CN112533318A - Dimming mapping module for scene-type classroom intelligent lighting control device - Google Patents

Abstract

The invention provides a dimming mapping module for a scene-based classroom intelligent lighting control device, which comprises a mapping input part, a mapping planning part, a mapping determination part, a mapping storage part and a mapping output part, wherein the mapping input part is used for inputting a scene-based classroom intelligent lighting control device; a BP neural network is established in the mapping determination part, the input quantity of the BP neural network is an LED string driving current parameter of the adjustable light bank, and the output quantity is a color temperature parameter and an illumination parameter of each measuring point in a classroom; the mapping planning part acquires a training sample set of the neural network when the dimmable lamp set is dimmed in a stepping mode; and the trained BP neural network maps the current value combination corresponding to the individual to be evaluated into the color temperature and illumination value of each test point in the multi-objective optimization processing process, so as to calculate the score value of the individual under different lighting scenes. The invention can carry out grading mapping on various current combinations, thereby ensuring that the optimized lighting conditions can meet the requirements of different learning scenes such as self-repair, projection, discussion, break in class and the like, and improving the pertinence of lighting control.

Description

Dimming mapping module for scene-type classroom intelligent lighting control device

The application is a divisional application with the application number of 201910287214.4, application date of 2019, 03/04/03 and the title of scene-based classroom intelligent lighting system, control device and optimization and control method.

Technical Field

The invention belongs to the field of intelligent lighting, and particularly relates to a dimming mapping module for a scene-type classroom intelligent lighting control device.

Background

Schools are learning places, lighting needs to fully consider enough illumination intensity needed when reading and writing, and appropriate light sources need to be selected and a comfortable and natural lighting environment needs to be designed and/or combined. The classroom is a place where students stay for the longest time every day, and improper light not only influences the learning of the students, but also can cause the shortsightedness of the students. In the past, the illumination in a classroom generally only needs to be bright enough, and the classroom is not easy to be broken. Nowadays, with social progress and development of lighting technology, lighting reaching standards is the minimum requirement of classroom lighting, and further, indexes such as eyesight protection, energy conservation, environmental protection, user experience and the like are considered.

At present, most of the existing classrooms, whether in middle and primary schools or colleges, are designed in advance. During the actual operation of the lighting system, the lamp can only be controlled by a few simple switches. The arrangement of the lamps is generally optimized according to the illumination requirement of the desk surface, and the assumed application is that all main areas of a classroom are all on or all off.

With the advent and development of intelligent lighting technology, independent, dimming lamp control means have been very easily obtained and implemented. Therefore, in recent years, regulations are provided in lighting design specifications of various classrooms, and it is required that a conditional school preferably selects an intelligent lighting control system in combination with natural lighting to realize constant illumination control and dimming control.

On the other hand, intelligent lamp control needs to consider the lighting effects of different lighting conditions. There are many aspects of the environmental lighting factors that affect human vision, of which the more important are: illumination level, illumination distribution, color appearance, light and shade colors, etc., which affect the operation differently. The human eye has two functions, namely a photosensitive function, namely, light reaches the eye ground through an optical system of the eye to form an object image on the retina; the other is a visual signal processing function, namely, the retina converts and processes the light energy of the object image into nerve impulses, and the impulses are transmitted into human eyes through ganglion cells, so that visual and non-visual effects are generated. After the third kind of photoreceptor cells on the retina of the human eye, namely the intrinsic photoreceptor retinal nerve knot cells, are discovered, it has been proved that the third kind of photoreceptor cells can control the human circadian rhythm, biological clock and human eye pupil size by generating a series of chemical and biological reactions to the visible radiation entering the human eye, thereby having influence on the human physiology, psychology and the like. In recent years, research results show that ambient lighting not only has an influence on various physiological parameters of people, such as blood pressure, heart rate, melatonin and the like, but also has a remarkable influence on the learning efficiency and visual function of people.

Currently, many schemes for automatic control have emerged for classroom lighting. Among the existing classroom lighting schemes, some focus on the illuminance distribution of a desk when an artificial light source is used for lighting at night, and as in the application of the invention patent in China with the patent number of 2016110057838, the illuminance distribution of a blackboard lamp is uniform through light distribution design at different angles for multiple times. In the aspect of lamp control, the number of people is detected by a camera and the illumination is adjusted according to the difference of the number of people in each subarea in the Chinese patent application with the patent number of 2017109662294; the chinese patent application No. 2014107085532 discloses that the light and curtain are controlled according to whether the projector is used in the multimedia classroom light. These control schemes are designed to intelligently control the lighting of a target such as illumination according to the requirements of different applications, but classroom lighting is a complex system, and the lighting itself is closely related to the psychological and physiological activities of people. According to the requirements of different activities in classroom learning on multiple aspects of illumination, an intelligent regulation and control device and system which comprehensively consider various indexes such as illumination, cold and warm color temperature, energy conservation and the like are needed.

SUMMARY OF THE PATENT FOR INVENTION

Aiming at the problem of influence of different lighting conditions on learning in the existing classroom lighting, the invention aims to provide a dimming mapping module for a scene-type classroom intelligent lighting control device, the scene-type classroom intelligent lighting control device is applied to, under an off-line state, the color temperature and the illumination of each test point under different driving current combinations of a dimmable lamp set in a classroom are obtained through sample collection, a dimming lighting distribution table is established, when the device is applied online, the current activity modes in the classroom, such as self-repair, projection, blackboard-writing teaching, discussion, break in class and the like, are identified based on classroom images and human body sensor signals, the photochromic scoring functions of the color temperature and the illumination are adjusted according to the characteristics of different activities, the scoring values based on 4 factors of the color temperature, the illumination and energy conservation are obtained, and the driving current values of each LED string of the dimmable lamp set are optimized by a multi-target optimization algorithm, the optimizing result is transmitted to the driver to change the light emission of the lamp group, so that intelligent illumination suitable for various learning activities in a classroom is realized.

The technical solution of the present invention is to provide a dimming mapping module for a scene-based classroom intelligent lighting control device, which comprises a mapping input part, a mapping planning part, a mapping determination part, a mapping storage part and a mapping output part, wherein the mapping input part further comprises a photochromic parameter input port and a driving current parameter input port,

the light color parameter input port receives color temperature and illumination parameters of multiple test points acquired after sampling processing by the external light color sensing module, the driving current parameter input port receives LED string driving current parameters which correspond to the color temperature and the illumination values and are sent to the external adjustable light bank,

the mapping determination part is provided with a BP neural network, the input quantity of the BP neural network is the LED string driving current parameter, the output quantity is the color temperature and illumination parameter,

the mapping planning part receives the training sample of the BP neural network through the mapping input part, trains the BP neural network by using the sample, and stores the connection weight of the BP neural network after training in the mapping storage module part.

