CN117094230B - Urban lamplight optimization model training system and method - Google Patents
Urban lamplight optimization model training system and method Download PDFInfo
- Publication number
- CN117094230B CN117094230B CN202311340384.7A CN202311340384A CN117094230B CN 117094230 B CN117094230 B CN 117094230B CN 202311340384 A CN202311340384 A CN 202311340384A CN 117094230 B CN117094230 B CN 117094230B
- Authority
- CN
- China
- Prior art keywords
- data
- parameters
- illumination
- lamp
- target area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012549 training Methods 0.000 title claims abstract description 121
- 238000005457 optimization Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000009877 rendering Methods 0.000 claims abstract description 26
- 238000005286 illumination Methods 0.000 claims description 113
- 239000013598 vector Substances 0.000 claims description 38
- 238000009434 installation Methods 0.000 claims description 28
- 238000012544 monitoring process Methods 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 26
- 238000005070 sampling Methods 0.000 claims description 24
- 239000000463 material Substances 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000036544 posture Effects 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 description 14
- 238000012795 verification Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000009471 action Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/50—Lighting effects
- G06T15/506—Illumination models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/35—Categorising the entire scene, e.g. birthday party or wedding scene
- G06V20/38—Outdoor scenes
- G06V20/39—Urban scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Computer Graphics (AREA)
- Circuit Arrangement For Electric Light Sources In General (AREA)
Abstract
The invention provides a training system and a training method for an urban light optimization model, which are characterized in that a candidate parameter sequence comprising lighting parameters and lamp layout parameters is generated by acquiring spatial layout data and a spatial three-dimensional model of a target area, the lighting parameter range and the lamp layout parameter range of a lighting lamp, and is respectively input into a light rendering engine to simulate generation of brightness distribution data under the corresponding lighting parameters and lamp layout parameters in the spatial three-dimensional model, lighting scores of the brightness distribution data corresponding to each element in the candidate parameter sequence are calculated, elements with the lighting scores larger than a preset scoring threshold in the candidate parameter sequence are determined as target elements, training sample data is generated by using the lighting parameters, the lamp layout parameters and the spatial layout data corresponding to the target elements, and the urban light optimization model is trained by using the training sample data, so that the urban light optimization model for outputting the urban light optimization parameters can be trained efficiently.
Description
Technical Field
The invention relates to the technical field of urban illumination, in particular to an urban light optimization model training system and method.
Background
The urban public lighting facilities are important components of urban lights and are also one of infrastructure of urban public construction, and with the advancement of urban construction, the public lighting facilities of all cities are more and more, so that necessary safety guarantee is provided for people going out at night. Although the basic function of urban lights is to provide night lighting, the lighting fixtures are not more densely arranged and the brightness of the lights is not as bright as possible. On one hand, the deployment cost of dense illumination facilities is high, the overall energy consumption of huge quantity of urban illumination facilities is quite remarkable, and excessive illumination can cause larger energy waste; on the other hand, the light pollution can also cause harm to the ecological environment, and the balance of the biological rhythm and the ecological system is destroyed. Conversely, insufficient illumination includes an excessively small number of arrangements of illumination facilities, or insufficient brightness of illumination facilities, or the like, which affects safety of pedestrians and vehicles. Therefore, how to balance energy consumption, ecological environment and night lighting needs by reasonably arranging urban lighting facilities is a problem to be solved.
Disclosure of Invention
Based on the problems, the invention provides a city light optimization model training system and a city light optimization model training method, which can be used for efficiently training a city light optimization model for outputting city light optimization parameters.
In view of this, a first aspect of the present invention proposes an urban lighting optimization model training system comprising a monitoring image acquisition unit connected to an urban public monitoring system for acquiring urban public area monitoring images, a luminaire parameter acquisition unit connected to a database for acquiring lighting parameters and luminaire layout parameters of luminaires, a light rendering unit connected to a light rendering system for generating luminance distribution data by a light rendering engine in the light rendering system, and a model training unit configured to:
acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
Inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
and training the city light optimization model by using the training sample data.
The second aspect of the invention provides a training method for an urban light optimization model, which comprises the following steps:
acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
Inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
and training the city light optimization model by using the training sample data.
Further, in the above training method of the city light optimization model, the step of obtaining the spatial layout data and the spatial stereoscopic model of the target area specifically includes:
acquiring a monitoring image of the target area through a public monitoring system;
identifying and acquiring spatial layout data of the target area from the monitoring image, wherein the spatial layout data comprises types, positions, shapes, postures, materials and color data of objects in the target area;
Generating an object model in the target region based on the spatial layout data;
and integrating the object models to generate a spatial stereoscopic model of the target area.
