls filtered out
Decision model processed;And illumination determining module, for passing through the Lighting control in response to the illumination request from lighting system
Decision model calculates lighting parameter, and calculated lighting parameter is fed back to the lighting system.
Optionally, the data processing module includes: the first matching module, for carrying out semantic point to the text data
Analysis, and the emotion word to match with the lexical semantic in preconfigured sentiment dictionary is extracted, wherein in the sentiment dictionary
It is corresponding for showing the weight of the emotion intensity of the emotion word including preset multiple emotion words and each emotion word;Feelings
Inductance value determining module, for being weighted and averaged the corresponding weight of all emotion words extracted to obtain this article notebook data pair
The emotional value answered;And sample generation module, for foundation data obtaining time, by the text data and corresponding emotion
Value is with the device operational Data Integration in same sample set.
Optionally, including lamplight brightness, user's operation before time, text, emotional value, user's operation in the sample set
Lamplight brightness, user's operation headlight light color temperature and user's operation back light light color temperature afterwards.
Optionally, the data screening module includes: the second matching module, for by each sample set be pre-configured with
Lamps and lanterns and illumination information dictionary matched;First judgment module can be extracted for judging whether from the sample set
The vocabulary to match with the lexical semantic in the lamps and lanterns and illumination information dictionary;Sample reservation module, for can be from described
When extracting the vocabulary to match with the lexical semantic in the lamps and lanterns and illumination information dictionary in sample set, retain the sample set
It closes using as effective sample;Sample Filtration Block, for that cannot extracted from the sample set and the lamps and lanterns and illumination
When the vocabulary that the lexical semantic in information dictionary matches, the sample set is filtered out;And effective sample determining module, it is used for
The evaluation for determining user to the sample set retained is analyzed by emotion word, if front evaluation or neutral evaluation, then should
Sample set is recorded as effective sample, if unfavorable ratings, then user is carried out the record before light adjustment for unfavorable ratings
Retain as an effective sample, and corresponds to the sample record generated after adjusting and be also used as an effective sample, corresponding evaluation
Default to positive evaluation.
Optionally, the model construction module includes: model library module, is stored with preconfigured initial illumination mathematics
Model and illumination rule model, wherein the illumination data model is directed to fuzzy scene, the illumination rule module is for specified
Scene;Correction module, for handling all effective samples using multiple linear regression method or non-linear regression method,
To correct the illumination mathematical model;And decision model generation module, for by revised illumination mathematical model with it is described
Illumination rule model is integrated to form final Lighting control decision model.
Optionally, the illumination determining module includes: request processing module, for receiving the illumination request, and from institute
It states and extracts key message in illumination request;Second judgment module whether there is and the pass in the illumination rule for judging
The rule that key information matches;First choice module, for existing and the key message phase in the illumination rule model
When matched rule, the illumination rule model is selected to provide the lighting parameter for given scenario;And second selecting module,
For, there is no when the rule to match with the key message, selecting the illumination mathematical modulo in the illumination rule model
Type calculates lighting parameter compatible with current time and current emotional value.
On the other hand, the embodiment of the present invention also provides a kind of illumination control apparatus, applies in lighting system, and the illumination
Control device includes: data acquisition module, for acquiring the voice data and device operational data of designated user;Data pass
Defeated module, for the voice data of acquisition and the device operational data to be transmitted to server;Lighting control mould
Block, in response to user's operation, Xiang Suoshu server to send illumination request, and receives the server in response to illumination request
The lighting parameter of feedback is shone with controlling each light source in lighting system by the lighting parameter.Wherein, the server quilt
It is configured to: the voice data being converted into text data, then sentiment analysis is carried out to obtain this article to the text data
The emotional value of notebook data, and according to data obtaining time, the text data and corresponding emotional value are set with the illumination
Standby operation data is incorporated into same sample set;Effective sample relevant to illumination is filtered out from all sample sets;Base
Lighting control decision model is constructed in all effective samples filtered out;And in response to the illumination request, pass through the photograph
Bright control decision model calculates the lighting parameter, and the lighting parameter is fed back to the lighting system.