Preferably, the input quantity of the BP neural network is n paths of LED string driving current values, the output quantity is 2m parameters of color temperature and luminance value obtained after the test point signals of m light color sensing modules in the light color sensing unit are processed by the light color processing module,

the model of the BP neural network is as follows:

the jth node of the hidden layer outputs

The p-th node of the output layer outputs

Wherein the f () function is taken as sigmoid function, w_ijAnd v_jpRespectively the connection weight from the input layer to the hidden layer and the connection weight from the hidden layer to the output layer, a_jAnd b_pAnd (3) respectively representing hidden layer and output layer thresholds, wherein p is 1, 2m, and k is the number of nodes of the hidden layer, and the network training is carried out by adopting a gradient descent method.

Preferably, the dimming mapping module is disposed in a host unit of the intelligent lighting control device for a scene-based classroom, the host unit further includes an input module, a light color processing module, an image processing module, a lighting optimization processing module, an output module, and a storage module, and the host unit is configured to:

the input module respectively acquires color temperature and illumination signals, classroom scene images and personnel on-site signals in a preset area from a photochromic sensing unit, an image acquisition unit and a human body detection unit of the system,

based on the color temperature and the illumination signal, the light color processing module acquires a color temperature value, an illumination value and an illumination uniformity value of each region of the classroom,

the image processing module extracts image features and personnel position features based on a scene detector trained by a training image set, and identifies the lighting scene of the image according to the image features and the personnel position features,

the mapping planning part sends a dimming signal to the dimmable lamp set in a step-by-step changing mode in the dimming mapping module, acquires training sample sets of the neural network under different light environments through the light color parameter input port and the driving current parameter input port, trains the BP neural network by using the training sample sets in an off-line mode, adjusts the connection weight of the BP neural network in the mapping determination part, and stores the connection weight in the mapping storage module part;

based on the BP neural network, established light color grading functions of various color temperatures and illuminations under different lighting scenes of self-repair, projection, blackboard writing teaching, discussion and break in a classroom, and standard reaching levels and power consumption of the uniformity of the illuminations of various areas of the classroom relative to reference values, the lighting optimization processing module optimizes the LED strings in an available value space range of the driving current values of the LED strings of the adjustable light banks through a multi-objective optimization algorithm, and transmits optimization results to drivers of the corresponding LED strings through the dimming mapping module and the output module; in the multi-objective optimization processing process, the mapping input part transmits n paths of LED string driving current value combinations corresponding to the individual to be evaluated to the mapping determination part, the BP neural network trained in the mapping input part maps the current value combinations into color temperature and illumination values of all test points, and after the color temperature and illumination values are output through the mapping output part, the illumination optimization processing module calculates the score value of the individual to be evaluated according to a light color scoring function aiming at the color temperature and illumination values of all the test points.

Preferably, in the process of multi-objective optimization processing, the color temperature and the illuminance value of each test point output by the mapping output part calculate the score value of the individual to be evaluated according to the following photochromic score function:

for color temperature, when its maximum human expectation value belongs to a medium-high color temperature, its score function is,

when the color temperature is most expected by a plurality of persons to belong to the medium and low color temperatures, the scoring function is,

wherein, W is the current color temperature, bW and cW are the lower limit and the upper limit of the middle and high expected color temperature interval covering the number of people with the set proportion obtained according to statistics in the current scene, aW and dW are the other two preset lower limits and upper limits in the current scene, respectively, and hW is the upper limit of the middle and low expected color temperature interval covering the number of people with the set proportion obtained according to statistics in the current scene.

Preferably, in the process of multi-objective optimization processing, the color temperature and the illuminance value of each test point output by the mapping output part calculate the score value of the individual to be evaluated according to the following photochromic score function:

for illumination, the scoring function is,

wherein, E is the current illumination, bE and cE are the lower limit and the upper limit of the interval covering the expected illumination value of the people with the set proportion obtained according to statistics in the current scene, and aE and dE are the other two preset lower limits and upper limits in the current scene respectively.

Preferably, in the process of multi-objective optimization processing, the color temperature and the illuminance value of each test point output by the mapping output part calculate the score value of the individual to be evaluated according to the following photochromic score function:

for uniformity of illumination, the scoring function is,

wherein, U is the current illuminance uniformity, bU is a reference value set according to the standard, and aU is a preset lower limit value.

Preferably, the total score calculation formula corresponding to the individual to be evaluated is as follows:

f＝k₁·f₁+k₂·f₂+k₃·f₃+k₄·f₄，

wherein f is₁Is the color temperature score value, f₂Is an illuminance score value, f₃An illuminance uniformity score calculated from the standard level of illuminance uniformity for each region of the classroom relative to a reference value, f₄For calculating an energy saving score value k based on the power consumed_iAnd (i-1, 2, 3, 4) is a preset weighting coefficient.

Preferably, in the process of multi-objective optimization processing, the color temperature and the illuminance value of each test point output by the mapping output part calculate the score value of the individual to be evaluated according to the following photochromic score function:

for a self-repairing scene, an illumination value interval corresponding to the maximum score value of the illumination scoring function is limited to 300-500 Lx, and a color temperature value interval corresponding to the maximum score value of the color temperature scoring function is limited to 4500-6500K;

for a projection scene, from a projection cloth to a classroom back row, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination score function are gradually increased;

for a blackboard writing teaching scene, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination score function of a blackboard area are larger than those of other areas;

for the discussion scene, the color temperature value range corresponding to the maximum score value of the color temperature scoring function is set according to the middle-low expected color temperature range, and the upper limit value of the color temperature value range is limited below 4000K;

for a break scene in a class, an illumination value range corresponding to the maximum score value of the illumination scoring function is limited to 300-320 Lx, and a color temperature value range corresponding to the maximum score value of the color temperature scoring function is limited to 3300-5000K.

Preferably, the input quantity of the BP neural network is n paths of LED string driving current values, and the output quantity is 2m parameters of color temperature and luminance value obtained after the test point signals of m light color sensing modules in the light color sensing unit are processed by the light color processing module.

Preferably, the input quantity of the BP neural network further includes ambient light intensity, and color temperature and illuminance of ambient light at each test point in a classroom.

Compared with the prior art, the scheme of the invention has the following advantages: the invention evaluates various combinations of the LED string driving current in the dimmable lamp set by establishing the dimming illumination distribution table and the light color scoring function, and the light color scoring function can be adjusted according to different learning activity scenes in a classroom, so that the optimization solution obtained by multi-objective optimization processing can meet different requirements of different learning scenes on illumination, such as self-repair, projection, blackboard writing teaching, discussion, break in class and the like, the pertinence of illumination control is improved, the illumination effect is optimized, and a more humanized illumination environment is provided for classroom learning. Meanwhile, the illumination and the color temperature of classroom illumination can be optimized according to different weather conditions, a natural and comfortable illumination environment is created for the classroom, and the experience of learners is improved.