Further, in the above training method of the city light optimization model, the step of generating a preset number of candidate parameter sequences in the lighting parameter range and the lamp layout parameter range for each target area specifically includes:
determination ofA plurality of candidate lighting fixtures;
obtaining lighting parameters of the candidate lighting fixtures from a databasePreconfigured +.>Group lamp layout parameters->Wherein->；
The illumination parameters are setThe lamp layout parameters +.>Combining two by two to generate a product comprising +>Candidate parameter sequences of individual elements, each element in the candidate parameter sequences being。
Further, in the above training method of the city light optimization model, the step of calculating the illumination score of the brightness distribution data corresponding to each element in the candidate parameter sequence specifically includes:
sampling from each object surface of the target area at a preset sampling distance to obtain brightness values of sampling pixel pointsWherein->，/>For the number of objects in the target area, < > x- >Is->The number of pixel points obtained by sampling the surface of each object;
based on the brightness valueCalculating each object by deviation of optimal surface brightness value relative to the same type of object and scoring coefficientIntensity score ∈>；
Taking a brightness score for each object in the target areaAs an illumination score of the brightness distribution data: />。
Further, in the city light optimization model training method, the luminance value is based onCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>The method specifically comprises the following steps:
obtaining an optimal surface brightness value for each object in the target areaScoring coefficient->；
Calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness valueIs the first deviation of (2):
；
based on the first deviationCalculating a brightness score for each object:
。
further, in the city light optimization model training method, the luminance value is based onCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>The method specifically comprises the following steps:
Obtaining an optimal surface brightness value for each object in the target areaFirst scoring coefficient->And a second scoring coefficient->The first scoring coefficient->Luminance value for pixel at object surface +.>Greater than said optimal surface brightness value +.>Score calculation at the time, the second score coefficient +.>Luminance value for pixel at object surfaceLess than the optimal surfaceLuminance value->Scoring calculation at the time;
calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness valueIs the second deviation of (2):
；
based on the second deviationCalculating a brightness score for each object:
。
further, in the above training method of the city light optimization model, the step of generating training sample data by using the lighting parameters, the lamp layout parameters and the spatial layout data corresponding to the target elements specifically includes:
determining the lighting parameter, the lamp layout parameter and the space layout data corresponding to a target element as a data record;
the following processing steps are performed for each data record:
generating an input feature vector based on the spatial layout data of the data record;
generating an output tag vector based on the lighting parameters and the lamp layout parameters of the data record;
And storing the input characteristic vector and the output label vector in association as an input sample and an output sample in the same training sample data.
Further, in the training method of the city light optimization model, the step of generating the input feature vector based on the spatial layout data of the data record specifically includes:
extracting the attributes such as object types, quantity, positions, shapes, materials and the like in the space layout data as original features;
performing standardization processing on the numerical value type data in the original characteristics to obtain first original characteristic data;
performing quantization processing on the non-numerical type data in the original features to obtain second original feature data;
and combining the first original characteristic data and the second original characteristic data to obtain the input characteristic vector.
Further, in the training method of the city light optimization model, the step of generating the output tag vector based on the lighting parameters and the lamp layout parameters recorded by the data specifically includes:
sequentially extracting each data from the lighting parameters and the lamp layout parameters;
for the data which is the numerical value type, performing standardization processing on the data;
Determining the standardized numerical value as a corresponding label value;
for non-numeric types, mapping the non-numeric types into a pre-built dictionary to obtain the single-hot codes of the non-numeric types;
determining the one-time thermal code as a corresponding tag value;
and constructing the tag value into the tag vector according to the sequence of the corresponding data in the data record.
The invention provides a training system and a training method for an urban light optimization model, which are characterized in that a candidate parameter sequence comprising lighting parameters and lamp layout parameters is generated by acquiring spatial layout data and a spatial three-dimensional model of a target area, the lighting parameter range and the lamp layout parameter range of a lighting lamp, and is respectively input into a light rendering engine to simulate generation of brightness distribution data under the corresponding lighting parameters and lamp layout parameters in the spatial three-dimensional model, lighting scores of the brightness distribution data corresponding to each element in the candidate parameter sequence are calculated, elements with the lighting scores larger than a preset scoring threshold in the candidate parameter sequence are determined as target elements, training sample data is generated by using the lighting parameters, the lamp layout parameters and the spatial layout data corresponding to the target elements, and the urban light optimization model is trained by using the training sample data, so that the urban light optimization model for outputting the urban light optimization parameters can be trained efficiently.
Drawings
FIG. 1 is a schematic diagram of an urban lighting optimization model training system according to an embodiment of the present invention;
fig. 2 is a flowchart of a training method for an optimization model of urban lighting according to an embodiment of the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced otherwise than as described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
In the description of the present invention, the term "plurality" means two or more, unless explicitly defined otherwise, the orientation or positional relationship indicated by the terms "upper", "lower", etc. are based on the orientation or positional relationship shown in the drawings, merely for convenience of description of the present invention and to simplify the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. The terms "coupled," "mounted," "secured," and the like are to be construed broadly, and may be fixedly coupled, detachably coupled, or integrally connected, for example; can be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first", "a second", etc. may explicitly or implicitly include one or more such feature. In the description of the present invention, unless otherwise indicated, the meaning of "a plurality" is two or more.
In the description of this specification, the terms "one embodiment," "some implementations," "particular embodiments," and the like, mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
An urban lighting optimization model training system and method provided according to some embodiments of the present invention are described below with reference to the accompanying drawings.
As shown in fig. 1, a first aspect of the present invention proposes an urban lighting optimization model training system, which includes a monitoring image acquisition unit connected to an urban public monitoring system for acquiring monitoring images of an urban public area, a lighting parameter acquisition unit connected to a database for acquiring lighting parameters and lighting layout parameters of a lighting fixture, a light rendering unit connected to a light rendering system for generating luminance distribution data by a light rendering engine in the light rendering system, and a model training unit. As shown in fig. 2, the model training unit is configured to:
Acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
Generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
and training the city light optimization model by using the training sample data.
Specifically, before the step of obtaining the spatial layout data and the spatial stereoscopic model of the target area, the method further includes dividing the public area of the city into a plurality of target areas with a preset size, for example, the public area may be divided by taking one or a plurality of roads as a division unit or taking artificial or natural objects such as bridges, rivers or mountain as boundaries.