On the other hand, the embodiment of the invention also provides a kind of lighting control system, which includes: to include
Preceding claim is applied to the server of the illumination control apparatus of server end;And it is applied to lighting system end comprising above-mentioned
Illumination control apparatus lighting system.
Through the above technical solutions, the beneficial effect of the embodiment of the present invention is: the scheme of the embodiment of the present invention is available to be gone through
The optimal illumination parameter that the study of history data enables it be satisfied with to user in some period, solves existing lamp lighting system user's body
The problem for testing difference realizes and carries out intelligent control to light to meet users ' individualized requirement automatically, and is conducive to save electric energy.
The other feature and advantage of the embodiment of the present invention will the following detailed description will be given in the detailed implementation section.
Specific embodiment
It is described in detail below in conjunction with specific embodiment of the attached drawing to the embodiment of the present invention.It should be understood that this
Locate described specific embodiment and be merely to illustrate and explain the present invention embodiment, is not intended to restrict the invention embodiment.
" first, second " for being previously mentioned in embodiments of the present invention etc. is used for description purposes only, and should not be understood as indicating
Or it implies relative importance or implicitly indicates the quantity of indicated technical characteristic.
Fig. 1 is the flow diagram for the illumination control method of the embodiment of the present invention realized in the server.Wherein, it services
Device can be preferably Cloud Server.As shown in Figure 1, the illumination control method may comprise steps of:
Step S100 obtains the voice data and device operational data of designated user from lighting system.
Wherein, the lighting system can be it is conventional include light source, control switch, lighting controller, communication module (such as
Wifi module, 2G/3G/4G module, NB-Iot module) etc. Intelligent House Light system.
Preferably, making the lighting system can also include pickup module, such as microphone, to acquire when user talks
And the voice data of cache user can transmit to server and be protected when the voice data of user runs up to a certain amount of
It deposits.
In addition, device operational data refer to that user adjusts brightness, the colour temperature etc. of light by control switch etc.
Operation log data caused by saving, speech database and lamps and lanterns log database can be established in the server come store respectively from
The voice data and device operational data that lighting system obtains.It should be noted that the voice data and the illumination are set
It include corresponding data obtaining time in standby operation data, in order to the storaged voice number in chronological order in the server
According to device operational data.
Step 200, the voice data is converted into text data, then sentiment analysis is carried out to obtain to the text data
This article notebook data emotional value, and according to data obtaining time, by the text data and corresponding emotional value with it is described
Device operational Data Integration is in same sample set.
Wherein, in more preferred embodiment, as shown in Fig. 2, carrying out sentiment analysis to the text data to obtain
The emotional value of this article notebook data may include:
Step S201 pre-processes the text data.
Wherein, the pretreatment may include carrying out denoising etc. to text data.It should be noted that in some implementations
In example, pre-treatment step can also be not provided with.
Step S202, to the text data carry out semantic analysis, and extract in preconfigured sentiment dictionary
The emotion word that lexical semantic matches.
Wherein, carrying out semantic analysis to the text data includes being segmented, part-of-speech tagging and going to text data
Except stop words etc. is handled, relatively accurately to extract the emotion word being directed to.
It wherein, include that preset multiple emotion words and each emotion word are corresponding for showing the feelings in the sentiment dictionary
Feel the weight of the emotion intensity of word.It is " front respectively for example, presetting the emotion language of three ranks in the sentiment dictionary
Evaluation, unfavorable ratings or neutral evaluation ", such as front evaluation includes " happiness ", " liking ", " comfortable ", " special stick " emotion
Word, unfavorable ratings include the emotion words such as " disagreeable ", " not liking ", and neutrality evaluation includes the words such as " general ", " fine ", and corresponding
Weight may be configured as [- 1,1], wherein -1 is minimum unfavorable ratings, 0 is evaluated for neutral evaluation, 1 for highest front, and from as low as
Emotion intensity shown in high weight is also that from low to high, such as 0.8 relative to 0.6, and the emotion intensity shown is " more
Add and like ".Wherein, the specific size of weight can determine that the embodiment of the present invention is not intended to limit according to semantic dictionary and empirical value
In this.