Drawings

Fig. 1 is a structural diagram of a scene-type classroom intelligent lighting control device and a scene-type classroom intelligent lighting system;

FIG. 2 is a block diagram of the light color processing module and the image processing module;

FIG. 3 is a schematic diagram of a classroom lighting environment;

FIG. 4 is a schematic diagram of the distribution of lamp sets and detection points in a classroom;

FIG. 5a is a diagram illustrating an illuminance scoring function;

FIG. 5b is a graph showing an illuminance uniformity score function;

FIG. 5c is a diagram illustrating a mid-to-high color temperature scoring function;

FIG. 5d is a diagram illustrating a mid-to-low color temperature scoring function;

FIG. 6 is a diagram of a neural network based dimming mapping module;

FIG. 7 is a schematic diagram of a neural network architecture for dimming mapping;

fig. 8 is a flow chart of a scene-based classroom intelligent lighting optimization;

fig. 9 is a workflow diagram of a scene-based classroom intelligent lighting control method.

Wherein:

2000 scene type classroom intelligent lighting system, 100 host unit, 200 user interface unit, 300 light color sensing unit, 400 image acquisition unit, 500 human body detection unit, 600 adjustable light lamp group, 700 server, 1000 scene type classroom intelligent lighting control device,

110 input module, 120 light color processing module, 130 image processing module, 140 illumination optimization processing module, 150 dimming mapping module, 160 output module, 170 storage module,

a 121 illumination detector, a 122 color temperature detector,

an eye opening detector of 131, a scene detector of 132,

151 mapping input part, 1511 light color scoring input part, 1512 driving current parameter input part, 152 mapping planning part, 153 mapping determination part, 154 mapping storage part, 155 mapping output part, 156 light modulation output part,

210, a display screen, 220 an operation panel,

310 light color sensing module, 410 image acquisition module, 510 human body detection module, 610 driver, 620LED lamp, 800 desk, 900 projection cloth.

Detailed Description

Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention is not limited to only these embodiments. The invention is intended to cover alternatives, modifications, equivalents and alternatives which may be included within the spirit and scope of the invention.

The invention is described in more detail in the following paragraphs by way of example with reference to the accompanying drawings. It should be noted that the drawings are in simplified form and are not to precise scale, which is only used for convenience and clarity to assist in describing the embodiments of the present invention.

Example 1

With reference to fig. 1 and fig. 9, the present embodiment provides a method for controlling intelligent lighting in a scene-based classroom, including the following steps:

s1, initializing, establishing a dimming illumination distribution model and a light color grading model in the host unit,

a dimmable light bank arranged in a classroom comprises a plurality of LED lights, each LED light has at least one LED string with adjustable color temperature and brightness,

the dimming illumination distribution model is the corresponding relation from all n paths of LED string driving current values of the dimming lamp set to color temperature values and illuminance values at each test point of a classroom, the test point is the test position of m light color sensing modules in the light color sensing unit,

the photochromic scoring model is a photochromic scoring function of various color temperatures and illuminations under different lighting scenes such as self-repair, projection, blackboard writing teaching, discussion, break in class and the like in a classroom;

s2, sending dimming signals to each LED string of the dimmable lamp set in a stepping change mode through the output module of the host unit, collecting signal samples of each lighting environment after dimming change, extracting n paths of LED string driving current values in the dimming signals, collecting and processing to obtain color temperature, illuminance value and illuminance uniformity value at each test point, recording and storing the current, color temperature and illuminance values in a dimming lighting distribution model,

meanwhile, under various lighting environments of different scenes, a learner in the classroom scores the lighting conditions, and the scores are counted to respectively adjust the light color scoring functions of color temperature and illumination intensity under different lighting scenes;

s3, during online control, based on a scene detector trained by a training image set, extracting image features and personnel position features aiming at a current classroom scene image and a personnel on-site signal corresponding to the image, and identifying a current lighting scene according to the image features and the personnel position features;

s4, optimizing the current parameters of each LED string by adopting a multi-objective optimization algorithm,

firstly, determining a strategy for coding and decoding the driving current parameters of each LED string, determining respective value intervals of the driving current parameters,

determining parameters such as population scale, cross probability, mutation probability and the like in optimization calculation;

s5, randomly generating an initial population;

s6, aiming at each individual body in the evolution group in the search space, decoding to obtain n paths of LED string driving current parameters, searching a dimming illumination distribution model through multi-dimensional interpolation according to the current parameters to obtain corresponding color temperature and illumination value, and respectively calculating color temperature scoring values f based on color temperature and illumination light color scoring functions corresponding to the currently identified illumination scene according to the color temperature and illumination value₁Illuminance score value f₂Calculating the grading value f of the illuminance uniformity according to the standard level of the illuminance uniformity of each region of the classroom relative to the reference value₃Calculating the energy-saving score f according to the power consumption₄Weighting and summing the 4 scoring values to calculate a total scoring value f corresponding to the individual,

s7, performing cross inheritance and mutation operations based on the total score value, and updating an evolutionary population;

s8, turning to the step S6, iterating repeatedly until the search is finished, and outputting a Pareto optimization solution;

and S9, the dimming mapping module transmits the optimizing result to the driver of each corresponding LED string through the output module, and the driver performs dimming by changing the driving current of the LED string.

Example 2

The embodiment provides a scene-based classroom intelligent lighting control device and a dimming mapping module for the scene-based classroom intelligent lighting control device.

As shown in fig. 1, the scenic classroom intelligent lighting control apparatus 1000 includes a host unit 100, a user interface unit 200. The user interface unit 200 includes an operation panel 220 including keys and a display screen 210 for parameter input and manual control operations. The host unit 100 includes an input module 110, a light color processing module 120, an image processing module 130, a lighting optimization processing module 140, a dimming mapping module 150, an output module 160, and a storage module 170.

The input module 110 respectively obtains setting parameters and user operation instructions, color temperature and illumination signals, classroom scene images and on-site personnel signals in a preset area through the user interface unit 200, the light color sensing unit 300, the image acquisition unit 400 and the human body detection unit 500.

Referring to fig. 2, the light color processing module 120 includes an illuminance detector 121 and a color temperature detector 122, based on the color temperature and the illuminance signal input from the input module 110, the color temperature detector 122 in the light color processing module 120 processes the color temperature signal to obtain a color temperature value of each area in the classroom, and the illuminance detector 121 processes the color temperature signal to obtain an illuminance value and an illuminance uniformity value.

The image processing module 130 includes a scene detector 132 trained by using a training image set, and the scene detector 132 extracts image features of the current classroom scene image collected by the image collecting unit, and identifies the current lighting scene according to the image features and the personnel position features identified according to the personnel presence signals detected by the human body detecting unit.