The lighting fixtures are a plurality of lighting fixtures used for urban public lighting and preset in a database, the lighting fixtures are distinguished according to the type, the brand, the model and the like of the lighting fixtures, and the lighting parameters of each candidate lighting fixture, including the power, the brightness, the color, the working voltage, the working current, the beam angle and the like, are stored in the database. The illumination parameter range of the candidate illumination lamp is a range formed by the numerical values corresponding to each illumination parameter of all the candidate illumination lamps in the database. The number of the installed lighting fixtures and the installation distance of the lighting fixtures in the target area can be obtained through the installation positions of the lighting fixtures.
The brightness distribution data is a set of brightness data at each pixel point on the surface of each object in the target area when the lighting fixtures corresponding to the lighting parameters are installed according to the corresponding fixture layout parameters.
Further, in the above city light optimizing model training system, in the step of acquiring the spatial layout data and the spatial stereoscopic model of the target area, the model training unit is configured to:
acquiring a monitoring image of the target area through a public monitoring system;
identifying and acquiring spatial layout data of the target area from the monitoring image, wherein the spatial layout data comprises types, positions, shapes, postures, materials and color data of objects in the target area;
generating an object model in the target region based on the spatial layout data;
and integrating the object models to generate a spatial stereoscopic model of the target area.
In some embodiments of the present invention, after the step of integrating the object model to generate the spatial stereo model of the target area, it is further periodically determined by the monitoring image whether the spatial environment of the target area changes, and when the spatial environment of the target area changes significantly, the spatial layout data and the spatial stereo model are updated.
Further, in the above city light optimizing model training system, in the step of generating a preset number of candidate parameter sequences within the illumination parameter range and the lamp layout parameter range for each target area, the model training unit is configured to:
determination ofA plurality of candidate lighting fixtures;
obtaining lighting parameters of the candidate lighting fixtures from a databasePreconfigured +.>Group lamp layout parameters->Wherein->；
The illumination parameters are setThe lamp layout parameters +.>Combining two by two to generate a product comprising +>Candidate parameter sequences of individual elements, each element in the candidate parameter sequences being。
In particular, the lighting parametersThe lamp layout parameters +.>In each element in the form of key-value pairs, for example: { brightness: 1000lm, color: 5000K, beam angle: 120 degrees }.
Further, in the above city light optimizing model training system, in the step of calculating the illumination score of the luminance distribution data corresponding to each element in the candidate parameter sequence, the model training unit is configured to:
sampling from each object surface of the target area at a preset sampling distance to obtain brightness values of sampling pixel points Wherein->，/>For the number of objects in the target area, < > x->Is->The number of pixel points obtained by sampling the surface of each object;
based on the brightness valueCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>；
Taking a brightness score for each object in the target areaAs an illumination score of the brightness distribution data: />。
Specifically, the sampling distance is a pixel distance in the spatial stereo model, and the size of the sampling distance is configured according to different precision of the spatial stereo model.
Further, in the city light optimizing model training system, the luminance value is based onCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>In the step (a), the model training unit is configured to:
obtaining an optimal surface brightness value for each object in the target areaScoring coefficient->；
Calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness valueIs the first deviation of (2):
；
based on the first deviationCalculating a brightness score for each object:
。
Specifically, the optimal surface brightness valueAnd the scoring coefficient->For values of each object type pre-configured and stored in a database, which are associated with classes of objectsType information is stored in association with each other, and optimal surface brightness of each object in the target area is obtained>Scoring coefficient->In the step (a), the type of each object in the target area is firstly identified, and then the corresponding optimal surface brightness is read from the database according to the type>Scoring coefficient->. In the technical solution of the foregoing embodiment, different scoring coefficients are configured for each different type of object in the target area, so that the scoring result may be adjusted more finely according to the type of the object, for example, for a road surface made of different materials, the road surface is identified as a different type of the object, and different scoring coefficients are configured to reflect the reflective capability of the material of the road surface made of different materials on the light in the scoring result. It should be appreciated that the same scoring coefficients may be used for different object types in order to save the effort of scoring coefficient configuration.
Further, in the city light optimizing model training system, the luminance value is based on Calculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>In the step (a), the model training unit is configured to:
obtaining an optimal surface brightness value for each object in the target areaFirst scoring coefficient->And a second scoring coefficient->The first scoring coefficient->Luminance value for pixel at object surface +.>Greater than said optimal surface brightness value +.>Score calculation at the time, the second score coefficient +.>Luminance value for pixel at object surfaceLess than said optimal surface brightness value +.>Scoring calculation at the time;
calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness value +.>Is the second deviation of (2):
；
based on the second deviationCalculating a brightness score for each object:
。
because the illumination lamp is too bright and too dark to influence the degree different, in the technical scheme of the above-mentioned implementation mode, use different scoring coefficients to score to too bright and too dark condition respectively.
Further, in the city light optimizing model training system, in the step of generating training sample data using the lighting parameters, the lamp layout parameters, and the spatial layout data corresponding to the target elements, the model training unit is configured to:
Determining the lighting parameter, the lamp layout parameter and the space layout data corresponding to a target element as a data record;
the following processing steps are performed for each data record:
generating an input feature vector based on the spatial layout data of the data record;
generating an output tag vector based on the lighting parameters and the lamp layout parameters of the data record;
and storing the input characteristic vector and the output label vector in association as an input sample and an output sample in the same training sample data.