Further, so-called semanteme matches, and refers to that the distance of the term vector of two vocabulary is less than the threshold value of setting, wherein
Term vector is one real vector of acquisition based on corpus training.Accordingly, the textual data can be judged by calculating term vector
The emotion word to match in the presence or absence of the meaning of a word with the vocabulary in sentiment dictionary.
Step S203 is weighted and averaged to obtain this article notebook data pair the corresponding weight of all emotion words extracted
The emotional value answered.
For a certain text data, multiple emotion words may be extracted after semantic analysis is carried out to it.In this regard, can to institute
There is the corresponding weight of emotion word to be weighted and averaged processing, and using weighted average treated result as this article notebook data correspondence
Emotional value.
According to emotional value obtained, text data can be recorded as gathering: (time, text, emotional value).
In addition, the device operational data can be converted by format, in lamps and lanterns log number in a manner of behavior vector
According to being stored in library.It can be set by device operational data record according to the format of behavior vector: (the time, before user's operation
Lamplight brightness, lamplight brightness after user's operation, user's operation headlight light color temperature, user's operation back light light color temperature).
Therefore, according to data obtaining time, above-mentioned text data set and lighting apparatus can be grasped by " time " dimension
1 set is integrated into as data acquisition system, it may be assumed that (time, text, emotional value, lamplight brightness before user's operation, user's operation back light
Brightness, user's operation headlight light color temperature, user's operation back light light color temperature), the set can be addressed further under for (time, text,
Emotional value, the lamplight brightness of current steady, the light colour temperature of current steady).In this way, realizing the text data and right
The emotional value answered and the device operational Data Integration are in same sample set.
Step S300 filters out effective sample relevant to illumination from all sample sets.
For multiple sample sets of acquisition, the emotion word in some sample sets may be unrelated with illumination, and this kind of sample
This set significant discomfort is preferably used to carry out Lighting control.
In this regard, in a preferred embodiment, as shown in figure 3, effective sample can be screened by following steps:
Step S301 matches each sample set with preconfigured lamps and lanterns and illumination information dictionary.
It wherein, include pre-set multiple and lamps and lanterns information and/or illumination in the lamps and lanterns and illumination information dictionary
The relevant vocabulary of information, such as " lamp ", " light ", " brightness ", " colour temperature ", " turning on light ", " turning off the light ", " brighter again ".If according to
Any vocabulary not being related in lamps and lanterns and illumination information dictionary in the sample set that the voice data at family obtains, then show user very
It may not be to be evaluated to current lighting state.
Step S302, judging whether can be from extraction in the sample set and the word in the lamps and lanterns and illumination information dictionary
The vocabulary that the semanteme that converges matches, it is no to then follow the steps S304 if so then execute step S303.
Wherein, it is also to be determined by term vector calculating that semanteme here, which matches, and details are not described herein.
Step S303 retains the sample set.
Step S304 filters out the sample set.
In this way, for example, the sample set finally retained may include following data:
Table 1
Here, the effective sample that finishing screen is selected can be supplied to subsequent modeling and use, and need before modeling to data
It is normalized, all data is transformed into (0,1) section.
Step S305 analyzes the evaluation for determining user to the sample set retained by emotion word;
Step S306, if front evaluation or neutral evaluation, then be recorded as effective sample for the sample set;
User is then carried out the record before light adjustment as one for unfavorable ratings if unfavorable ratings by step S307
A effective sample, and correspond to the sample record generated after adjusting and be also used as effective sample, corresponding evaluation defaults to positive evaluation.
Wherein, corresponding evaluation defaults to front evaluation and comments it can be appreciated that corresponding emotional value dimension is denoted as front
Valence.
Specifically, the side of effective sample is determined referring to the corresponding sample set of table 1 for step S305 to step S307
Method can be such that
1) it is analyzed by emotion word, user involved in a sample set is being recorded to " the front of current lighting state
When evaluation or neutral evaluation ", an effectively record is generated: (t, l, c, s);
2) if it is unfavorable ratings, illuminating effect can be adjusted in general user, first will be " time, bright before adjustment
Degree, coloration, emotional value " is remained as an effective sample, then, user is subjected to light adjustment for unfavorable ratings
Corresponding " time, brightness, coloration, the emotional value " generated is also stored as effective sample afterwards, and wherein emotional value has corresponded to just
Face emotion.