The dimming mapping module 150 obtains a dimming illumination distribution table of each LED string by changing a driving current value of each LED string of the dimmable light set 600 in a classroom and recording a color temperature value, an illuminance value and an illuminance uniformity value of each area when each corresponding LED string is illuminated by combination lighting; and establish the photochromic scoring function of various color temperatures and illuminations under different lighting scenes such as self-repair, projection, blackboard writing teaching, discussion, break in class and the like in a classroom; the dimmed lighting distribution table and the light color scoring function are also maintained in a memory module. And calculating the grading value of different value combinations of each LED string in the dimmable lamp set through the dimming illumination distribution table and the light color grading function.

Therefore, based on the dimming illumination distribution table, the light color scoring function, the identified illumination scene, and the standard reaching level and the power consumption of the illuminance uniformity of each region of the classroom relative to the reference value, the illumination optimization processing module 140 optimizes the driving current value of each LED string of the dimmable lamp bank within the space range of the allowable value through a multi-objective optimization algorithm, and transmits the optimization result to the driver of the corresponding LED string through the dimming mapping module and the output module. The multiple targets comprise 4 grading indexes of color temperature, illumination uniformity, energy conservation and the like of the current lighting scene.

Referring to fig. 1 and 3, the dimming mapping module 150 outputs a dimming command to the dimmable light set 600 through the output module 160, where the dimming parameter set includes a driving current value of each LED string. Preferably, the dimmable light bank 600 has n LED strings in common.

Referring to fig. 4, after the LED strings emit light, light on the desk surface of the desk 800 is detected by the light color sensing modules 310 distributed at m test points, and the sensing signals are processed by the light color processing module 120 to obtain color temperature and illuminance values at each test point. Preferably, the light color sensing module 310 may be disposed on a desk surface or suspended on a lamp housing of the LED lamp 620 through a bracket. When the desk is hung, the light color sensing module can adopt a silicon photocell as a sensor, and the focusing lens is aligned to the target desk surface to detect the reflected light intensity sensing signal of the target desk surface, and the light color processing module also calibrates the reflected light intensity and the illumination of the target desk surface through experiments and detects the illumination of the measuring point of the desk surface on line according to the calibration relation.

In the dimming illumination distribution table, each record comprises n paths of LED string driving current values of the dimming lamp group and color temperature and luminance values obtained after test point signals of m light color sensing modules in the light color sensing units are processed by the corresponding light color processing modules.

And under the state without background light, sending dimming signals to the dimmable lamp set in a stepping change mode through the dimming mapping module and the output module, detecting the changed light environment, acquiring the color temperature and the illuminance value at the test point, and continuously repeating until the recorded sample covers the value interval of the driving current of each LED string.

Preferably, the dimming illumination distribution table can be represented by a BP neural network to estimate the light color corresponding to each driving current combination.

As shown in fig. 6 and 7, the dimming map module 150 includes a map input unit 151, a map planning unit 152, a map determining unit 153, a map storage unit 154, a map output unit 155, and a dimming output unit 156, and the map input unit 151 further includes a photochromic parameter input port 1511 and a driving current parameter input port 1512. The dimming output part 156 outputs a dimming instruction signal to the dimmable lamp set through the output module.

The photochromic parameter input port 1511 receives color temperature and illuminance values obtained after test point signals of the m photochromic sensing modules are processed from the input module; the driving current parameter input port 1512 transmits n-channel LED string driving current parameters corresponding to the color temperature and illuminance values, which are sent to the dimmable light set. The map determining unit 153 establishes a BP neural network, where the input quantity of the BP neural network is the parameter value of the driving current of the n LED strings, and the output quantity is the color temperature and illuminance value of the m test points.

The mapping planning unit 152 is configured to collect training sample sets of the neural network in different light environments through the light color parameter input port 1511 and the driving current parameter input port 1512 when the dimming mapping module sends a dimming signal or other dimming operations to the dimmable lamp set in a step-by-step changing manner; and off-line train the neural network by using the training sample set, adjust the connection weight of the neural network in the mapping determination section 153, and store the connection weight in the mapping storage module section 154.

In the multi-objective optimization process, the mapping input unit 151 transmits the n LED string driving current value combinations to be evaluated to the mapping determination unit 153, the current value combinations are mapped to the color temperature and illuminance values of the test points by the BP neural network trained therein, and the values are output by the mapping output unit 155, and then the lighting optimization processing module calculates the values of the scores based on the light color scoring function.

As shown in fig. 7, the model of the BP neural network is:

the jth node of the hidden layer outputs

The p-th node of the output layer outputs

Wherein the f () function is taken as sigmoid function, w_ijAnd v_jpRespectively the connection weight from the input layer to the hidden layer and the connection weight from the hidden layer to the output layer, a_jAnd b_pAnd (3) respectively representing hidden layer and output layer thresholds, wherein p is 1, 2m, and k is the number of nodes of the hidden layer, and the network training is carried out by adopting a gradient descent method.

Based on the dimmed lighting distribution table, the host unit is further configured to:

in the processing process of the multi-objective optimization algorithm, firstly, initialization is carried out, a strategy for encoding n paths of LED string driving current parameters is determined, and respective value intervals are determined; secondly, aiming at each individual in the evolution group in the search space, searching a dimming illumination distribution table to obtain the color temperature and the illumination value of each test point corresponding to the individual based on the n-path driving current parameter values, and respectively calculating the color temperature scoring value f of the color temperature and the illumination value of each test point obtained by searching according to the color temperature and the illumination color scoring function of the identified illumination scene₁Illuminance score value f₂Calculating the grading value f of the illuminance uniformity according to the standard level of the illuminance uniformity of each region of the classroom relative to the reference value₃Calculating the energy-saving score f according to the power consumption₄And carrying out weighted summation on the 4 scoring values to calculate a total scoring value f-k corresponding to the individual₁·f₁+k₂·f₂+k₃·f₃+k₄·f₄Wherein k is_i(i1, 2, 3, 4) is a preset weighting coefficient, and performs inheritance, crossover and mutation operations according to the total score value to update an evolutionary population; and then, repeatedly evolving the population until the optimization is finished, and outputting an optimization result.

Under different lighting scenes such as self-repair, projection, blackboard writing teaching, discussion, break between classes and the like, the lighting conditions expected by learners in a classroom are different. For example, for self-repair scenes, a higher color temperature of the illumination is desirable to improve alertness and attention; in projection teaching, in order to improve resolving power, it is desirable that the ambient light near the projection cloth be kept at a dark level; when the blackboard writing is taught, the blackboard is expected to have higher illumination; in the discussion scenario, the color temperature may be turned lower for lively atmosphere; when the student has a rest in class, the light can be generally dimmed to relax the thought that the student is continuously tensed for a period of time, so that the combination of work and rest is realized.

For this reason, it is desirable that the lighting control device be able to automatically identify the scene and adjust the light color scoring function according to the scene.

As shown in connection with fig. 4, the image acquisition unit 400 includes one or more image acquisition modules 410 suspended below the ceiling of a classroom. Preferably, the camera of the image capturing module 410 may adopt a wide-angle lens, and may also adopt a plurality of cameras to capture images of different target areas respectively.