In the technical scheme of the embodiment, the spatial layout data are processed into input sample data, and the lighting parameters and the lamp layout parameters are processed into output sample data, so that the urban lighting optimization model trained based on the sample data can output the optimal lighting parameters and lamp layout parameters when the spatial layout data of a certain area are input, and a constructor can optimize urban lighting configuration through the lighting parameters and the lamp layout parameters.
Further, in the city light optimization model training system, in the step of generating the input feature vector based on the spatial layout data of the data record, the model training unit is configured to:
Extracting the attributes such as object types, quantity, positions, shapes, materials and the like in the space layout data as original features;
performing standardization processing on the numerical value type data in the original characteristics to obtain first original characteristic data;
performing quantization processing on the non-numerical type data in the original features to obtain second original feature data;
and combining the first original characteristic data and the second original characteristic data to obtain the input characteristic vector.
Specifically, for the numerical value type data in the original feature, such as the number of objects, the numerical values can be directly used for carrying out normalization processing to obtain first original feature data, wherein the normalization processing comprises unit normalization, numerical value range normalization, value accuracy normalization and the like, and the position of the object can be processed by adopting x, y and z coordinates in a specific coordinate system or longitude and latitude in GPS positioning coordinates and the like to obtain the first original feature data.
For non-numerical type data in the original feature, such as the type of the object, the shape of the object, the material of the object, etc., it may be quantized in a one-hot (one-hot) coding manner, that is, a pre-built dictionary is searched for one-hot codes corresponding to the numerical type data, and the codes are used as the second original feature data. The method comprises the steps of extracting feature description shapes such as outline features, polygonal approximation and the like for the shape of an object, and converting the feature description shapes into corresponding independent thermal codes. Likewise, for the material of the object, the characteristics of the object such as color, texture, reflection characteristic and the like can be extracted and converted into corresponding independent thermal codes. Further, in the technical solutions of some embodiments of the present invention, the spatial layout data further includes a posture feature of the non-stationary object.
Further, in the city light optimizing model training system, in the step of generating the output tag vector based on the lighting parameters and the lamp layout parameters of the data record, the model training unit is configured to:
sequentially extracting each data from the lighting parameters and the lamp layout parameters;
for the data which is the numerical value type, performing standardization processing on the data;
determining the standardized numerical value as a corresponding label value;
for non-numeric types, mapping the non-numeric types into a pre-built dictionary to obtain the single-hot codes of the non-numeric types;
determining the one-time thermal code as a corresponding tag value;
and constructing the tag value into the tag vector according to the sequence of the corresponding data in the data record.
Specifically, for each data record, data such as brightness (unit: lumen), color (unit: kelvin), beam angle (unit: degree), position (x, y, z coordinates in a specific coordinate system, longitude and latitude in GPS positioning coordinates, or the like), mounting height (unit: meter), irradiation angle (pitch angle and azimuth angle, unit: degree), and the like, which are required as a label, are extracted, and for irradiation angle, sine and cosine thereof may be extracted as labels or converted into a representation of similar spherical coordinates, and for data of which unit is inconsistent, standardized processing is performed, that is, converted into standard unit values. In addition, the normalization processing is performed on the data, and the normalization processing further includes unifying the numerical range, unifying the numerical accuracy of the decimal, and the like. For non-quantized data, such as luminaire type, it can be converted into a single thermal code as its tag. The quantized tag values are constructed into tag vectors in a fixed order.
In some embodiments of the present invention, a fully connected neural network may be selected as a training model, where the number of nodes at the input layer is the same as the length of the feature vector, and the number of nodes at the output layer is the same as the length of the tag vector. Of course, other supervision learning models such as support vector machines or random forests can also be selected. In the step of training the city light optimization model using the training sample data, the model training unit is configured to:
dividing the training sample data into a training data set and a verification data set, wherein the training data set is used for training to adjust parameters of the urban light optimization model, and the verification data set is used for detecting overfitting and evaluating generalization capability of the urban light optimization model;
configuring training parameters, wherein the training parameters comprise iteration times, batch size, learning rate, loss function and optimizer;
inputting the training data set for iterative training;
inputting verification data of the verification data set to obtain a loss function value of the city lamplight optimization model after every preset iteration step number;
and stopping inputting the training data set to perform iterative training when the loss function value is no longer reduced.
As shown in fig. 2, a second aspect of the present invention proposes a training method for an optimization model of urban lighting, including:
acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
Determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
and training the city light optimization model by using the training sample data.
Specifically, before the step of obtaining the spatial layout data and the spatial stereoscopic model of the target area, the method further includes dividing the public area of the city into a plurality of target areas with a preset size, for example, the public area may be divided by taking one or a plurality of roads as a division unit or taking artificial or natural objects such as bridges, rivers or mountain as boundaries.
The lighting fixtures are a plurality of lighting fixtures used for urban public lighting and preset in a database, the lighting fixtures are distinguished according to the type, the brand, the model and the like of the lighting fixtures, and the lighting parameters of each candidate lighting fixture, including the power, the brightness, the color, the working voltage, the working current, the beam angle and the like, are stored in the database. The illumination parameter range of the candidate illumination lamp is a range formed by the numerical values corresponding to each illumination parameter of all the candidate illumination lamps in the database. The number of the installed lighting fixtures and the installation distance of the lighting fixtures in the target area can be obtained through the installation positions of the lighting fixtures.
The brightness distribution data is a set of brightness data at each pixel point on the surface of each object in the target area when the lighting fixtures corresponding to the lighting parameters are installed according to the corresponding fixture layout parameters.