Wherein, the side of " front evaluation, unfavorable ratings or the neutral evaluation " for current lighting state is expressed for user
Formula can have following two situation:
A) user expresses the satisfactory feedback to current lighting state by way of voice, records an effective sample
(t, l, c, s);
B) user feeds back not over voice mode, since current time t, if (l, c) (can also be in 1 minute
Other setting times) it remains unchanged, then default user is satisfied to current lighting state, records last fixed value (l, c), enables s
=0.95, obtain effective sample.
C) user is expressed dissatisfied to current lighting state by voice mode, and (is also possible to other to set in 1 minute
Fix time) light is adjusted, then corresponding sample set (t ', l ', c ', s ') is effective sample after recording adjustment.
Here, the effective sample finally obtained can be stored into the sample database of server for subsequent modeling calling.
Step S400 constructs Lighting control decision model based on all effective samples filtered out.
In a preferred embodiment, as shown in figure 4, the construction method of Lighting control decision model may include following step
It is rapid:
Step S401 is pre-configured with initial illumination mathematical model and illumination rule model.
Wherein the illumination data model is directed to fuzzy scene, and the illumination rule module is directed to given scenario.Specified field
Scape can be regarded as needing the special scenes of specific illumination scheme, such as " sleep scene ", which is generally defaulted as closing all
Light, without adjusting light according to the emotional value of user.In this way, illumination rule model refers to according to the user monitored
With the specific condition of scene, activate specific illumination scheme, for example, receive user sending " sleep " phonetic order after, root
According to rule model, light-off program will start.Fuzzy scene shows the lighting demand that user does not determine under the scene, belong to
The opposite scene of given scenario may be assumed that " illumination scheme (selection of brightness and colour temperature) " by time and use under fuzzy scene
The influence of family emotion.In this way, illumination mathematical model refers to and can be closed after the emotion or mood for knowing active user according to user
In the historical data of illumination, provide the favorite lighting method of user, i.e., its can predict " certain user when certain time,
If possessing certain mood, what the illumination scheme most possibly selected is ".
In addition, initial illumination mathematical model and illumination rule model is storable in the model library of server.Wherein, just
The illumination mathematical model of beginning be arranged according to the light status that generally adapts to of masses, such as " Emotion expression is just hair excessive fire,
Lamps and lanterns issue warm light, are conducive to relieving mood ", it is subsequent to learn initial illumination number by the effective sample for designated user again
Model is learned, to obtain the illumination mathematical model for meeting users ' individualized requirement;It is preset in initial illumination rule model multiple
Rule, these rules are from common sense and domain knowledge summary, such as " should close all lamps during sleep ".
Step S402 handles all effective samples, to correct the illumination mathematical model.
Wherein it is possible to all effective samples are handled using multiple linear regression method or non-linear regression method,
To correct the illumination mathematical model, but the embodiment of the present invention and it is limited to multiple linear regression method and nonlinear regression side
Method.
For example, the step of correcting illumination mathematical model using multiple linear regression method can be such that
Assuming that L (X), T (X) respectively indicate luminance function and colour temperature function, wherein X is the feature vector of description illumination, X
=(x1,x2,…,xn).Specifically, x1Indicate affective variable value, x2It indicates the time, then has formula:
Lα(X)=α0+α1x1+α2x2=α0x0+α1x1+α2x2;
It is exactly with vector representation: Lα(X)=αTX, wherein α indicates parameter to be estimated.
Further, α is solved by gradient descent methodT, in which:
Wherein, m is the number of samples in sample set;γ is learning rate, an adjustable parameter;"y(i)" it is sample set
In i-th of sample brightness value;Respectively indicate the result of parameter after+1 iteration of iteration j and jth (below
Formula have the same meaning).Work as α0、α1、α2Change be less than certain threshold value when, it is believed that iteration convergence.