Sample images under various lighting scenes such as self-repairing, projection, blackboard writing teaching, discussion, break in class and the like in a classroom are collected, and a training image set is formed by the sample images to train the scene detector.

Preferably, the scene detector may utilize a support vector machine, SVM, of linear kernel functions as a classifier for the different lighting scenes. For various lighting scene classes, a classifier is trained based on features of the respective target regions, in particular human activity or location features. And forming a feature vector of the SVM by using the position of the person in each specific area in the image in the classroom, the aggregation state and the feature value of the related sensor data.

The training image set is derived from a segmented image from the wide-angle image or the region image itself acquired for each different target region. The identification of the lighting scene is performed for each sample of the training image set, and the untrained recognition portion is trained on the basis of the training image set.

The specific lighting scene identification is defined according to the personnel distribution characteristics and the human activity characteristics of the scene. For example, for a self-repairing scene, the scene features are that most learners sit on chairs beside desks, and for the self-repairing scene, the scene detector can mainly identify the image features of the proportion of the sitting personnel to the total number of people in the classroom.

Referring to fig. 4, for the break scene, the human body detection module 510 facing the pedestrian passageway in the human body detection unit 500 detects the movement of multiple persons, and the number of standing persons in the image can be greatly increased compared with the number of persons in self-repair, and the image features and the person position features of the human body detection module together form the feature vector of the scene.

In another example, for a projected teaching scene, features of the projected cloth 900 falling may be detected in the image, and features of the image of the scene where most people are sitting may be obtained. Preferably, a projector startup signal detection unit connected with the host unit can be further arranged, and when the projector is started, the host unit can acquire the information and automatically identify the current scene as the projection teaching scene.

For discussion scenes, the image features of one or several people standing in turn can be detected.

Preferably, the feature of the feature vector may be a static image feature at a time point, or a human body movement feature detected in a plurality of continuous images. For example, several people in a discussion scenario take turns standing.

Preferably, a part of the sample images of the training image set can be left as a verification set to verify the trained scene detector.

Preferably, the scene detector comprises a plurality of recognition modules, one recognition module of the plurality of recognition modules corresponds to human body detection in a local area of a classroom, each recognition module corresponds to a classifier, for example, a podium area can be detected independently, and image characteristics of the area can be used as a main basis for teaching mode judgment.

After a scene detector is trained by a training image set, images of a classroom and sensing data corresponding to the images are obtained, image features and feature values of human body sensing data are extracted, and lighting scenes of the images are determined according to the image features and the feature values of the sensing data.

On the basis of scene detection, various light color scoring functions are defined according to the requirements of classroom learning on illumination.

For illumination, as shown in fig. 5a, the scoring function is,

wherein, E is the current illumination, bE and cE are the lower limit and the upper limit of the interval covering the expected illumination value of the people with the set proportion obtained according to statistics in the current scene, and aE and dE are the other two preset lower limits and upper limits in the current scene respectively. Wherein, the set proportion is a value between 0.85 and 0.95.

For illuminance uniformity, as shown in fig. 5b, the scoring function is,

wherein, U is the current illuminance uniformity, bU is a reference value set according to the standard, and aU is a preset lower limit value;

the illuminance uniformity is taken as the ratio of the minimum illuminance to the average illuminance of the target class desktop in the classroom, and according to the general standard, the value of bU is 0.7 or higher, and the value of aU is between 0.55 and 0.6.

For color temperature, when its maximum human expectation value belongs to a medium-high color temperature, as shown in fig. 5c, its score function is,

when the color temperature is most expected by a person to belong to the medium-low color temperature, as shown in fig. 5d, the scoring function is,

wherein, W is the current color temperature, bW and cW are the lower limit value and the upper limit value of the middle and high expected color temperature interval which is obtained according to statistics and covers the number of people with the set proportion under the current scene, aW and dW are the other two preset lower limit values and the upper limit values under the current scene respectively, and hW is the upper limit value of the middle and low expected color temperature interval which is obtained according to statistics and covers the number of people with the set proportion under the current scene; wherein the set ratio is between 0.85 and 0.95.

For the energy saving index, the score value f₄Is defined as:

wherein u is_k、i_kRespectively the driving voltage, current, P of n-way LED string₀Is the sum of the rated powers of all the LED strings.

Preferably, the photochromic scoring function is adjusted based on the current identified scene:

for a self-repairing scene, an illumination value interval corresponding to the maximum score value of the illumination scoring function is limited to 300-500 Lx, and a color temperature value interval corresponding to the maximum score value of the color temperature scoring function is limited to 4500-6500K;

for a projection scene, from a projection cloth to a parallel strip-shaped area at the back row of a classroom, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination scoring function are gradually increased, and in the scene, scoring calculation is carried out according to the strip-shaped area and the scores of the areas are summed;

for a blackboard writing teaching scene, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination score function of a blackboard area are larger than those of other areas;

for the discussion scene, the color temperature value range corresponding to the maximum score value of the color temperature scoring function is set according to the middle-low expected color temperature range, and the upper limit value of the color temperature value range is limited below 4000K;

for a break scene in a class, an illumination value interval corresponding to the maximum score value of the illumination scoring function is limited between 300 Lx and 320Lx, and a color temperature value interval corresponding to the maximum score value of the color temperature scoring function is limited between 3300K and 5000K.

Preferably, the definition of the photochromic scoring function is further adjusted using the following rule:

in rainy days or other weather, when the natural illumination in the classroom is insufficient, the supplementary illumination is carried out through the adjustable light set, if the current time belongs to the range from 8 am to 5 pm, the total score value f of each individual in the evolution group in the optimized search space is adjusted according to the similarity between the color temperature to be scored, the illumination light color parameter and the current sunlight color:

f′＝f·(1+η)，

η＝α·sim(W，Wnow)+(1-α)·sim(e，enow)，

wherein, α is a set coefficient, and W, e is a color temperature value to be scored, a ratio of the illumination intensity to the maximum illumination intensity, namely two values of the relative illumination intensity; wnow and enow are the color temperature and the relative illuminance of the sunlight at the current time in the weather forecast obtained from the weather forecast server 700, respectively, the relative illuminance is the ratio of the brightness of the current sunlight to the brightness of the sunlight at noon, the similarity function sim (,) adopts a normal distribution function or a triangular distribution function with the second parameter as the center, the distribution amplitude is set according to the value range of the first parameter, η is an adjustment coefficient, and f' are the scores before and after adjustment, respectively.