Further, in the above training method of the city light optimization model, the step of obtaining the spatial layout data and the spatial stereoscopic model of the target area specifically includes:
acquiring a monitoring image of the target area through a public monitoring system;
identifying and acquiring spatial layout data of the target area from the monitoring image, wherein the spatial layout data comprises types, positions, shapes, postures, materials and color data of objects in the target area;
generating an object model in the target region based on the spatial layout data;
and integrating the object models to generate a spatial stereoscopic model of the target area.
In some embodiments of the present invention, after the step of integrating the object model to generate the spatial stereo model of the target area, it is further periodically determined by the monitoring image whether the spatial environment of the target area changes, and when the spatial environment of the target area changes significantly, the spatial layout data and the spatial stereo model are updated.
Further, in the above training method of the city light optimization model, the step of generating a preset number of candidate parameter sequences in the lighting parameter range and the lamp layout parameter range for each target area specifically includes:
determination ofA plurality of candidate lighting fixtures;
obtaining lighting parameters of the candidate lighting fixtures from a databasePreconfigured +.>Group lamp layout parameters->Wherein->；
The illumination parameters are setThe lamp layout parameters +.>Combining two by two to generate a product comprising +>Candidate parameter sequences of individual elements, each element in the candidate parameter sequences being。
In particular, the lighting parametersThe lamp layout parameters +.>In each element in the form of key-value pairs, for example: { brightness: 1000lm, color: 5000K, beam angle: 120 degrees }.
Further, in the above training method of the city light optimization model, the step of calculating the illumination score of the brightness distribution data corresponding to each element in the candidate parameter sequence specifically includes:
sampling from each object surface of the target area at a preset sampling distance to obtain brightness values of sampling pixel pointsWherein->，/>For the number of objects in the target area, < > x- >Is->The number of pixel points obtained by sampling the surface of each object;
based on the brightness valueCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>；
Taking a brightness score for each object in the target areaAs an illumination score of the brightness distribution data: />。
Specifically, the sampling distance is a pixel distance in the spatial stereo model, and the size of the sampling distance is configured according to different precision of the spatial stereo model.
Further, in the city light optimization model training method, the luminance value is based onCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>The method specifically comprises the following steps:
obtaining an optimal surface brightness value for each object in the target areaScoring coefficient->；
Calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness valueIs the first deviation of (2):
；
based on the first deviationCalculating a brightness score for each object:
。
specifically, the optimal surface brightness valueAnd the scoring coefficient- >Values corresponding to each object type, which are pre-configured and stored in a database, are stored in association with type information of the object, and optimal surface brightness +/of each object in the target area is obtained>Scoring coefficient->In the step (a), the type of each object in the target area is firstly identified, and then the corresponding optimal surface brightness is read from the database according to the type>Scoring coefficient->. In the technical solution of the foregoing embodiment, different scoring coefficients are respectively configured for each different type of object in the target area, so that the scoring result may be more finely adjusted according to the object type, for example, for a road surface made of different materials, the road surface is identified as a different object type and different scoring coefficients are configured to be used in the target areaThe reflecting capacity of the road surface materials with different materials to the light is reflected in the grading result. It should be appreciated that the same scoring coefficients may be used for different object types in order to save the effort of scoring coefficient configuration.
Further, in the city light optimization model training method, the luminance value is based onCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object >The method specifically comprises the following steps:
obtaining an optimal surface brightness value for each object in the target areaFirst scoring coefficient->And a second scoring coefficient->The first scoring coefficient->Luminance value for pixel at object surface +.>Greater than said optimal surface brightness value +.>Score calculation at the time, the second score coefficient +.>Luminance value for pixel at object surfaceLess than said optimal surface brightness value +.>Scoring calculation at the time;
calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness valueIs the second deviation of (2):
；
based on the second deviationCalculating a brightness score for each object:
。
because the illumination lamp is too bright and too dark to influence the degree different, in the technical scheme of the above-mentioned implementation mode, use different scoring coefficients to score to too bright and too dark condition respectively.
Further, in the above training method of the city light optimization model, the step of generating training sample data by using the lighting parameters, the lamp layout parameters and the spatial layout data corresponding to the target elements specifically includes:
determining the lighting parameter, the lamp layout parameter and the space layout data corresponding to a target element as a data record;
The following processing steps are performed for each data record:
generating an input feature vector based on the spatial layout data of the data record;
generating an output tag vector based on the lighting parameters and the lamp layout parameters of the data record;
and storing the input characteristic vector and the output label vector in association as an input sample and an output sample in the same training sample data.
In the technical scheme of the embodiment, the spatial layout data are processed into input sample data, and the lighting parameters and the lamp layout parameters are processed into output sample data, so that the urban lighting optimization model trained based on the sample data can output the optimal lighting parameters and lamp layout parameters when the spatial layout data of a certain area are input, and a constructor can optimize urban lighting configuration through the lighting parameters and the lamp layout parameters.
Further, in the training method of the city light optimization model, the step of generating the input feature vector based on the spatial layout data of the data record specifically includes:
extracting the attributes such as object types, quantity, positions, shapes, materials and the like in the space layout data as original features;
Performing standardization processing on the numerical value type data in the original characteristics to obtain first original characteristic data;
performing quantization processing on the non-numerical type data in the original features to obtain second original feature data;
and combining the first original characteristic data and the second original characteristic data to obtain the input characteristic vector.