Similarly, colour temperature function T (X) can be indicated are as follows:
Tβ(X)=β0+β1x1+β2x2;
It is exactly with vector representation: Lβ(X)=βT X。
It is equally possible that can seek parameter beta using gradient descent method, details are not described herein.
So far, by the correlation formula of above-mentioned L (X) and T (X), given time, emotional value is can use, brightness is sought
Value, color temperature value.
Revised illumination mathematical model and the illumination rule model are integrated to form final photograph by step S403
Bright control decision model.
In this way, illumination mathematical model and illumination rule model are combined, unified Lighting control decision is formd
Model.
Step S500 is calculated in response to the illumination request from lighting system by the Lighting control decision model
Lighting parameter, and calculated lighting parameter is fed back into the lighting system.
Wherein, illumination request is generated by lighting system in response to user's operation, and its object is to wake servers execution
The calculating of lighting parameter.
In a preferred embodiment, as shown in figure 5, lighting parameter is calculated by the Lighting control decision model can be with
The following steps are included:
Step S501 extracts key message from the illumination request.
For example, in the illumination request being correspondingly formed, vocabulary " is slept for the phonetic order " starting to sleep " that user issues
Feel " then obviously belong to key message.
Step S502 judges in the illumination rule with the presence or absence of the rule to match with the key message.
Step S503 selects institute if there is the rule to match with the key message in the illumination rule model
It states illumination rule model and provides lighting parameter for given scenario.
Step S504, if being selected in the illumination rule model there is no the rule to match with the key message
The illumination mathematical model calculates lighting parameter compatible with current time and current emotional value.
Specifically, if successful match, matched rule is activated in real time, chromatic value and temperature value needed for providing illumination;
If matching is unsuccessful, show the lighting demand (fuzzy scene) that user does not determine, it can be according to current time and current emotion
Value provides lighting parameter, and the lighting parameter is the illumination that user most possibly selects under current time and current emotional value
Parameter.For example, if phonetic order " sleep " matches with " sleep " rule in illumination rule model, according to the regular mould of illumination
Type starts light-off program, if receiving other kinds of instruction, according to user feeling state, scene and the time of acquisition
Etc. factors, using illumination mathematical model calculate most suitable one group of lighting parameter.
In conclusion above-mentioned illumination control method is applied to server end, relative to existing lighting system, have following several
The advantages of a aspect:
1) learn the optimal illumination parameter (brightness and the color that enable it be satisfied in some period to user using historical data
Temperature), solve the problems, such as existing lamp lighting system poor user experience, and realize based on user feeling intelligent control light
Purpose.
2) lamps and lanterns are allow " to observe and learn " feelings of people by experience so that the adjusting of colour temperature and brightness is more scientific by correction model parameter
Thread, and by the change to light come the mood for the people that releives, so that user be made to possess more healthy and comfortable experience.
3) personalized illumination scheme is more advantageous to energy conservation, does not waste electricity.
4) there is certain entertainment, the enjoyment of life at home can be increased.
Fig. 6 is the flow diagram for the illumination control method of the embodiment of the present invention realized in the illumination system, wherein according to
Bright system can be it is conventional include light source, control switch, lighting controller, communication module (such as wifi module, 2G/3G/4G
Module, NB-Iot module) etc. Intelligent House Light system.As shown in fig. 6, the illumination control method may include following step
It is rapid:
Step S601 acquires the voice data and device operational data of designated user.
Wherein, pickup module can be configured in the illumination system to acquire voice data, such as at the wall near wall lamp
Microphone is installed.Device operational data are then generated in user's operation control switch or voice control light and are cached to photograph
The lamps and lanterns log memory module of bright system.
The voice data of acquisition and the device operational data are transmitted to server by step S602.
Wherein, lighting system can pass after storing a certain amount of voice data and device operational data, then by data
Transport to server.
Step S603, in response to user's operation, Xiang Suoshu server sends illumination request, and receives the server response
It is shone in the lighting parameter of illumination request feedback with controlling each light source in lighting system by the lighting parameter.
Step S603 can be executed by the lighting controller of lighting system, and the lighting parameter is by executing above-mentioned answer
It is obtained for the illumination control method of the server, details are not described herein.
Further, it after lighting system receives the lighting parameter that server end transmits, is generated before this based on lighting parameter
Illumination scheme, then by illumination scheme be converted to signal light control instruction, the lighting part of light source execute the signal light control instruction with into
Row shines.In addition, illumination scheme is also possible to be generated by user's operation lighting system, such as user adjusts lamp by touch switch
The illumination scheme that lighting system automatically generates when brightness.
Above-mentioned illumination control method is applied to lighting system, which can be based on machine learning structure by server end
The Lighting control decision model built obtains the lighting parameter suitable for user feeling demand, can satisfy users ' individualized requirement,
To realize the purpose based on user feeling intelligent control light.
The above are the exemplary illustrations to method in the embodiment of the present invention, it will be understood by those skilled in the art that on
Stating step or process in method can concurrently or sequentially execute under the premise of can be realized the object of the invention, in above description
Context does not represent stringent timing limitation.
Fig. 7 is the structural schematic diagram of the illumination control apparatus according to an embodiment of the present invention realized in the server.Such as Fig. 7
Shown, the illumination control apparatus includes: data collection module 10, for obtaining the voice data of designated user from lighting system
With device operational data;Data processing module 20, for the voice data to be converted to text data, then to the text
Notebook data carries out sentiment analysis to obtain the emotional value of this article notebook data, and according to data obtaining time, by the text data
And corresponding emotional value and the device operational Data Integration are in same sample set;Data screening module 30 is used
In filtering out effective sample relevant to illumination from all sample sets;Model construction module 40, for being based on all screenings
Effective sample out constructs Lighting control decision model;And illumination determining module 50, in response to from lighting system
Illumination request calculates lighting parameter by the Lighting control decision model, and calculated lighting parameter is fed back to institute
State lighting system.
In a preferred embodiment, as shown in figure 8, the data processing module 20 may include: the first matching module 21,
For carrying out semantic analysis to the text data, and extracts and match with the lexical semantic in preconfigured sentiment dictionary
Emotion word, wherein including that preset multiple emotion words and each emotion word are corresponding for showing the feelings in the sentiment dictionary
Feel the weight of the emotion intensity of word;Emotional value determining module 22, for the corresponding weight of all emotion words extracted
It is weighted and averaged to obtain the corresponding emotional value of this article notebook data;And sample generation module 23, for according to data acquisition
Time, by the text data and corresponding emotional value and the device operational Data Integration in same sample set
In.
In a preferred embodiment, as shown in figure 9, the data screening module 30 may include: the second matching module 31,
For each sample set to be matched with preconfigured lamps and lanterns and illumination information dictionary;First judgment module 32, is used for
Judge whether the word that can be matched from extraction in the sample set with the lexical semantic in the lamps and lanterns and illumination information dictionary
It converges;Sample reservation module 33, for can be from being extracted in the sample set and the word in the lamps and lanterns and illumination information dictionary
Converge the semantic vocabulary to match when, retain the sample set using as effective sample;Sample Filtration Block 34, for cannot be from
When extracting the vocabulary to match with the lexical semantic in the lamps and lanterns and illumination information dictionary in the sample set, this is filtered out
Sample set;And effective sample determining module 35, determine user to the sample set retained for analyzing by emotion word
Evaluation if unfavorable ratings, then will if front evaluation or neutral evaluation, then be recorded as effective sample for the sample set
User carries out the record before light adjustment as an effective sample for unfavorable ratings, and the sample record generated is corresponded to after adjustment
Also as being effective sample, and default it is denoted as positive evaluation.
In a preferred embodiment, as shown in Figure 10, the model construction module 40 may include: model library module 41,
It is stored with preconfigured initial illumination mathematical model and illumination rule model, wherein the illumination data model is for fuzzy
Scene, the illumination rule module are directed to given scenario;Correction module 42, for using multiple linear regression method or non-
Linear regression method handles all effective samples, to correct the illumination mathematical model;And decision model forms module 43,
For being integrated revised illumination mathematical model and the illumination rule model to form final Lighting control decision model
Type.
In a preferred embodiment, as shown in figure 11, the illumination determining module 50 may include: request processing module
51, key message is extracted for receiving the illumination request, and from the illumination request;Second judgment module 52, for sentencing
With the presence or absence of the rule to match with the key message in the illumination rule of breaking;First choice module 53, for described
When there is the rule to match with the key message in illumination rule model, the illumination rule model is selected to provide for finger
Determine the lighting parameter of scene;And second selecting module 54, for being not present in the illumination rule model and the key
When the rule that information matches, the illumination mathematical model is selected to calculate photograph compatible with current time and current emotional value
Bright parameter.
Above-mentioned apparatus is corresponding with the illumination control method being previously used in server, before specific illustrative description can refer to
It states in method and illustrates, details are not described herein.
Figure 12 is the structural schematic diagram for the illumination control apparatus of the embodiment of the present invention realized in the illumination system.Such as Figure 12
Shown, the illumination control apparatus may include: data acquisition module 61, for acquiring voice data and the illumination of designated user
Equipment operation data;Data transmission module 62, the voice data and the device operational data for that will acquire pass
Transport to server;Lighting control module 63, in response to user's operation, Xiang Suoshu server to send illumination request, and receives
The server is joined in response to the lighting parameter that illumination request is fed back with controlling each light source in lighting system by the illumination
Number shines.
Wherein, for data acquisition module 61, it may include pickup module can also include to acquire the voice data of user
Lamps and lanterns log memory module is to cache device operational data.
It wherein, is preferably wifi module for data transmission module 62.
Wherein, raw based on lighting parameter after the lighting parameter for receiving server end transmission for lighting control module 63
At illumination scheme, then by illumination scheme be converted to signal light control instruction, the lighting part of light source execute the signal light control instruction with
It carries out luminous.In addition, illumination scheme is also possible to the scheme as caused by user's operation lighting system, such as user passes through touch-control
The scheme generated when switch adjustment lamplight brightness, so that lighting control module is in the lighting parameter for receiving server end feedback
And while carrying out luminous, also it can change lamplight brightness or colour temperature according to the execute-in-place of user.
It should be noted that lighting parameter is obtained by the above-mentioned illumination control apparatus applied to the server,
This is repeated no more.
In addition, above-mentioned apparatus is corresponding with the illumination control method being previously used in lighting system, specific illustrative description
It can refer in preceding method and illustrate, details are not described herein.
Figure 13 is the structural schematic diagram of the lighting control system of the embodiment of the present invention.As shown in figure 13, the Lighting control system
System includes: server 70, and it includes the above-mentioned illumination control apparatus applied to server;And lighting system 80, packet
Containing the above-mentioned illumination control apparatus applied to lighting system.In addition, the lighting control system can also include gateway 90,
It is responsible for receiving the data from lighting system, and uploaded onto the server 70, and data of the reception from server 70, and
Lighting system 80 is passed to, to realize the communication between server 70 and lighting system 80.Here, server 70 and illumination
System 80 can also realize distribution by other means, and the embodiment of the present invention is not restricted to this.
The optional embodiment of the embodiment of the present invention is described in detail in conjunction with attached drawing above, still, the embodiment of the present invention is simultaneously
The detail being not limited in above embodiment can be to of the invention real in the range of the technology design of the embodiment of the present invention
The technical solution for applying example carries out a variety of simple variants, these simple variants belong to the protection scope of the embodiment of the present invention.
It is further to note that specific technical features described in the above specific embodiments, in not lance
In the case where shield, it can be combined in any appropriate way.In order to avoid unnecessary repetition, the embodiment of the present invention pair
No further explanation will be given for various combinations of possible ways.
It will be appreciated by those skilled in the art that implementing the method for the above embodiments is that can pass through
Program is completed to instruct relevant hardware, which is stored in a storage medium, including some instructions are used so that single
Piece machine, chip or processor (processor) execute all or part of the steps of each embodiment the method for the application.And it is preceding
The storage medium stated includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), magnetic or disk.
In addition, any combination can also be carried out between a variety of different embodiments of the embodiment of the present invention, as long as it is not
The thought of the embodiment of the present invention is violated, equally should be considered as disclosure of that of the embodiment of the present invention.