Preferably, the definition of the photochromic scoring function is further adjusted using the following rule:

when the curtain is pulled up to block direct sunlight in rainy days or sunny days, if natural illumination in a classroom is insufficient, supplementary illumination is carried out by the adjustable light bank in the daytime:

for the color temperature, in rainy days, the color temperature value range corresponding to the highest score value of the color temperature scoring function is limited to 4000-5000K; in sunny days, the color temperature value range corresponding to the highest score value of the color temperature scoring function is limited to be 5000-6500K,

for the illumination, the total score value f of each individual in the evolutionary population in the optimized search space is adjusted according to the similarity between the illumination parameter to be scored and the current sunlight:

f′＝f·(1+η)，

η＝α·sim(e，enow)，

wherein alpha is a set coefficient, and e is the ratio of the illumination to be scored to the maximum illumination, namely the relative illumination; enow is the relative illuminance of sunlight at the current time in the weather forecast obtained from the weather forecast server 700, the relative illuminance of sunlight is the ratio of the current sunlight brightness to the midday sunlight brightness, the similarity function sim (,) adopts a normal distribution function or a triangular distribution function with the second parameter as the center, the distribution amplitude is set according to the value range of the first parameter, η is an adjustment coefficient, and f' are the score values before and after adjustment, respectively.

For the study on the desk, the detection and analysis of the study efficiency factors can be carried out by an image processing method. Different from the state when the illumination is moderate and the attention is concentrated, when the illumination is unreasonable, the physiological parameters can change to different degrees, and the physiological parameters can be used as the basis for monitoring the learning efficiency state. When the light is too dark or too bright to influence the learning efficiency of the learner, the eyelids of the human are closed downwards, and the opening degree of the eyes is obviously reduced.

As shown in fig. 2 and 3, an eye opening detector 131 is disposed in the image processing module 130, and the eye opening detector 131 obtains the eye opening of the learner in the target area through image processing. On the basis of the segmentation definition of the scoring function based on statistics, the eye opening of the learners in the classroom is detected on line through image processing, and the total scoring value f of each individual in the evolutionary population in the optimized search space is adjusted according to the eye opening:

f′＝f·(1+η)，

wherein η is an adjustment coefficient, and the value of η is half power of the ratio of the current average eye opening to the maximum eye opening.

Preferably, when the scheme of the invention is used for illumination automatic optimization control, the illumination switching between different scenes is transited in a stepping mode, such as break between classes, and the switching between the classes is completed in a mode that the driving current of each LED string is gradually changed in a set time.

Example 3

The embodiment provides a scene-based classroom intelligent lighting system. As shown in fig. 1, the scene-based classroom intelligent lighting system 2000 includes a user interface unit 200, a light color sensing unit 300, an image acquisition unit 400, a human body detection unit 500, a dimmable light set 600, a server 700, and a host unit 100 connected to the user interface unit 200, the light color sensing unit 300, the image acquisition unit 400, the human body detection unit 500, the dimmable light set 600, and the server 700.

Wherein, the user interface unit 200 includes a display screen 210 and an operation panel 220 for entering parameters and initiating operations; the light color sensing unit 300 includes a plurality of light color sensing modules, and is used for detecting light color parameters such as illumination, color temperature, etc.; the image acquisition unit 400 is used for acquiring an image of a lighting scene in a classroom; the human body detection unit 500 is used for detecting human body information of a preset area in a classroom; the dimmable light set 600 comprises a plurality of LED lights 620, the LED lights 620 have at least one LED string with adjustable color temperature and brightness, and each LED light adjusts the driving current of the LED string inside the LED light through one driver 610.

The host unit 100 includes an input module 110, a light color processing module 120, an image processing module 130, a lighting optimization processing module 140, a dimming mapping module 150, an output module 160, and a storage module 170. The host unit 100 is configured to:

the input module receives setting parameters and user operation instructions through the user interface unit, respectively acquires color temperature and illumination signals, classroom scene images and personnel on-site signals in a preset area from the photochromic sensing unit, the image acquisition unit and the human body detection unit,

based on the color temperature and the illumination signal, the light color processing module acquires a color temperature value, an illumination value and an illumination uniformity value of each region of the classroom,

the image processing module extracts image characteristics and personnel position characteristics aiming at the scene image and the personnel on-position signals corresponding to the image based on a scene detector trained by a training image set, and identifies the illumination scene of the image according to the image characteristics and the personnel position characteristics,

the dimming mapping module obtains a dimming illumination distribution table of each LED string by changing the driving current value of each LED string of the dimming lamp set in a classroom and recording the color temperature value, the illumination value and the illumination uniformity value of each area when each corresponding LED string is combined to illuminate; and establish the photochromic scoring function of various color temperatures and illuminations under different lighting scenes such as self-repair, projection, blackboard writing teaching, discussion, break in class and the like in a classroom; the dimmed lighting profile table and the light color scoring function are also maintained in a memory module,

based on the dimming illumination distribution table, the light color scoring function, the identified illumination scene, and the standard reaching level and the power consumption of the illuminance uniformity of each region of the classroom relative to the reference value, the illumination optimization processing module optimizes the driving current value of each LED string of the dimmable lamp bank within the available value space range through a multi-objective optimization algorithm, and transmits the optimization result to the driver of the corresponding LED string through the dimming mapping module and the output module, and the driver performs dimming by changing the driving current of the LED string.

Preferably, each record of the dimming illumination distribution table includes n-channel LED string driving current values of the dimming lamp set, and color temperature and illuminance values obtained by the corresponding color processing module after signal processing of the test points of the m color sensing modules in the color sensing unit,

the host unit is further configured to:

and sending dimming signals to the dimmable lamp set in a stepping change mode through the dimming mapping module and the output module, detecting the changed light environment, acquiring the color temperature and the illuminance value at the test point, and continuously repeating until the recorded samples cover the value intervals of the driving currents of the LED strings.

In the processing process of the multi-objective optimization algorithm, firstly, initialization is carried out, a strategy for encoding n paths of LED string driving current parameters is determined, and respective value intervals are determined; secondly, aiming at each individual in the evolution group in the search space, searching a dimming illumination distribution table to obtain the color temperature and the illumination value of each test point corresponding to the individual based on the n-path driving current parameter values, and respectively calculating the color temperature scoring value f of the color temperature and the illumination value of each test point obtained by searching according to the color temperature and the illumination color scoring function of the identified illumination scene₁Illuminance score value f₂Calculating the grading value f of the illuminance uniformity according to the standard level of the illuminance uniformity of each region of the classroom relative to the reference value₃Calculating the energy-saving score f according to the power consumption₄And carrying out weighted summation on the 4 scoring values to calculate a total scoring value f-k corresponding to the individual₁·f₁+k₂·f₂+k₃·f₃+k₄·f₄Wherein k is_i(i is 1, 2, 3, 4) is a preset weighting coefficient, and the inheritance, intersection and variation operations are carried out according to the total score value to update the evolutionary population; and then, repeatedly evolving the population until the optimization is finished, and outputting an optimization result. Wherein the weighting coefficient k_iAccording to the setting of the desired effect of illumination, if high uniformity of illumination is desired, the corresponding k can be set₃Increasing; when the color temperature of the light is expected to be more in accordance with the expected interval, the corresponding k can be adjusted₁The other scoring factors are similar.