Specifically, for the numerical value type data in the original feature, such as the number of objects, the numerical values can be directly used for carrying out normalization processing to obtain first original feature data, wherein the normalization processing comprises unit normalization, numerical value range normalization, value accuracy normalization and the like, and the position of the object can be processed by adopting x, y and z coordinates in a specific coordinate system or longitude and latitude in GPS positioning coordinates and the like to obtain the first original feature data.
For non-numerical type data in the original feature, such as the type of the object, the shape of the object, the material of the object, etc., it may be quantized in a one-hot (one-hot) coding manner, that is, a pre-built dictionary is searched for one-hot codes corresponding to the numerical type data, and the codes are used as the second original feature data. The method comprises the steps of extracting feature description shapes such as outline features, polygonal approximation and the like for the shape of an object, and converting the feature description shapes into corresponding independent thermal codes. Likewise, for the material of the object, the characteristics of the object such as color, texture, reflection characteristic and the like can be extracted and converted into corresponding independent thermal codes. Further, in the technical solutions of some embodiments of the present invention, the spatial layout data further includes a posture feature of the non-stationary object.
Further, in the training method of the city light optimization model, the step of generating the output tag vector based on the lighting parameters and the lamp layout parameters recorded by the data specifically includes:
sequentially extracting each data from the lighting parameters and the lamp layout parameters;
for the data which is the numerical value type, performing standardization processing on the data;
determining the standardized numerical value as a corresponding label value;
for non-numeric types, mapping the non-numeric types into a pre-built dictionary to obtain the single-hot codes of the non-numeric types;
determining the one-time thermal code as a corresponding tag value;
and constructing the tag value into the tag vector according to the sequence of the corresponding data in the data record.
Specifically, for each data record, data such as brightness (unit: lumen), color (unit: kelvin), beam angle (unit: degree), position (x, y, z coordinates in a specific coordinate system, longitude and latitude in GPS positioning coordinates, or the like), mounting height (unit: meter), irradiation angle (pitch angle and azimuth angle, unit: degree), and the like, which are required as a label, are extracted, and for irradiation angle, sine and cosine thereof may be extracted as labels or converted into a representation of similar spherical coordinates, and for data of which unit is inconsistent, standardized processing is performed, that is, converted into standard unit values. In addition, the normalization processing is performed on the data, and the normalization processing further includes unifying the numerical range, unifying the numerical accuracy of the decimal, and the like. For non-quantized data, such as luminaire type, it can be converted into a single thermal code as its tag. The quantized tag values are constructed into tag vectors in a fixed order.
In some embodiments of the present invention, a fully connected neural network may be selected as a training model, where the number of nodes at the input layer is the same as the length of the feature vector, and the number of nodes at the output layer is the same as the length of the tag vector. Of course, other supervision learning models such as support vector machines or random forests can also be selected. The step of training the city light optimizing model by using the training sample data specifically comprises the following steps:
dividing the training sample data into a training data set and a verification data set, wherein the training data set is used for training to adjust parameters of the urban light optimization model, and the verification data set is used for detecting overfitting and evaluating generalization capability of the urban light optimization model;
configuring training parameters, wherein the training parameters comprise iteration times, batch size, learning rate, loss function and optimizer;
inputting the training data set for iterative training;
inputting verification data of the verification data set to obtain a loss function value of the city lamplight optimization model after every preset iteration step number;
and stopping inputting the training data set to perform iterative training when the loss function value is no longer reduced.
It should be noted that in this document relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Embodiments in accordance with the present invention, as described above, are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention and various modifications as are suited to the particular use contemplated. The invention is limited only by the claims and the full scope and equivalents thereof.
Claims (5)
1. An urban lighting optimization model training system, comprising a monitoring image acquisition unit connected with an urban public monitoring system and used for acquiring monitoring images of an urban public area, a lamp parameter acquisition unit connected with a database and used for acquiring lighting parameters and lamp layout parameters of lamps, a light rendering unit connected with a light rendering system and used for generating brightness distribution data through a light rendering engine in the light rendering system, and a model training unit, wherein the model training unit is configured to:
acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
Inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
training the city light optimization model by using the training sample data;
in the step of acquiring spatial layout data and a spatial stereoscopic model of the target region, the model training unit is configured to:
acquiring a monitoring image of the target area through a public monitoring system;
identifying and acquiring spatial layout data of the target area from the monitoring image, wherein the spatial layout data comprises types, positions, shapes, postures, materials and color data of objects in the target area;
Generating an object model in the target region based on the spatial layout data;
integrating the object models to generate a spatial stereoscopic model of the target area;
in the step of generating a preset number of candidate parameter sequences within the illumination parameter range and the luminaire layout parameter range for each target area, the model training unit is configured to:
determination ofCandidate lightingA lamp;
obtaining lighting parameters of the candidate lighting fixtures from a databasePreconfigured +.>Group lamp layout parameters->Wherein->；
The illumination parameters are setThe lamp layout parameters +.>Combining two by two to generate a product comprisingCandidate parameter sequences of individual elements, each element of the candidate parameter sequences being +.>；
In the step of calculating an illumination score of the luminance distribution data corresponding to each element in the candidate parameter sequence, the model training unit is configured to:
sampling from each object surface of the target area at a preset sampling distance to obtain brightness values of sampling pixel pointsWherein->,/>For the number of objects in the target area, < > x->Is->The number of pixel points obtained by sampling the surface of each object;
Luminance value based on pixel pointCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>；
Taking a brightness score for each object in the target areaAs an illumination score of the brightness distribution data:
；
at pixel-based luminance valuesCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>In the step (a), the model training unit is configured to:
obtaining an optimal surface brightness value for each object in the target areaScoring coefficient->；
Calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness value +.>Is the first deviation of (2):
；
based on the first deviationCalculating a brightness score for each object:
；
at pixel-based luminance valuesCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>In the step (a), the model training unit is configured to:
obtaining an optimal surface brightness value for each object in the target areaFirst scoring coefficient->And a second scoring coefficient->The first scoring coefficient- >Luminance value for pixel at object surface +.>Greater than said optimal surface brightness value +.>Score calculation at the time, the second score coefficient +.>Luminance value for pixel at object surface +.>Less than said optimal surface brightness value +.>Scoring calculation at the time;
calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness value +.>Is the second deviation of (2):
；
based on the second deviationCalculating a luminance score for each object：
。
2. The city lamplight optimization model training method is characterized by comprising the following steps of:
acquiring space layout data and a space stereoscopic model of a target area, wherein the target area is an illumination area of city lamplight;
acquiring an illumination parameter range of an illumination lamp and a preset lamp layout parameter range, wherein the lamp layout parameters comprise the installation position and the installation height of the illumination lamp;
generating a preset number of candidate parameter sequences in the illumination parameter range and the lamp layout parameter range for each target area, wherein each element in the candidate parameter sequences comprises a group of illumination parameters and a group of lamp layout parameters, the illumination parameters comprise the brightness, the color and the beam angle of the illumination lamp, and the lamp layout parameters comprise the installation position, the installation height and the illumination angle of the illumination lamp;
Inputting each element in the candidate parameter sequence to a ray rendering engine respectively, so that the ray rendering engine simulates and generates brightness distribution data of a target area under corresponding illumination parameters and lamp layout parameters in a space stereoscopic model of the corresponding target area;
calculating illumination scores of brightness distribution data corresponding to each element in the candidate parameter sequence;
determining elements with illumination scores greater than a preset scoring threshold value in the candidate parameter sequence as target elements;
generating training sample data by using the lighting parameters, the lamp layout parameters and the space layout data corresponding to the target elements;
training the city light optimization model by using the training sample data;
the step of obtaining the spatial layout data and the spatial stereoscopic model of the target area specifically comprises the following steps:
acquiring a monitoring image of the target area through a public monitoring system;
identifying and acquiring spatial layout data of the target area from the monitoring image, wherein the spatial layout data comprises types, positions, shapes, postures, materials and color data of objects in the target area;
generating an object model in the target region based on the spatial layout data;
Integrating the object models to generate a spatial stereoscopic model of the target area;
the step of generating a preset number of candidate parameter sequences within the illumination parameter range and the lamp layout parameter range for each target area specifically comprises the following steps:
determination ofA plurality of candidate lighting fixtures;
obtaining lighting parameters of the candidate lighting fixtures from a databasePreconfigured +.>Group lamp layout parameters->Wherein->；
The illumination parameters are setThe lamp layout parameters +.>Combining two by two to generate a product comprisingCandidate parameter sequences of individual elements, each element of the candidate parameter sequences being +.>；
The step of calculating the illumination score of the brightness distribution data corresponding to each element in the candidate parameter sequence specifically comprises the following steps:
sampling from each object surface of the target area at a preset sampling distance to obtain brightness values of sampling pixel pointsWherein->,/>For the number of objects in the target area, < > x->Is->The number of pixel points obtained by sampling the surface of each object;
luminance value based on pixel pointCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object >；
Taking a brightness score for each object in the target areaAs said average value of (2)Illumination score of luminance distribution data:
；
luminance value based on pixel pointCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>The method specifically comprises the following steps:
obtaining an optimal surface brightness value for each object in the target areaScoring coefficient->；
Calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness value +.>Is the first deviation of (2):
；
based on the first deviationCalculating a brightness score for each object:
；
luminance value based on pixel pointCalculating a luminance score +.for each object with respect to a deviation of optimal surface luminance values for the same type of object>The method specifically comprises the following steps:
obtaining an optimal surface brightness value for each object in the target areaFirst scoring coefficient->And a second scoring coefficient->The first scoring coefficient->Luminance value for pixel at object surface +.>Greater than said optimal surface brightness value +.>Score calculation at the time, the second score coefficient +.>Luminance value for pixel at object surface +.>Less than said optimal surface brightness value +. >Scoring calculation at the time;
calculating the brightness value of each pixel point of each object surfaceRelative to the optimal surface brightness value +.>Is the second deviation of (2):
；
based on the second deviationCalculating a brightness score for each object:
。
3. the city light optimizing model training method of claim 2, wherein the step of generating training sample data using the lighting parameters, the lamp layout parameters, and the spatial layout data corresponding to the target elements specifically comprises:
determining the lighting parameter, the lamp layout parameter and the space layout data corresponding to a target element as a data record;
the following processing steps are performed for each data record:
generating an input feature vector based on the spatial layout data of the data record;
generating an output tag vector based on the lighting parameters and the lamp layout parameters of the data record;
and storing the input characteristic vector and the output label vector in association as an input sample and an output sample in the same training sample data.
4. A method of training an urban lighting optimization model according to claim 3, wherein the step of generating the input feature vector based on the spatial layout data of the data record comprises:
Extracting object types, quantity, positions, shapes and material properties in the space layout data as original features;
performing standardization processing on the numerical value type data in the original characteristics to obtain first original characteristic data;
performing quantization processing on the non-numerical type data in the original features to obtain second original feature data;
and combining the first original characteristic data and the second original characteristic data to obtain the input characteristic vector.