Preferably, the dimmable light set in the system comprises two LED strings of high color temperature and low color temperature, each LED string corresponds to one driving current channel, the dimming illumination distribution table is a mapping table combining the two-channel current values (i1, i2) to the color temperature and illuminance value of each test point, and in the processing process of the multi-objective optimization algorithm, the color temperature and illuminance value of each test point are obtained by performing multi-dimensional interpolation search in the mapping table for the combination of the two-channel current values (i01, i02) corresponding to each individual in the evolution group.

Preferably, the lamp group adjusts the driving current value of each LED string in the lamp group through a driver, and the optimization result of the multi-objective optimization algorithm is the PWM wave duty ratio value of the driving current of the LED strings.

Preferably, the invention can also be combined with image processing to perform partition management on a classroom, when no person exists in one partition, the target illumination is set to be zero or a maintenance illumination, and the maintenance illumination can be set to be 50-100 Lx.

Preferably, a human body detection sensor can be additionally arranged in each partition to detect the presence information of people.

Example 4

In the present embodiment, a scenic classroom intelligent lighting optimization method is provided.

With reference to fig. 1 and 8, a scene-based classroom intelligent lighting optimization method includes the following steps:

s1, determining the evaluation standard,

in a classroom, assuming that the dimmable light set comprises n LED strings, each LED string corresponds to a driving current, m light color sensing modules are arranged in the light color sensing unit, each light color sensing module corresponds to a desk test point, a sensing signal of each light color sensing module is processed by the light color processing module and respectively obtains a color temperature parameter and an illumination parameter,

establishing a dimming illumination distribution table in a host unit, wherein each record of the dimming illumination distribution table comprises n paths of LED string driving current values of a dimming lamp group and color temperature and luminance values obtained after test point signals of m light color sensing modules in a light color sensing unit are processed by a light color processing module corresponding to the n paths of LED string driving current values; also aiming at different lighting scenes such as self-repair, projection, blackboard writing teaching, discussion, break in class and the like in a classroom, a light color grading function of color temperature and illumination under each scene is established after statistics according to scores of different learners,

aiming at the driving current values of n paths of LED strings of the adjustable light lamp group, an illumination overall evaluation function F of intelligent illumination of a scene type classroom is established,

in the formula, w_iTo set weighting coefficients, f_iFor each factor evaluation value, i is 1, 2, 3, 4, and f is_iThe light source set composed of n paths of LED strings of the adjustable light lamp group and the combination of two light color parameters of total emergent illumination and total emergent color temperature when the environment natural light is mixed for illumination, and the corresponding single factor score value;

wherein f is₁Is the color temperature score value, f₂Is an illuminance score value, f₃Value of illuminance uniformity scale, f₄In order to provide a value of the energy saving rating,

s2, initializing parameters such as evolution population scale, crossover probability, variation probability and the like, and determining the value interval and the coding strategy of the N-path LED string driving current parameters of the dimmable light group and the number N of global Pareto optimal solution replacements in each generation of population_rp；

S3, generating an initial population P (0) randomly for the current drive current set to be optimized, wherein k is 0;

s4, let k equal to k + 1; if the end condition is reached, turning to the S11, otherwise, turning to the next step;

s5, decoding all individuals in the current group P (k-1) to obtain n paths of LED string driving current values, obtaining corresponding color temperature and illumination value through searching a dimming distribution table and multi-dimensional interpolation according to the current values, and respectively calculating color temperature scoring values f based on color temperature and illumination light color scoring functions corresponding to the currently identified lighting scene according to the color temperature and illumination value₁Illuminance score value f_GCalculating the grading value f of the illuminance uniformity according to the standard level of the illuminance uniformity of each region of the classroom relative to the reference value₃Calculating the energy-saving score f according to the power consumption₄According to each f_iAnd comparing to obtain the Pareto optimal solution set PT of the present generation_kAnd updating the global Pareto optimal solution set PT_g；

S6, if PT_kNumber of individuals in set N (PT)_k) For odd numbers, randomly selecting one individual to add to PT_kSet up to match each other and calculate the current generation groupPT_kCollecting the F value of the overall evaluation function of each individual, and selecting other (N (gp) -N (PT) according to the F value of each individual by roulette method_k) B)/2 pairs of fathers; the obtained parent population is P' (k);

s7, carrying out crossover and mutation operations on the individuals in the P '(k) to generate a population P' (k);

s8, for PT in P' (k)_kThe filial individuals of the set are substituted back by the parents if the value of the overall evaluation function F cannot be better than that of the parents, so that a population P' (k) is obtained;

s9, putting the PT in P' (k) not_kN of collection filial generation_rpRandomly replacing the individuals with global Pareto optimal solution individuals to generate a next generation group P (k);

s10, turning to step S4;

and S11, finishing optimization, selecting a solution with the optimal value of the overall evaluation function F based on the finally obtained Pareto optimal solution set, and storing and outputting the optimal solution.

Preferably, the illuminance and the color temperature of the total outgoing light in step S1 are calculated and processed as follows:

t1, converting the color temperature of the light corresponding to the n paths of LED strings and the color temperature of the environmental natural light into xyz color coordinates based on the conversion relation from the color temperature to the color coordinates;

t2, converting color coordinates XYZ of the artificial light source light emission corresponding to the n-path LED strings and the ambient natural light and the brightness corresponding to the independent light emission into XYZ tristimulus values, and adding the two X, Y, Z tristimulus transformation values to obtain total XYZ tristimulus values;

t3, adding two illuminances of the artificial light source and the environment natural light corresponding to the n-path LED strings to obtain total illuminance; meanwhile, converting the total XYZ tristimulus values into total XYZ color coordinates, and further converting the total XYZ tristimulus values into total color temperature; and calculating the grading value of each factor according to the total illumination and the total color temperature.

When the illumination is insufficient in the daytime, the light is supplemented by classroom lamps; ambient light generally forms an ambient light distribution in a classroom by diffusion and reflection, and in this case, the ambient light forms mixed light illumination with the illumination light formed by the lamp groups.

Preferably, the found illumination is scaled to luminance to calculate the tristimulus value.

Preferably, in addition to the LED string driving current parameters, the input quantity of the BP neural network further includes ambient light intensity, and color temperature and illuminance of ambient light at each test point in the classroom, and the parameter configuration of the BP neural network model is modified accordingly.

While the embodiments of the present invention have been described above, these embodiments are presented as examples and do not limit the scope of the invention. These embodiments may be implemented in other various ways, and various omissions, substitutions, combinations, and changes may be made without departing from the spirit of the invention. These embodiments and modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalent scope thereof.

Claims (9)

1. A light modulation mapping module for scene type classroom intelligent lighting control device comprises a mapping input part, a mapping planning part, a mapping determination part, a mapping storage part and a mapping output part, wherein the mapping input part comprises a photochromic parameter input port and a driving current parameter input port,

the light color parameter input port receives color temperature and illumination parameters of multiple test points acquired after sampling processing by the external light color sensing module, the driving current parameter input port receives LED string driving current parameters which correspond to the color temperature and the illumination values and are sent to the external adjustable light bank,

the mapping determination part is provided with a BP neural network, the input quantity of the BP neural network is the LED string driving current parameter, the output quantity is the color temperature and illumination parameter,

the mapping planning part receives the training sample of the BP neural network through the mapping input part, trains the BP neural network by using the sample, and stores the connection weight of the BP neural network after training in the mapping storage module part.

2. The dimming mapping module as claimed in claim 1, wherein the input quantity of the BP neural network is n-way LED string driving current values, the output quantity is 2m parameters of color temperature and illuminance values obtained after the color processing module processes the test point signals of m color sensing modules in the color sensing unit,

the model of the BP neural network is as follows:

the jth node of the hidden layer outputs

The p-th node of the output layer outputs

Wherein the f () function is taken as sigmoid function, w_ijAnd v_jpRespectively the connection weight from the input layer to the hidden layer and the connection weight from the hidden layer to the output layer, a_jAnd b_pAnd (3) respectively representing hidden layer and output layer thresholds, wherein p is 1, 2m, and k is the number of nodes of the hidden layer, and the network training is carried out by adopting a gradient descent method.

3. The dimming mapping module of claim 1, wherein the dimming mapping module is disposed in a host unit of the scenic classroom intelligent lighting control device, the host unit further comprises an input module, a photochromic processing module, an image processing module, a lighting optimization processing module, an output module, and a storage module, and the host unit is configured to:

the input module respectively acquires color temperature and illumination signals, classroom scene images and personnel on-site signals in a preset area from a photochromic sensing unit, an image acquisition unit and a human body detection unit of the system,

based on the color temperature and the illumination signal, the light color processing module acquires a color temperature value, an illumination value and an illumination uniformity value of each region of the classroom,

the image processing module extracts image features and personnel position features based on a scene detector trained by a training image set, and identifies the lighting scene of the image according to the image features and the personnel position features,

the mapping planning part sends a dimming signal to the dimmable lamp set in a step-by-step changing mode in the dimming mapping module, acquires training sample sets of the neural network under different light environments through the light color parameter input port and the driving current parameter input port, trains the BP neural network by using the training sample sets in an off-line mode, adjusts the connection weight of the BP neural network in the mapping determination part, and stores the connection weight in the mapping storage module part;

based on the BP neural network, established light color grading functions of various color temperatures and illuminations under different lighting scenes of self-repair, projection, blackboard writing teaching, discussion and break in a classroom, and standard reaching levels and power consumption of the uniformity of the illuminations of various areas of the classroom relative to reference values, the lighting optimization processing module optimizes the LED strings in an available value space range of the driving current values of the LED strings of the adjustable light banks through a multi-objective optimization algorithm, and transmits optimization results to drivers of the corresponding LED strings through the dimming mapping module and the output module; in the multi-objective optimization processing process, the mapping input part transmits n paths of LED string driving current value combinations corresponding to the individual to be evaluated to the mapping determination part, the BP neural network trained in the mapping input part maps the current value combinations into color temperature and illumination values of all test points, and after the color temperature and illumination values are output through the mapping output part, the illumination optimization processing module calculates the score value of the individual to be evaluated according to a light color scoring function aiming at the color temperature and illumination values of all the test points.

4. The dimming mapping module as claimed in claim 1, wherein in the multi-objective optimization process, the color temperature and the illuminance value of each test point output by the mapping output unit calculate the value of the individual to be evaluated according to the following light color grading function:

for color temperature, when its maximum human expectation value belongs to a medium-high color temperature, its score function is,

when the color temperature is most expected by a plurality of persons to belong to the medium and low color temperatures, the scoring function is,

wherein, W is the current color temperature, bW and cW are the lower limit and the upper limit of the middle and high expected color temperature interval covering the number of people with the set proportion obtained according to statistics in the current scene, aW and dW are the other two preset lower limits and upper limits in the current scene, respectively, and hW is the upper limit of the middle and low expected color temperature interval covering the number of people with the set proportion obtained according to statistics in the current scene.

5. The dimming mapping module as claimed in claim 1, wherein in the multi-objective optimization process, the color temperature and the illuminance value of each test point output by the mapping output unit calculate the value of the individual to be evaluated according to the following light color grading function:

for illumination, the scoring function is,

wherein, E is the current illumination, bE and cE are the lower limit and the upper limit of the interval covering the expected illumination value of the people with the set proportion obtained according to statistics in the current scene, and aE and dE are the other two preset lower limits and upper limits in the current scene respectively.

6. The dimming mapping module as claimed in claim 1, wherein in the multi-objective optimization process, the color temperature and the illuminance value of each test point output by the mapping output unit calculate the value of the individual to be evaluated according to the following light color grading function:

for uniformity of illumination, the scoring function is,

wherein, U is the current illuminance uniformity, bU is a reference value set according to the standard, and aU is a preset lower limit value.

7. The dimming mapping module for the intelligent lighting control device for the scene-based classroom according to claims 4-6, wherein the total score calculation formula corresponding to the individual to be evaluated is:

f＝k₁·f₁+k₂·f₂+k₃·f₃+k₄·f₄，

wherein f is₁Is the color temperature score value, f₂Is an illuminance score value, f₃An illuminance uniformity score calculated from the standard level of illuminance uniformity for each region of the classroom relative to a reference value, f₄For calculating an energy saving score value k based on the power consumed_iAnd (i-1, 2, 3, 4) is a preset weighting coefficient.

8. The dimming mapping module as claimed in claim 1, wherein in the multi-objective optimization process, the color temperature and the illuminance value of each test point output by the mapping output unit calculate the value of the individual to be evaluated according to the following light color grading function:

for a self-repairing scene, an illumination value interval corresponding to the maximum score value of the illumination scoring function is limited to 300-500 Lx, and a color temperature value interval corresponding to the maximum score value of the color temperature scoring function is limited to 4500-6500K;

for a projection scene, from a projection cloth to a classroom back row, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination score function are gradually increased;

for a blackboard writing teaching scene, the upper limit value and the lower limit value of an illumination value interval corresponding to the highest score value of the illumination score function of a blackboard area are larger than those of other areas;

for the discussion scene, the color temperature value range corresponding to the maximum score value of the color temperature scoring function is set according to the middle-low expected color temperature range, and the upper limit value of the color temperature value range is limited below 4000K;

for a break scene in a class, an illumination value range corresponding to the maximum score value of the illumination scoring function is limited to 300-320 Lx, and a color temperature value range corresponding to the maximum score value of the color temperature scoring function is limited to 3300-5000K.

9. The dimming mapping module as claimed in claim 1, wherein the input of the BP neural network further comprises ambient light intensity, and color temperature and illumination of the ambient light at each test point in the classroom.

Images