5. The city light optimization model training method of claim 3, wherein the step of generating the output tag vector based on the recorded lighting parameters and the lamp layout parameters comprises:
sequentially extracting each data from the lighting parameters and the lamp layout parameters;
for the data which is the numerical value type, performing standardization processing on the data;
determining the standardized numerical value as a corresponding label value;
for non-numeric types, mapping the non-numeric types into a pre-built dictionary to obtain the single-hot codes of the non-numeric types;
determining the one-time thermal code as a corresponding tag value;
and constructing the tag value into the tag vector according to the sequence of the corresponding data in the data record.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311340384.7A CN117094230B (en) | 2023-10-17 | 2023-10-17 | Urban lamplight optimization model training system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311340384.7A CN117094230B (en) | 2023-10-17 | 2023-10-17 | Urban lamplight optimization model training system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117094230A CN117094230A (en) | 2023-11-21 |
| CN117094230B true CN117094230B (en) | 2024-02-13 |
Family
ID=88783628
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311340384.7A Active CN117094230B (en) | 2023-10-17 | 2023-10-17 | Urban lamplight optimization model training system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117094230B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107895395A (en) * | 2017-11-08 | 2018-04-10 | 成都生活家网络科技有限公司 | A kind of VRay light autoplacement method |
| CN110889029A (en) * | 2018-08-17 | 2020-03-17 | 北京京东金融科技控股有限公司 | City target recommendation method and device |
| CN112752375A (en) * | 2018-11-02 | 2021-05-04 | 中国计量大学 | Commercial hotel guest room lighting controller and lighting system based on scene automatic identification |
| CN114186735A (en) * | 2021-12-10 | 2022-03-15 | 沭阳鸿行照明有限公司 | Fire-fighting emergency illuminating lamp layout optimization method based on artificial intelligence |
-
2023
- 2023-10-17 CN CN202311340384.7A patent/CN117094230B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107895395A (en) * | 2017-11-08 | 2018-04-10 | 成都生活家网络科技有限公司 | A kind of VRay light autoplacement method |
| CN110889029A (en) * | 2018-08-17 | 2020-03-17 | 北京京东金融科技控股有限公司 | City target recommendation method and device |
| CN112752375A (en) * | 2018-11-02 | 2021-05-04 | 中国计量大学 | Commercial hotel guest room lighting controller and lighting system based on scene automatic identification |
| CN114186735A (en) * | 2021-12-10 | 2022-03-15 | 沭阳鸿行照明有限公司 | Fire-fighting emergency illuminating lamp layout optimization method based on artificial intelligence |
Non-Patent Citations (2)
| Title |
|---|
| 城市景观照明质量评价和风险管控;赵浩之;陈俊喜;刘菊;徐华伟;;电子产品可靠性与环境试验(第01期);第66-71页 * |
| 基于最小二乘支持向量机的城市路灯开关时间预测;郑维宏;黄本雄;王非;;计算机与数字工程(第06期);第48-51+109页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117094230A (en) | 2023-11-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113065000B (en) | Multisource heterogeneous data fusion method based on geographic entity | |
| CN112634292B (en) | Asphalt pavement crack image segmentation method based on deep convolutional neural network | |
| Pijanowski et al. | Modelling urbanization patterns in two diverse regions of the world | |
| Long et al. | Mapping block-level urban areas for all Chinese cities | |
| RU2771515C1 (en) | Method and system for determining the location of a high-speed train in a navigation blind spot based on meteorological parameters | |
| CN115482355A (en) | Many-source data driven LOD 2-level city building model enhanced modeling algorithm | |
| CN117015813A (en) | Apparatus, system, method, and medium for adaptively enhancing point cloud data sets for training | |
| CN114818056A (en) | Traffic data integration method, device, equipment and medium based on BIM technology | |
| Baynev et al. | Simulation and study of optical systems based on LEDs | |
| CN117274650A (en) | Urban functional area identification method based on remote sensing-crowd source semantic deep clustering | |
| CN115879630A (en) | Method and device for immediately characterizing and predicting carbon emission based on land utilization | |
| CN116863097A (en) | Space environment data model construction method and device | |
| Zhang et al. | Inferring building function: A novel geo-aware neural network supporting building-level function classification | |
| CN115880466A (en) | Urban engineering surveying and mapping method and system based on unmanned aerial vehicle remote sensing | |
| Ogando-Martínez et al. | Maintenance factor identification in outdoor lighting installations using simulation and optimization techniques | |
| CN117094230B (en) | Urban lamplight optimization model training system and method | |
| Espey | Empirical modelling of public lighting emission functions | |
| CN114168795A (en) | Building three-dimensional model mapping and storing method and device, electronic equipment and medium | |
| Lin et al. | An MIU-based deep embedded clustering model for urban functional zoning from remote sensing images and VGI data | |
| CN115392137B (en) | Three-dimensional simulation system based on karst water and soil coupling effect that sinks | |
| CN110174714A (en) | Street spacial sight sunshine time mass measurement method and system based on machine learning | |
| Terrich et al. | The Effect of Spill Light on Street Lighting Energy Efficiency and Light Pollution | |
| Kippers et al. | Automatic modelling of 3D trees using aerial LIDAR point cloud data and deep learning | |
| Hemmes | Classification of large scale outdoor point clouds using convolutional neural networks | |
| CN114943693B (en) | Jetson Nano bridge crack detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |