An intelligent factory building lighting control method and system

Abstract

The present invention relates to the technical field of lighting control, and particularly to an intelligent lighting control method and system for a factory building. The method includes the following steps: collecting lighting brightness data for multiple time periods in the area irradiated by the light source; taking images for multiple time periods and preprocessing them to obtain image preprocessing data; generating an attenuation map of the brightness within the area to obtain brightness attenuation map data; matching the movement characteristics of personnel in adjacent lighting areas for the lighting area distribution data to obtain a movement detail matching data set; constructing a Bayesian probability classification model to obtain a lighting probability classification model; performing dynamic analysis of the lighting brightness to obtain multi-dimensional brightness change data; performing time series correction modeling to obtain a time series brightness correction model; predicting the brightness of personnel flow based on the lighting probability classification model, and correcting the brightness of natural light based on the time series brightness correction model. The present invention can efficiently perform intelligent lighting control for an intelligent factory building.

Description

Technical Field

[0001] The present invention relates to the technical field of lighting control, and in particular, to an intelligent factory building lighting control method and system. Background Art

[0002] The lighting control of intelligent factories is one of the important technologies in the field of building intelligence, and is widely used in industrial factories, office buildings, shopping malls, hotels and other places. According to factors such as ambient light, personnel activities, and time plans, it automatically or remotely adjusts parameters such as the on / off, brightness, and color temperature of lamps, so as to achieve energy saving, comfort, safety, and beauty of lighting. However, the existing intelligent lighting control methods have problems such as poor flexibility and inability to respond to changes in personnel behavior. The problem of slow brightness adjustment response caused by the long-term use of light sensors or sensor failure problems are difficult to detect and solve in a timely manner. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide an intelligent factory building lighting control method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, an intelligent factory building lighting control method includes the following steps:

[0005] Step S1: Collect the lighting brightness of the light source irradiation area in the intelligent factory building in multiple time periods to obtain lighting brightness data; analyze the lighting area distribution of the lighting brightness data to obtain lighting area distribution data; take pictures of the lighting area distribution data in multiple time periods and preprocess them to obtain image preprocessing data; generate a brightness attenuation map of the brightness within the area according to the image preprocessing data for the lighting area distribution data to obtain brightness attenuation map data;

[0006] Step S2: Match the personnel movement characteristics of adjacent lighting areas for the lighting area distribution data to obtain a moving detail matching data set; construct a Bayesian probability classification model for the moving detail matching data set to obtain a lighting probability classification model;

[0007] Step S3: Perform dynamic analysis of the lighting brightness on the image preprocessing data according to the brightness attenuation map data to obtain multi-dimensional brightness change data; perform time series correction modeling on the multi-dimensional brightness change data to obtain a time series brightness correction model;

[0008] Step S4: Predict the regional brightness change during personnel flow based on the lighting probability classification model to obtain flow brightness prediction data; predict the regional brightness correction during natural light change based on the time series brightness correction model to obtain natural light correction time series data.

[0009] First, the present invention can collect the illumination brightness data of the light source irradiation area in the intelligent factory building in multiple time periods, so as to obtain the illumination brightness data, understand the illumination conditions in different time periods, and provide a basis for subsequent illumination control. By analyzing the illumination area distribution of the illumination brightness data, the illumination area distribution data can be obtained, and different illumination areas can be divided, providing a reference for subsequent illumination optimization. By taking and preprocessing images of the illumination area distribution data in multiple time periods, the image preprocessing data can be obtained, eliminating noise and interference in the images, and providing clear image data for subsequent image analysis. By generating the attenuation map of the brightness within the area based on the image preprocessing data for the illumination area distribution data, the brightness attenuation map data can be obtained, reflecting the brightness distribution and attenuation conditions in different areas. By matching the personnel movement characteristics of adjacent illumination areas for the illumination area distribution data, the moving detail matching data set can be obtained, capturing the movement rules and characteristics of personnel between different illumination areas, and providing data support for subsequent personnel flow prediction. By constructing a Bayesian probability classification model for the moving detail matching data set, the illumination probability classification model can be obtained, using the Bayesian theory to classify and predict the moving probability of personnel between different illumination areas. By performing dynamic analysis of the illumination brightness on the image preprocessing data based on the brightness attenuation map data, the multi-dimensional brightness change data can be obtained, analyzing the dynamic change of the brightness in different illumination areas over time, and providing dynamic data support for subsequent brightness adjustment. By performing time series correction modeling on the multi-dimensional brightness change data, the time series brightness correction model can be obtained, using the time series analysis method to correct and optimize the multi-dimensional brightness change data. By predicting the regional brightness change during personnel flow based on the illumination probability classification model, the flow brightness prediction data can be obtained, predicting the brightness change requirements in different areas according to the probability of personnel flow, and providing predictive data support for subsequent brightness control. By predicting the regional brightness correction during natural light change based on the time series brightness correction model, the natural light correction time series data can be obtained, predicting the brightness correction requirements in different areas according to the change of natural light.

[0010] Preferably, the present invention also provides an intelligent factory building lighting control system for executing the intelligent factory building lighting control method as described above. The intelligent factory building lighting control system includes:

[0011] An attenuation map generation module, which is used to collect the illumination brightness of the light source irradiation area in the intelligent factory building at multiple time periods, so as to obtain illumination brightness data; analyze the illumination area distribution of the illumination brightness data, so as to obtain illumination area distribution data; take pictures of the illumination area distribution data at multiple time periods and perform preprocessing, so as to obtain image preprocessing data; generate an attenuation map of the brightness within the area according to the image preprocessing data for the illumination area distribution data, so as to obtain brightness attenuation map data;

[0012] An illumination probability classification module, which is used to match the personnel movement characteristics of adjacent illumination areas for the illumination area distribution data, so as to obtain a mobile detail matching data set; construct a Bayesian probability classification model for the mobile detail matching data set, so as to obtain an illumination probability classification model;

[0013] A timing correction modeling module, which is used to perform dynamic analysis of the illumination brightness on the image preprocessing data according to the brightness attenuation map data, so as to obtain multi-dimensional brightness change data; perform timing correction modeling on the multi-dimensional brightness change data, so as to obtain a timing brightness correction model;

[0014] A regional brightness prediction module, which is used to predict the regional brightness change during personnel flow based on the illumination probability classification model, so as to obtain mobile brightness prediction data; perform regional brightness correction prediction during natural light change based on the timing brightness correction model, so as to obtain natural light correction timing data.

[0015] In summary, the present invention provides an intelligent factory building lighting control system, which is composed of an attenuation map generation module, an illumination probability classification module, a timing correction modeling module, and a regional brightness prediction module, and can implement any one of the intelligent factory building lighting control methods described in the present invention. By analyzing the images in the time and distance dimensions of the light source irradiation area including artificial light and other natural light according to the different distribution of the number of people, the personnel movement law in the edge of the effective illumination area is found, so as to quickly judge the lighting requirements of the personnel moving to the next illumination area in the busy intelligent factory building and make corresponding adjustments. And according to the brightness change law and brightness requirements of different regions at different times, a timing model is established to ensure that the brightness of the area can be adjusted in time even when the light sensor responds slowly or fails, and the data can be summarized to solve possible fault problems in time. The internal structure of the system cooperates with each other, thus simplifying the operation process of the intelligent factory building lighting control system. Description of the Drawings

[0016] By reading the detailed description of the non-restrictive embodiments with reference to the following drawings, other features, purposes, and advantages of the present invention will become more obvious:

[0017] Figure 1Schematic diagram of the step flow of the intelligent factory building lighting control method described in the present invention;

[0018] Figure 2 is Figure 1 detailed step flow diagram of step S1 in

[0019] Figure 3 is Figure 2 detailed step flow diagram of step S2 in Specific implementation manner

[0020] The technical method of the present invention patent will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments in the present invention without creative efforts belong to the scope of protection of the present invention.

[0021] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0022] It should be understood that although terms such as "first" and "second" may be used here to describe each unit, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit can be called the second unit, and similarly the second unit can be called the first unit. The term "and / or" used here includes any and all combinations of one or more of the listed related items.

[0023] To achieve the above object, please refer to Figures 1 to 3 , the present invention provides an intelligent factory building lighting control method, including the following steps:

[0024] Step S1: Collect the lighting brightness of the light source irradiation area in the intelligent factory building in multiple time periods to obtain lighting brightness data; analyze the lighting area distribution of the lighting brightness data to obtain lighting area distribution data; take pictures and preprocess the lighting area distribution data in multiple time periods to obtain image preprocessing data; generate an attenuation map of the brightness within the area based on the image preprocessing data for the lighting area distribution data to obtain brightness attenuation map data;

[0025] Step S2: Match the personnel movement characteristics of adjacent lighting areas against the lighting area distribution data to obtain a moving detail matching data set; construct a Bayesian probability classification model for the moving detail matching data set to obtain a lighting probability classification model.

[0026] Step S3: Perform dynamic analysis of the lighting brightness on the image preprocessing data according to the brightness attenuation map data to obtain multi-dimensional brightness change data; perform time series correction modeling on the multi-dimensional brightness change data to obtain a time series brightness correction model.

[0027] Step S4: Predict the regional brightness change during personnel flow based on the lighting probability classification model to obtain flow brightness prediction data; predict the regional brightness correction during natural light change based on the time series brightness correction model to obtain natural light correction time series data.

[0028] In the embodiment of the present invention, please refer to Figure 1 As shown, it is a schematic diagram of the step flow of the intelligent factory building lighting control method described in the present invention. In this example, the steps of the intelligent factory building lighting control method include:

[0029] Step S1: Collect the lighting brightness of the light source irradiation area in the intelligent factory building for multiple periods to obtain lighting brightness data; perform lighting area distribution analysis on the lighting brightness data to obtain lighting area distribution data; take pictures and preprocess the lighting area distribution data for multiple periods to obtain image preprocessing data; generate a brightness attenuation map data of the brightness attenuation within the area according to the image preprocessing data and the lighting area distribution data.

[0030] In the embodiment of the present invention, light sensors are installed in each lighting area of the factory building, and the collected lighting brightness data is sent to the intelligent gateway through a wireless communication module; the intelligent gateway aggregates and transmits the lighting brightness data to the centralized controller, and the centralized controller analyzes the data, divides the factory building into several lighting areas according to different lighting requirements and standards, and generates lighting area distribution data; cameras are installed in each lighting area of the factory building, and the captured image data is sent to the intelligent gateway through a wireless communication module; the intelligent gateway aggregates and transmits the image data to the centralized controller, and the centralized controller preprocesses the data, including operations such as denoising, enhancement, and segmentation, to improve the quality and usability of the image, and generates image preprocessing data; the centralized controller performs brightness attenuation analysis on the lamps in each lighting area according to the image preprocessing data and the lighting area distribution data, calculates the change rule of the lamp brightness with distance, and generates brightness attenuation map data.

[0031] Step S2: Match the personnel movement characteristics of adjacent lighting areas against the lighting area distribution data to obtain a moving detail matching data set; construct a Bayesian probability classification model for the moving detail matching data set to obtain a lighting probability classification model;

[0032] In the embodiment of the present invention, human body induction sensors are installed in each lighting area of the factory building, and the detected personnel movement data is sent to the intelligent gateway through the wireless communication module; the intelligent gateway aggregates and transmits the personnel movement data to the centralized controller, and the centralized controller analyzes the data. According to the lighting area distribution data, to judge the possibility of personnel moving from one lighting area to another, a moving detail matching data set is generated; according to the moving detail matching data set, the centralized controller uses the Bayesian probability classification model to predict the lighting demand of each lighting area (that is, the probability of personnel moving from one lighting area to another corresponds to the lighting demand of that area), and generates a lighting probability classification model. This model can calculate the lighting probability of personnel entering an adjacent lighting area from one area based on the historical feature data and current data of personnel movement.

[0033] Step S3: Perform dynamic analysis of the lighting brightness on the image preprocessing data according to the brightness attenuation map data to obtain brightness multi-dimensional change data; perform time series correction modeling on the brightness multi-dimensional change data to obtain a time series brightness correction model;

[0034] In the embodiment of the present invention, the centralized controller performs dynamic analysis of the lighting brightness of the lamps and the natural light or other reflected light sources in each lighting area according to the brightness attenuation map data and the image preprocessing data, calculates the change rules of its brightness with factors such as time, space, and angle, and generates brightness multi-dimensional change data; based on the brightness multi-dimensional change data, the attenuation rules of brightness in multiple dimensions with time and distance are obtained, and the time period when the brightness needs to be corrected is found according to the actual brightness adaptation situation in the lighting area, and a time series brightness correction model is generated accordingly. This model can calculate the brightness correction value of the lamp according to the historical data and current data of the brightness change of the lamp, so as to realize the adaptive dimming of the lamp.

[0035] Step S4: Predict the regional brightness change during personnel flow based on the lighting probability classification model to obtain flow brightness prediction data; predict the regional brightness correction during natural light change based on the time series brightness correction model to obtain natural light correction time series data.

[0036] In the embodiments of the present invention, the centralized controller predicts the regional brightness change during personnel flow for the lamps in each lighting area according to the lighting probability classification model, and generates the mobile brightness prediction data. This data can predict the on / off or brightness state of the lamps in the next area according to the probability of personnel moving to another lighting area, so as to realize the intelligent control of the lamps; the centralized controller predicts the regional brightness correction during natural light change for the lamps in each lighting area according to the timing brightness correction model, and generates the natural light correction timing data. This data can predict the brightness correction value of the lamps according to the change of natural light or other reflected light source brightness, so as to realize the adaptive dimming of the lamps.

[0037] First, the present invention can collect the lighting brightness data by collecting the lighting brightness in the light source irradiation area of the intelligent factory building in multiple time periods, so as to understand the lighting conditions in different time periods and provide a basis for subsequent lighting control. By analyzing the lighting area distribution of the lighting brightness data, the lighting area distribution data can be obtained, so as to divide different lighting areas and provide a reference for subsequent lighting optimization. By taking pictures and preprocessing the lighting area distribution data in multiple time periods, the image preprocessing data can be obtained, so as to eliminate the noise and interference in the image and provide clear image data for subsequent image analysis. By generating the brightness attenuation map of the regional brightness in the lighting area distribution data according to the image preprocessing data, the brightness attenuation map data can be obtained, so as to reflect the brightness distribution and attenuation in different areas. By matching the personnel movement characteristics of adjacent lighting areas in the lighting area distribution data, the mobile detail matching data set can be obtained, so as to capture the movement rules and characteristics of personnel between different lighting areas and provide data support for subsequent personnel flow prediction. By constructing the Bayesian probability classification model for the mobile detail matching data set, the lighting probability classification model can be obtained, so as to classify and predict the movement probability of personnel between different lighting areas by using the Bayesian theory. By dynamically analyzing the lighting brightness of the image preprocessing data according to the brightness attenuation map data, the brightness multi-dimensional change data can be obtained, so as to analyze the dynamic change of the brightness in different lighting areas over time and provide dynamic data support for subsequent brightness adjustment. By performing timing correction modeling on the brightness multi-dimensional change data, the timing brightness correction model can be obtained, so as to correct and optimize the brightness multi-dimensional change data by using the method of timing analysis. By predicting the regional brightness change during personnel flow based on the lighting probability classification model, the mobile brightness prediction data can be obtained, so as to predict the brightness change requirements of different areas according to the probability of personnel flow and provide predictive data support for subsequent brightness control. By predicting the regional brightness correction during natural light change based on the timing brightness correction model, the natural light correction timing data can be obtained, so as to predict the brightness correction requirements of different areas according to the change of natural light.

[0038] Preferably, step S1 includes the following steps:

[0039] Step S11: Collect the lighting brightness at multiple time periods in the illumination area of the light source in the intelligent factory building to obtain lighting brightness data;

[0040] Step S12: Analyze the distribution of the lighting area based on the lighting brightness data to obtain lighting area distribution data;

[0041] Step S13: Take images of the lighting area distribution data at multiple time periods to obtain lighting image data; perform image preprocessing on the lighting image data to obtain image preprocessing data;

[0042] Step S14: Detect the lighting edge from the image preprocessing data to obtain lighting edge data;

[0043] Step S15: Measure the brightness of the image preprocessing data based on the lighting edge data to obtain brightness measurement data;

[0044] Step S16: Analyze the local brightness attenuation of the brightness measurement data to obtain the local brightness attenuation coefficient;

[0045] Step S17: Generate an attenuation map of the brightness within the area based on the local brightness attenuation coefficient for the lighting area distribution data to obtain brightness attenuation map data.

[0046] As an embodiment of the present invention, referring to Figure 2 shown, it is Figure 1 a detailed step flow schematic diagram of step S1 in

[0047] Step S11: Collect the lighting brightness at multiple time periods in the illumination area of the light source in the intelligent factory building to obtain lighting brightness data;

[0048] In the embodiment of the present invention, a light sensor, a wireless communication module, and an intelligent gateway are used to collect and transmit lighting brightness data;

[0049] Step S12: Analyze the distribution of the lighting area based on the lighting brightness data to obtain lighting area distribution data;

[0050] In the embodiment of the present invention, a centralized controller is used to analyze and divide the lighting brightness data to generate lighting area distribution data;

[0051] Step S13: Take images of the lighting area distribution data at multiple time periods to obtain lighting image data; perform image preprocessing on the lighting image data to obtain image preprocessing data;

[0052] In an embodiment of the present invention, a camera, a wireless communication module, and an intelligent gateway are used to collect and transmit image data; a centralized controller is used to preprocess the image data to generate preprocessed image data;

[0053] Step S14: Perform illumination edge detection on the preprocessed image data to obtain illumination edge data;

[0054] In an embodiment of the present invention, the centralized controller is used to perform edge detection on the preprocessed image data. Starting from the edge corners of the image, the degree of change in the image brightness gradient is calculated towards the center of the image. When the change gradient of the illumination brightness starts to increase significantly, the pixel point is recorded, thereby finding the boundary of the illumination area and generating illumination edge data;

[0055] Step S15: Perform brightness measurement on the preprocessed image data according to the illumination edge data to obtain brightness measurement data;

[0056] In an embodiment of the present invention, the centralized controller is used to perform brightness measurement on the preprocessed image data. Using the grayscale value of the image, the brightness of each pixel point is calculated within the edge range of the image, and brightness measurement data is generated;

[0057] Step S16: Perform local brightness attenuation analysis on the brightness measurement data to obtain a local brightness attenuation coefficient;

[0058] In an embodiment of the present invention, by performing local brightness attenuation analysis on the brightness measurement data, using the brightness measurement data, each illumination area is divided into local areas according to the approximate gradient area of the brightness, and the attenuation coefficient of the brightness of each local area within each illumination area with respect to distance is calculated, and a local brightness attenuation coefficient is generated;

[0059] Step S17: Generate an in-region brightness attenuation map for the illumination area distribution data according to the local brightness attenuation coefficient to obtain brightness attenuation map data.

[0060] In an embodiment of the present invention, according to the local brightness attenuation coefficient and the illumination area distribution data, a brightness attenuation map is generated for the lamps within each illumination area, that is, a curve of the change in regional brightness with respect to distance is plotted according to the summary of the local brightness attenuation coefficient, and brightness attenuation map data is generated.

[0061] The present invention first collects and transmits lighting brightness data by using a light sensor, a wireless communication module and an intelligent gateway, improving the accuracy of lighting brightness data collection and the reliability of transmission. By using a centralized controller to analyze and divide the lighting brightness data, lighting area distribution data is generated, which can quickly and effectively analyze and divide the lighting brightness data to generate lighting area distribution data, improving the generation efficiency and quality of the lighting area distribution data. The image data is preprocessed to generate image preprocessing data, which can improve the generation efficiency and quality of the image preprocessing data. The brightness of the image preprocessing data is measured, and through local brightness attenuation analysis of the brightness measurement data, the brightness attenuation analysis can be targeted at a part of the local areas with strong gradient feature correlation first, improving the overall operation efficiency for subsequent brightness attenuation analysis extending to the entire area.

[0062] Preferably, step S2 includes the following steps:

[0063] Step S21: According to the lighting area distribution data and using an infrared sensor to capture infrared human portraits in different areas of the intelligent factory building, thereby obtaining infrared human portrait movement data;

[0064] Step S22: Conduct time series fluctuation analysis on the infrared human portrait movement data, thereby obtaining the fluctuation data of the number of people;

[0065] Step S23: Conduct the distribution analysis of the number of people on the lighting area distribution data according to the fluctuation data of the number of people, thereby obtaining the distribution data of the number of people;

[0066] Step S24: Label the distribution data of the number of people according to a preset sensitive threshold of the number of people, thereby obtaining the distribution label data of the number of people;

[0067] Step S25: According to the distribution label data of the number of people and using a millimeter wave sensor to capture the details of personnel movement in different areas of the intelligent factory building, thereby obtaining a set of movement detail data;

[0068] Step S26: Match the movement characteristics of personnel in adjacent lighting areas for the set of movement detail data and the infrared human portrait movement data, thereby obtaining a set of movement detail matching data;

[0069] Step S27: Conduct lighting correlation marking on the lighting area distribution data according to the set of movement detail matching data, thereby obtaining lighting correlation marking data;

[0070] Step S28: Construct a Bayesian probability classification model for the lighting correlation marking data, thereby obtaining a lighting probability classification model.

[0071] As an embodiment of the present invention, refer to Figure 3As shown, it is a detailed step - by - step schematic diagram of step S2. In this embodiment, step S2 includes the following steps:

[0072] Step S21: According to the lighting area distribution data and using an infrared sensor, capture infrared human portraits in different areas of the intelligent factory building to obtain infrared human portrait movement data;

[0073] In the embodiment of the present invention, infrared sensors are installed in each lighting area of the factory building, and the collected infrared image data is sent to the intelligent gateway through a wireless communication module; the intelligent gateway aggregates and transmits the infrared image data to the centralized controller. The centralized controller performs human portrait detection and tracking on the data, judges the area where each human portrait is located according to the lighting area distribution data, and calculates its movement distance and direction, so as to obtain infrared human portrait movement data.

[0074] Step S22: Conduct time - series fluctuation analysis on the infrared human portrait movement data to obtain the pedestrian flow fluctuation data;

[0075] In the embodiment of the present invention, by conducting time - series fluctuation analysis on the infrared human portrait movement data, using statistical methods and performing Fourier transform analysis on the changing data, calculate the change trend and fluctuation range of the pedestrian flow in each lighting area over time, so as to obtain the pedestrian flow fluctuation data;

[0076] Step S23: Conduct pedestrian flow distribution analysis on the lighting area distribution data according to the pedestrian flow fluctuation data to obtain the pedestrian flow distribution data;

[0077] In the embodiment of the present invention, according to the pedestrian flow fluctuation data and the lighting area distribution data, conduct distribution analysis on the pedestrian flow in each lighting area, use the clustering method to divide the areas with similar pedestrian flows into one category, and generate the pedestrian flow distribution data.

[0078] Step S24: Perform data tagging on the pedestrian flow distribution data according to the preset pedestrian flow sensitivity threshold to obtain the pedestrian flow distribution tag data;

[0079] In the embodiment of the present invention, according to the preset pedestrian flow sensitivity threshold and the pedestrian flow distribution data, perform data tagging on the pedestrian flow in each lighting area. Using the classification method, mark the areas with pedestrian flow higher than the threshold as high - sensitivity areas, mark the areas with pedestrian flow lower than the threshold as low - sensitivity areas, and generate the pedestrian flow distribution tag data. This data is used in step S25 and can also be used for subsequent collection of lighting brightness and image shooting in multiple time periods. Specifically, when collecting lighting brightness and shooting in the lighting areas during the time periods with more pedestrian flows, because the lighting experience in these time periods is generally more sensitive, and the pedestrian flow sensitivity threshold also needs to be statistically analyzed and modified based on continuous experience and personnel feedback.

[0080] Step S25: According to the crowd flow distribution label data, use millimeter-wave sensors to capture the personnel movement details in different areas of the intelligent factory building, so as to obtain a movement detail data set;

[0081] In the embodiment of the present invention, millimeter-wave sensors are installed in each lighting area of the factory building, and the collected millimeter-wave image data is sent to the intelligent gateway through the wireless communication module; the intelligent gateway aggregates and transmits the millimeter-wave image data to the centralized controller, and the centralized controller captures the personnel movement details of the data, and according to the crowd flow distribution label data, captures the personnel movement details in the highly sensitive area with higher precision, so as to obtain a movement detail data set.

[0082] Step S26: Match the personnel movement characteristics in adjacent lighting areas for the movement detail data set and the infrared portrait movement data, so as to obtain a movement detail matching data set;

[0083] In the embodiment of the present invention, by matching the personnel movement characteristics in adjacent lighting areas for the movement detail data set and the infrared portrait movement data, using the image processing method, the millimeter-wave image and the infrared image are aligned and fused, so as to obtain a movement detail matching data set.

[0084] Step S27: Perform lighting association marking on the lighting area distribution data according to the movement detail matching data set, so as to obtain lighting association marking data;

[0085] In the embodiment of the present invention, according to the movement detail matching data set and the lighting area distribution data, each lighting area is subjected to lighting association marking. Using the association rule method, the relevant areas that affect the lighting demand during personnel movement are found, and lighting association marking data is generated, that is, the degree of association between the personnel movement characteristics in each lighting area and the lighting state of that area during that period.

[0086] Step S28: Construct a Bayesian probability classification model for the lighting association marking data, so as to obtain a lighting probability classification model.

[0087] In the embodiment of the present invention, according to the lighting association marking data, probability classification of the entry trend of important personnel movement characteristics in adjacent areas is performed. For example, according to the movement direction, speed and behavior action characteristics of personnel, the Bayesian probability model is used to train the characteristics of the probability classification of personnel entering its adjacent lighting area in each lighting area, so as to generate a lighting probability classification model.

[0088] First, based on the pedestrian flow distribution data and lighting association marker data of the lighting area, the present invention can more accurately analyze and optimize the lighting requirements of each lighting area. By understanding the movement trends and behavior patterns of people in different areas, the lighting settings can be adjusted according to real-time needs, providing appropriate lighting brightness and coverage, improving the lighting effect and comfort. By fusing infrared image data and millimeter-wave image data, the movement details and position information of people can be captured more accurately. By continuously collecting and analyzing pedestrian flow data, movement detail data, and lighting requirement data, historical data and statistical models can be established to more accurately predict future pedestrian flow trends and lighting requirements. This helps with long-term energy planning and management decisions, optimizing the energy efficiency and sustainability of the lighting system. By continuously updating the pedestrian flow sensitivity threshold and lighting association marker data, and combining with people's feedback and experience, the performance and accuracy of the lighting system can be continuously improved and optimized. This enables the system to adapt to changes in different environments and needs, and provide a more intelligent and personalized lighting experience.

[0089] Preferably, in step S28, the Bayesian probability classification model construction for the lighting association marker data adopts the following formula:

[0090]

[0091] In the formula, P is the probability of regional lighting adjustment, A is the number of pedestrian movement characteristics of the associated marker, B is the number of pedestrian movement characteristics of the associated marker in the adjacent lighting area, f a,b is the similarity between the pedestrian movement characteristic value of the a-th associated marker and the pedestrian movement characteristic value of the b-th associated marker in the adjacent lighting area, μ b is the mean of the pedestrian movement characteristics of the associated marker in the adjacent lighting area, β b is the importance weight value of the pedestrian movement characteristics of the b-th associated marker.

[0092] The present invention constructs a formula for the probability of regional lighting adjustment. By setting different pedestrian flow sensitivity thresholds and importance weight values of pedestrian movement characteristics, this formula can flexibly adapt to different scenarios and needs, improving the flexibility and adaptability of lighting control. This formula consists of the sum of two parts. The first part is the variance of the pedestrian movement characteristics, and the second part is the ratio of the similarity of the pedestrian movement characteristics. These two parts respectively reflect the degree of change and consistency of the pedestrian movement characteristics, and have different effects on the probability of lighting adjustment. In the first part, the larger the variance, the more unstable the pedestrian movement characteristics are, and the more the lighting needs to be adjusted. Therefore, the coefficient β b of the variance is positive. β b also represents the importance of the pedestrian movement characteristics of the b-th associated marker, β bThe larger it is, the greater the impact of the feature on lighting adjustment, and vice versa. In the second part, the greater the similarity, the more consistent the personnel movement features are, and the less lighting adjustment is required. Therefore, the coefficient of similarity is negative. The calculation of similarity adopts the form of geometric mean, which can effectively avoid distortion caused by some feature values being too large or too small. Specifically, P is a number between 0 and 1, indicating the possibility of whether lighting adjustment is required given the personnel movement features and lighting area distribution data. The closer P is to 1, the more necessary it is to adjust the lighting, and vice versa. A is a positive integer, indicating how many different personnel movement features are associated and marked in a lighting area. The larger A is, the more complex the personnel movement features are, and the more comprehensive consideration is required, and vice versa. B is also a positive integer, indicating how many different personnel movement features are associated and marked in the adjacent area of a lighting area. The larger B is, the greater the impact of the adjacent lighting area is, and the more coordinated adjustment is required, and vice versa, f a,b is a number between 0 and 1, indicating the degree of similarity between two feature values. f a,b The closer it is to 1, the more similar the two feature values are, and the less lighting adjustment is required, and vice versa, μ b is a non - negative number, indicating the average level of all associated and marked personnel movement feature values in an adjacent lighting area. μ b The larger it is, the higher the personnel movement features in the adjacent lighting area are, and the more lighting adjustment is required, and vice versa, β b is a positive number, indicating the importance of the b - th associated and marked personnel movement feature in the calculation of the probability of lighting adjustment. β bThe larger it is, the greater the impact of this feature on lighting adjustment, and vice versa. This formula can comprehensively consider multiple factors, such as the quantity, variation, consistency, importance of personnel movement features, as well as lighting area distribution data, so as to obtain a reasonable probability value for lighting adjustment, guiding the lighting control system of an intelligent factory building to perform dynamic adjustment. This formula can be combined with a Bayesian probability classification model as the input of the model. According to historical data and current data, it can predict future lighting requirements, improve the intelligent level of lighting control, and achieve energy-saving and comfortable lighting effects. It should be noted that in the present invention, it is also possible not to use this formula. In constructing the Bayesian probability classification model for lighting-related marker data in step S28, it is also possible to directly perform vector quantization on the personnel movement direction features, and then construct a model according to the Bayesian probability classification principle. The advantage of doing this is higher efficiency and lower model training cost. By directly predicting only by considering the personnel movement direction in the lighting edge area, but too few features are considered in a complex environment, and the obtained prediction model is difficult to accurately predict complex actual situations. While the above formula can consider more personnel movement features exposed in complex actual situations and their associated data for entering another lighting area, so as to better handle some complex scenarios. However, in some areas, a dual-model method can also be considered for adaptive prediction.

[0093] Preferably, step S26 includes the following steps:

[0094] Step S261: Select important features from the movement detail data set and the infrared portrait movement data, so as to obtain movement angle feature data, movement speed feature data, and movement action feature data;

[0095] Step S262: Use the FLANN algorithm to perform feature ratio filtering for adjacent lighting areas on the movement angle feature data, movement speed feature data, and movement action feature data, so as to obtain movement feature ratio data;

[0096] Step S263: Construct a convolution model for the movement feature ratio data, so as to obtain a convolution model; use the RANSAC algorithm to perform iterative optimization on the convolution model, so as to obtain an optimized convolution model;

[0097] Step S264: Based on the optimized convolution model, perform personnel movement feature matching for adjacent lighting areas on the movement detail data set, so as to obtain movement detail matching data.

[0098] In the embodiments of the present invention, important feature selection is performed on the mobile detail dataset and the infrared portrait mobile data. Using feature extraction methods such as SIFT1, HOG2, HAAR3, etc., features such as moving angle, moving speed, and moving action are extracted from each data point, and moving angle feature data, moving speed feature data, and moving action feature data are generated. The FLANN algorithm is used to perform feature ratio filtering of adjacent illumination regions on the moving angle feature data, moving speed feature data, and moving action feature data. Using the nearest neighbor search method, the feature points most similar to other regions in each illumination region are found, and their feature ratios are calculated to obtain moving feature ratio data. A convolutional model is constructed for the moving feature ratio data. Using the convolutional neural network method, the moving feature ratio data is used as input, and through operations such as multi-layer convolution, pooling, activation, and fully connected, a convolutional model capable of representing the relationship between illumination regions is obtained. The RANSAC algorithm is used to iteratively optimize the convolutional model. Using the random sample consensus method, a part of the data is randomly selected from the moving feature ratio data as inliers, and they are used to fit the parameters of the convolutional model, and the error of the model is calculated. Repeating multiple times, the convolutional model with the smallest error is selected as the optimized convolutional model. Based on the optimized convolutional model, the personnel movement feature matching of adjacent illumination regions in the mobile detail dataset is performed. Using the model prediction method, the mobile detail dataset is used as input, and through the optimized convolutional model, the output of the personnel movement features of each illumination region is obtained, and compared with the outputs of other regions to find the most matching region and generate mobile detail matching data.

[0099] First, by performing important feature selection on the mobile detail dataset and infrared portrait movement data, the present invention can improve the extraction efficiency and quality of movement feature data, reduce the extraction cost and complexity of movement feature data, and improve the extraction accuracy and stability of movement feature data. Using the FLANN algorithm to perform feature ratio filtering on adjacent illumination regions for movement angle feature data, movement speed feature data, and movement action feature data can improve the filtering efficiency and quality of movement feature ratio data, reduce the filtering cost and complexity of movement feature ratio data, and improve the filtering accuracy and stability of movement feature ratio data. Constructing a convolutional model for the movement feature ratio data can improve the construction efficiency and quality of the convolutional model, reduce the construction cost and complexity of the convolutional model, and improve the construction accuracy and stability of the convolutional model. Using the RANSAC algorithm to iteratively optimize the convolutional model can improve the optimization efficiency and quality of the optimized convolutional model, reduce the optimization cost and complexity of the optimized convolutional model, and improve the optimization accuracy and stability of the optimized convolutional model. Based on the optimized convolutional model, performing person movement feature matching for adjacent illumination regions on the mobile detail dataset can improve the matching efficiency and quality of mobile detail matching data, reduce the matching cost and complexity of mobile detail matching data, and improve the matching accuracy and stability of mobile detail matching data.

[0100] Preferably, step S3 includes the following steps:

[0101] Step S31: Perform linear correlation screening on the luminance attenuation map data to obtain linear attenuation standard data;

[0102] Step S32: Perform linear correlation analysis on the luminance attenuation map data according to the linear attenuation standard data to obtain linear luminance attenuation data; perform non-linear correlation analysis of luminance interference on the luminance attenuation map data according to the linear attenuation standard data to obtain non-linear luminance attenuation data;

[0103] Step S33: Perform linear fitting on the linear luminance attenuation data and the non-linear luminance attenuation data to obtain linearly fitted attenuation data;

[0104] Step S34: Perform high-dimensional feature dimensionality reduction on the multi-dimensional linear attenuation data to obtain low-dimensional linear attenuation data;

[0105] Step S35: Construct a time matrix for the low-dimensional linear attenuation data to obtain luminance attenuation matrix data;

[0106] Step S36: Perform dynamic analysis of illumination luminance on the image preprocessing data according to the luminance attenuation matrix data to obtain multi-dimensional luminance change data;

[0107] Step S37: Perform a time-series correction modeling on the multi-dimensional brightness change data to obtain a time-series brightness correction model.

[0108] In the embodiment of the present invention, the type of the illumination light source is selected from the brightness attenuation map data as the independent variable (i.e., artificial light source and other light sources), and the brightness of the illumination light source is used as the dependent variable, forming a set of data with the independent variable. Using a tool for linear correlation analysis, calculate the correlation coefficient between the independent variable and the dependent variable, such as Pearson correlation coefficient, Spearman correlation coefficient, etc. According to the magnitude and significance level of the correlation coefficient, determine whether there is a linear relationship between the independent variable and the dependent variable, that is, whether it conforms to the assumption of linear attenuation. Select the data composed of the independent variable and the dependent variable with a relatively high and significant correlation coefficient as the linear attenuation standard data. According to the linear attenuation standard data, use a tool for linear regression analysis to establish a linear regression model, obtain the regression equation and regression coefficients. Substitute the independent variable in the brightness attenuation map data into the regression equation to calculate the corresponding dependent variable value, that is, the linear brightness attenuation data. According to the non-linear factors in the brightness attenuation map data, use a tool for non-linear regression analysis to correct the linear regression model to obtain a corrected non-linear regression model. Substitute the independent variable and non-linear factors in the brightness attenuation map data into the corrected non-linear regression model to calculate the corresponding dependent variable value, that is, the non-linear brightness attenuation data. Take the linear brightness attenuation data and non-linear brightness attenuation data as scatter data respectively, and plot them on a coordinate plane, where the abscissa is time and the ordinate is brightness. Use a tool for linear fitting to perform linear fitting on the two sets of scatter data respectively to obtain the equations and fitting degrees of the two straight lines. Take the equations and fitting degrees of the two straight lines as the linear fitting attenuation data. Take the independent variable and the dependent variable in the multi-dimensional linear attenuation data as the column vectors of the matrix respectively to form a data matrix, perform principal component analysis or factor analysis on the data matrix to obtain the coefficients and contribution rates of the new variables, and select a certain number of new variables as the independent variables of the low-dimensional linear attenuation data according to the coefficients and contribution rates of the new variables, while retaining the original dependent variable. The low-dimensional linear attenuation data refers to the data in which the brightness of each pixel in the image changes linearly with time and can be represented by a vector, and the brightness attenuation matrix data refers to the brightness change rate of each row in the time matrix and can be represented by a vector. According to the brightness attenuation matrix data, perform a brightness clustering analysis on the image preprocessing data and use machine learning to analyze the illumination brightness change situation in different regions of the image. Performing a time-series correction modeling on the multi-dimensional brightness change data means comparing the change gradients of the brightness in multiple dimensions with the change gradient of each light source and the minimum brightness adaptation value of the region for the multi-dimensional brightness change data, so as to obtain the time points when the natural light brightness change gradient in the region is larger but the artificial light brightness has not changed, screening and marking these time points, and training these time points corresponding to the gradient change characteristics of the region through a memory model, so as to obtain a time-series brightness correction model.

[0109] First, through linear correlation analysis and linear regression analysis, the present invention can determine the brightness attenuation law and trend of the lighting source. This enables more precise control and adjustment of the lighting system to ensure the stability and durability of the light source. In addition, the linear attenuation standard data and the corrected non-linear regression model can be used to predict future brightness changes and perform maintenance and optimization of the lighting system in advance. By using a straight line fitting tool to fit the linear brightness attenuation data and non-linear brightness attenuation data, the trend and fitting degree of the attenuation data can be obtained. This helps to analyze and compare the attenuation speed and stability of different lighting sources and select appropriate light source types and brightness control strategies. Principal component analysis or factor analysis can transform multi-dimensional linear attenuation data into low-dimensional linear attenuation data while retaining the cause variables. This helps to reduce the complexity and dimension of the data and improve the efficiency of data processing and analysis. At the same time, selecting new variables can better capture the key features of brightness changes and provide a more accurate basis for subsequent analysis and decision-making. Through clustering analysis and machine learning analysis of lightness, the lighting brightness changes in different regions of the image can be identified. This helps to perform personalized lighting control and optimization for different regions and provide a lighting experience that better meets the user's needs. By using a memory model and training to obtain a time-series brightness correction model, the time points of light source changes can be identified, thereby realizing the adaptive adjustment of the lighting system and improving the continuity and smoothness of lighting.

[0110] Preferably, step S36 includes the following steps:

[0111] Step S361: Perform HSV color space conversion on the image preprocessing data and extract the lightness component to obtain lightness component data;

[0112] Step S362: Perform brightness-distance attenuation clustering analysis on the lightness component data according to the brightness attenuation matrix data to obtain brightness-distance attenuation clustering data; perform brightness-time attenuation clustering analysis on the lightness component data according to the brightness attenuation matrix data to obtain brightness-time attenuation clustering data;

[0113] Step S363: Perform brightness trend partitioning on the lightness component data according to the brightness-distance attenuation clustering data and the brightness-time attenuation clustering data to obtain brightness partitioning data;

[0114] Step S364: Perform gamma conversion on the brightness partitioning data to obtain gamma conversion partitioning data;

[0115] Step S365: Perform dynamic analysis of the lighting brightness on the gamma conversion partitioning data to obtain brightness multi-dimensional change data.

[0116] In an embodiment of the present invention, relevant information of image preprocessing data is obtained, and then the method and formula for RGB to HSV conversion are obtained. According to the formula, the RGB values of each pixel in the image preprocessing data are calculated to obtain the corresponding HSV values, which are stored as lightness component data. The image preprocessing data is converted into a two-dimensional array, where each element is an RGB color value. The RGB color space is converted into the HSV color space to obtain a new two-dimensional array, where each element is an HSV color value. According to the definition of the HSV color space, the lightness component, i.e., the V channel in the HSV color space, is extracted. Based on the brightness attenuation matrix data, the lightness component data is analyzed by brightness-distance attenuation clustering and brightness-time attenuation clustering, i.e., clustering according to the attenuation characteristics of the brightness weight values of different light sources on the lightness component data with respect to the distance (the change in the distance radiating downward from the artificial light source) and the change in time, so as to obtain two clustering analysis data; according to the brightness-distance attenuation clustering data and the brightness-time attenuation clustering data, the brightness values of each pixel in the lightness component data are divided into trends, and according to the formula of gamma conversion, the brightness values of each pixel in the brightness partition data are non-linearly transformed to amplify the change trend of the brightness. Finally, the dynamic analysis in time series is performed on the pixel brightness data after gamma conversion and partitioning, so as to obtain the brightness multi-dimensional change data.

[0117] First, through the RGB to HSV conversion, the present invention can convert the image preprocessing data from the RGB color space to the HSV color space. This enables better understanding and analysis of the color information in the image, especially the lightness (brightness) component. By extracting the V channel in the HSV color space, the lightness component data can be obtained for subsequent brightness analysis and processing. The brightness-distance attenuation clustering and brightness-time attenuation clustering analysis can cluster the lightness component data according to the brightness attenuation matrix data. This helps to identify different attenuation characteristics in the lightness component data, i.e., the degree to which the brightness is affected by the changes in distance and time. Through the clustering analysis, the lightness component data can be divided into different brightness trend regions, providing a basis for subsequent non-linear transformation. The non-linear transformation of gamma conversion can be performed according to the brightness value of each pixel in the brightness partition data. By applying the formula of gamma conversion, the change trend of the brightness can be amplified, enhancing the brightness difference in the image. This helps to better capture and analyze the change characteristics of the brightness, providing a more accurate data basis for subsequent dynamic analysis in time series. The dynamic analysis in time series can be performed on the pixel brightness data after gamma conversion and partitioning. By analyzing the brightness multi-dimensional change data, the time series change rules and trends of the brightness can be identified. This helps to understand the dynamic behavior and change patterns of the light source, providing more comprehensive information for the control and adjustment of the lighting system.

[0118] Preferably, step S37 includes the following steps:

[0119] Step S371: Use the optical flow method to perform change gradient division on the luminance multi-dimensional change data, so as to obtain luminance change gradient data;

[0120] Step S372: Sort the luminance change gradient data, so as to obtain luminance gradient sequence data;

[0121] Step S373: Perform time point marking on the luminance gradient sequence data according to the preset minimum luminance adaptation data, so as to obtain luminance change time marking data;

[0122] Step S374: Perform time series correction modeling on the luminance change time marking data, so as to obtain a time series luminance correction model.

[0123] In the embodiment of the present invention, first, the optical flow vectors between every two frames of images are calculated, that is, the displacement and velocity of each pixel point in the horizontal and vertical directions. Then, according to the magnitude and direction of the optical flow vector, the luminance change gradient of each pixel point is calculated, that is, the luminance change rate of each pixel point in the horizontal and vertical directions, and the characteristics of the gradient change of the luminance change rate are matched according to the luminance multi-dimensional change data, that is, according to the magnitude and direction of the luminance change gradient, the pixel points are divided into different categories according to different light source dimensions, such as stationary, slow moving, and fast moving. To sort the luminance change gradient data, a comparison function for the luminance change gradient needs to be defined to compare the magnitude and direction of the luminance change gradient. For example, the modulus length and included angle of the vector can be used as the comparison basis, and according to the sorting principle that the natural light change gradient is large but the artificial light source change gradient is small, the luminance gradient sequence data is obtained. Perform time point marking on the luminance gradient sequence data according to the preset minimum luminance adaptation data, where the minimum luminance adaptation data can mainly be fed back by personnel through the feedback system, and the controller determines the magnitude of the minimum acceptable luminance fluctuation in the lighting area according to the luminance feedback, so as to obtain the minimum luminance adaptation data. Performing time point marking on the luminance gradient sequence data means determining the luminance change time point, that is, the time point when the brightness of the light needs to be adjusted. Compare the luminance gradient value and the minimum luminance adaptation value at each time point. If the luminance gradient value is greater than the minimum luminance adaptation value, it means that the luminance change at this time point exceeds the luminance adaptation threshold of this area, so the brightness of the light needs to be adjusted, otherwise no adjustment is required. Perform time series correction modeling on the luminance change time marking data, that is, perform model training on the marked luminance change data corresponding to the time and its corresponding change gradient characteristics, and use the least squares method or other optimization methods to estimate the parameters of the time series model according to the luminance change time marking data. For example, the gradient descent method and the Newton method can be used to obtain a time series luminance correction model.

[0124] First, the present invention can be achieved by calculating the optical flow vector, which can be calculated by the pixel displacement and velocity between every two frames of images. By obtaining the optical flow vector, the motion information of objects in the image can be captured. This helps to identify the moving objects or regions in the image and provides basic data for subsequent brightness change analysis. The brightness change gradient calculation can be achieved according to the magnitude and direction of the optical flow vector. By calculating the brightness change rates of each pixel point in the horizontal and vertical directions, the brightness change gradient can be obtained. This helps to understand the brightness change situation in the image, especially the changes related to the light source. The sorting of the brightness gradient sequence data can be achieved according to the magnitude and direction of the brightness change gradient. By defining a comparison function for the brightness change gradient, the brightness gradient data can be sorted. According to the sorting principle that the natural light change gradient is larger while the artificial light source change gradient is smaller, the sorted brightness gradient sequence data can be obtained. This helps to better understand and sort the brightness changes of different light source dimensions in the image. The preset of the minimum brightness adaptation data can be determined by feedback through a feedback system. According to the brightness feedback of the lighting area, the controller can determine the magnitude of the fluctuation of the minimum acceptable brightness in this area, thereby obtaining the minimum brightness adaptation data. This helps to define the threshold of brightness change for judging whether it is necessary to adjust the brightness of the light. The marking of the brightness change time point can be achieved according to the brightness gradient sequence data and the minimum brightness adaptation value. By comparing the brightness gradient value and the minimum brightness adaptation value at each time point, the time point when the brightness change exceeds the brightness adaptation threshold of this area can be determined, thereby determining the time point when it is necessary to adjust the brightness of the light. This helps to accurately control the adjustment timing of the light to meet the lighting requirements and provide a comfortable visual experience. The estimation of the time-sequence brightness correction model can be achieved by training a model on the marked brightness change data and change gradient features corresponding to the time, which helps to establish an accurate time-sequence brightness correction model for the time-sequence adjustment and control of the light.

[0125] Preferably, step S4 includes the following steps:

[0126] Predict the regional brightness change during personnel flow based on the lighting probability classification model, so as to obtain the flow brightness prediction data and record it in the factory building lighting system; adjust the brightness of the prediction area according to the factory building lighting system;

[0127] Predict the regional brightness correction during natural light change based on the time-sequence brightness correction model, so as to obtain the natural light correction time-sequence data and record it in the factory building lighting system; monitor the correction time according to the factory building lighting system, and when the correction time arrives, actively collect the lighting brightness of the light source irradiation area to obtain the current lighting brightness data, and adjust the current regional brightness according to the current lighting brightness data.

[0128] In the embodiment of the present invention, first, personnel flow data is obtained from the factory building lighting system, including information such as the number of personnel, moving direction, and moving speed in each area. The personnel flow data is used as input and a lighting probability classification model is used for prediction. The model obtains the adjacent next area where the personnel in each area are most likely to move based on the most obvious and strongly correlated moving direction characteristics of the personnel in each area, and obtains the lighting brightness of this area according to the number of this area. Specifically, the sensor needs to first monitor the movement of personnel in the lighting edge area of this area. The model receives the main data of the personnel movement sent by the sensor. This data can filter out the most important source data according to the selection of the important characteristics of the previous personnel movement. Then, the movement characteristics of this personnel and the corresponding correlation probability are analyzed. The model makes a prediction and judgment based on this. According to the previously obtained personnel flow distribution data, the system can set a threshold based on this data to automatically ensure the brightness of these areas during high personnel flow periods without continuous monitoring. Moreover, the present invention retains the functions of the existing lighting system and can not adjust the brightness within a specified period, so there is no need for the sensor to collect monitoring data. When the factory building lighting system monitors the natural light correction timing data obtained based on the timing brightness correction model prediction, the system collects the lighting brightness of the specified light source irradiation area through the sensor and adjusts it to the correct brightness. If the brightness of this area already meets the brightness requirement size of the current area preset by the system at this time, no brightness adjustment is performed.

[0129] The present invention can provide real-time personnel location and movement information by obtaining personnel movement data from the factory building lighting system. By monitoring data such as the number of personnel, movement direction, and movement speed in different areas of the factory building, the distribution and movement trends of personnel can be understood. This helps to predict and adjust lighting requirements based on personnel movement. Using the lighting probability classification model for prediction can predict lighting requirements based on personnel movement data. By analyzing the movement direction characteristics of different personnel in each area, the adjacent area where personnel in each area are most likely to move can be predicted, and the lighting brightness of this area can be obtained. This helps to determine the areas where lighting adjustments are needed based on personnel movement and provide appropriate lighting brightness. The main data for sensor monitoring of personnel movement can be achieved by screening out the most important source data. Based on the selection of important characteristics of previous personnel movement, the key data of personnel movement can be obtained, providing accurate input for the model. This helps to reduce the complexity of data processing while ensuring accurate prediction and judgment of the model on key data. Setting thresholds based on the pedestrian flow distribution data can set lighting thresholds according to personnel movement. By analyzing the pedestrian flow distribution data, the areas that need to be automatically adjusted during high pedestrian flow periods can be determined, and the brightness of these areas can be ensured. This helps to achieve automated lighting adjustment and improve the efficiency and response speed of the lighting system. Retaining the functions of the existing lighting system can ensure the compatibility of the present invention with the existing lighting system. Through the present invention, brightness adjustment is not performed within a specified period, so there is no need for sensors to continuously collect monitoring data. This helps to reduce the modification of the existing lighting system and retain its original functions and operation methods. The natural light correction timing data predicted based on the timing brightness correction model can collect and adjust the lighting brightness of the designated light source irradiation area according to the sensor. When the natural light correction timing data is monitored, the system can adjust according to the lighting brightness data collected by the sensor to ensure the correct brightness of the lights. This helps to achieve dynamic response and adjustment to natural light changes. If the brightness of the area already meets the brightness requirements preset by the system, no brightness adjustment is performed, which can avoid unnecessary lighting adjustments. By judging whether the brightness of the current area meets the requirements, frequent brightness adjustments can be avoided, saving energy and improving the stability of the system. In summary, through the steps in the above embodiments, personnel movement data can be used to predict and adjust lighting requirements. This will help to adjust the lighting brightness in real time according to the personnel location and movement, provide an efficient and intelligent lighting system, and reduce energy waste.

[0130] Preferably, the present invention also provides an intelligent factory building lighting control system for performing the intelligent factory building lighting control method as described above. The intelligent factory building lighting control system includes:

[0131] An attenuation map generation module, configured to collect the illumination brightness of the light source irradiation area in the intelligent factory building at multiple time periods, so as to obtain illumination brightness data; analyze the illumination area distribution of the illumination brightness data, so as to obtain illumination area distribution data; take pictures of the illumination area distribution data at multiple time periods and perform preprocessing, so as to obtain image preprocessing data; generate an attenuation map of the brightness within the area for the illumination area distribution data according to the image preprocessing data, so as to obtain brightness attenuation map data;

[0132] An illumination probability classification module, configured to match the personnel movement characteristics of adjacent illumination areas for the illumination area distribution data, so as to obtain a mobile detail matching data set; construct a Bayesian probability classification model for the mobile detail matching data set, so as to obtain an illumination probability classification model;

[0133] A timing correction modeling module, configured to perform dynamic analysis of the illumination brightness on the image preprocessing data according to the brightness attenuation map data, so as to obtain multi-dimensional brightness change data; perform timing correction modeling on the multi-dimensional brightness change data, so as to obtain a timing brightness correction model;

[0134] A regional brightness prediction module, configured to predict the regional brightness change during personnel flow based on the illumination probability classification model, so as to obtain flow brightness prediction data; perform regional brightness correction prediction during natural light change based on the timing brightness correction model, so as to obtain natural light correction timing data.

[0135] In summary, the present invention provides an intelligent factory building lighting control system, which is composed of an attenuation map generation module, an illumination probability classification module, a timing correction modeling module, and a regional brightness prediction module, and can implement any one of the intelligent factory building lighting control methods described in the present invention. By analyzing the images in the time and distance dimensions of the light source irradiation area including artificial light and other natural light according to the different distribution of the number of people, the movement law of personnel at the edge of the effective illumination area is found, so as to quickly judge the lighting requirements of personnel moving to the next illumination area in the busy intelligent factory building and make corresponding adjustments. And a timing model is established according to the brightness change law and brightness requirements in different regions at different times to ensure that the brightness of the area can be adjusted in time even when the light sensor responds slowly or fails, and the data can be summarized to solve possible fault problems in time. The internal structure of the system cooperates with each other, thereby simplifying the operation process of the intelligent factory building lighting control system.

[0136] Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to include all changes within the meaning and scope of the equivalent elements of the application documents in the present invention.

[0137] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent factory lighting control method, characterized in that: The following steps are involved: Step S1: collecting the illumination brightness of the light source illumination area in the intelligent factory in multiple time periods, thereby obtaining illumination brightness data; Perform lighting area distribution analysis on lighting brightness data to obtain lighting area distribution data; Capturing and preprocessing images of the lighting area distribution data at multiple time periods, thereby obtaining image preprocessing data; Generate a brightness attenuation spectrum of the illumination area distribution data in the area according to the image preprocessing data, thereby obtaining brightness attenuation spectrum data; Step S2: matching the personnel movement features of adjacent lighting areas with the lighting area distribution data, thereby obtaining a movement detail matching data set; constructing a Bayesian probability classification model for the movement detail matching data set, thereby obtaining a lighting probability classification model; Step S3: Performing dynamic analysis of illumination brightness on the image preprocessing data according to the brightness attenuation spectrum data, thereby obtaining multi-dimensional brightness change data; Perform time series correction modeling on the multi-dimensional brightness change data to obtain a time series brightness correction model; Step S4: predicting the regional brightness change when people flow based on the lighting probability classification model, thereby obtaining flow brightness prediction data; Based on the time series brightness correction model, regional brightness correction prediction is performed when natural light changes, so as to obtain natural light correction time series data; wherein step S4 includes the following steps: Based on the lighting probability classification model, the regional brightness change prediction when personnel flow is carried out, so as to obtain the flow brightness prediction data and record it in the factory lighting system; the brightness of the predicted area is adjusted according to the factory lighting system; Based on the time-series brightness correction model, the regional brightness correction prediction is carried out when the natural light changes, so as to obtain the natural light correction time-series data and record it in the factory lighting system; the correction time is monitored according to the factory lighting system, and when the correction time arrives, the lighting brightness of the light source illumination area is actively collected to obtain the current lighting brightness data, and the current area brightness is adjusted according to the current lighting brightness data.

2. The intelligent factory lighting control method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting the illumination brightness of the light source illumination area in the intelligent factory in multiple time periods, thereby obtaining illumination brightness data; Step S12: performing lighting area distribution analysis on the lighting brightness data, thereby obtaining lighting area distribution data; Step S13: photographing the illumination area distribution data in multiple time periods to obtain illumination image data; performing image preprocessing on the illumination image data to obtain image preprocessing data; Step S14: performing illumination edge detection on the image preprocessing data to obtain illumination edge data; Step S15: performing brightness measurement on the image preprocessing data according to the illumination edge data, thereby obtaining brightness measurement data; Step S16: performing local brightness attenuation analysis on the brightness measurement data to obtain a local brightness attenuation coefficient; Step S17: Generate an attenuation spectrum of the brightness in the area based on the illumination area distribution data according to the local brightness attenuation coefficient, so as to obtain brightness attenuation spectrum data.

3. The intelligent factory lighting control method according to claim 2, characterized in that: Step S2 includes the following steps: Step S21: capturing infrared portraits of different areas of the intelligent factory building using infrared sensors according to the lighting area distribution data, thereby obtaining infrared portrait movement data; Step S22: performing time series fluctuation analysis on the infrared portrait movement data to obtain human flow fluctuation data; Step S23: performing a crowd flow distribution analysis on the lighting area distribution data according to the crowd flow fluctuation data, thereby obtaining crowd flow distribution data; Step S24: labeling the crowd flow distribution data according to a preset crowd flow sensitivity threshold, thereby obtaining crowd flow distribution label data; Step S25: capturing the movement details of personnel in different areas of the smart factory building according to the human flow distribution label data and using the millimeter wave sensor to obtain a movement detail data set; Step S26: performing personnel movement feature matching of adjacent lighting areas on the movement detail data set and the infrared portrait movement data, thereby obtaining a movement detail matching data set; Step S27: performing lighting association marking on the lighting area distribution data according to the moving detail matching data set, thereby obtaining lighting association marking data; Step S28: constructing a Bayesian probability classification model for the lighting-related tag data, thereby obtaining a lighting probability classification model.

4. The intelligent factory lighting control method according to claim 3 is characterized in that: In step S28, the Bayesian probability classification model for the lighting-related tag data is constructed using the following formula: ； In the formula, is the probability of regional lighting adjustment, is the number of associated labeled human movement features, is the number of associated tagged human movement features of adjacent lighting areas, It is The personnel movement feature value of the associated mark and the The similarity of the personnel movement feature values ​​of the associated tags of adjacent lighting areas, is the mean of the personnel movement characteristics of the associated markers in the adjacent lighting areas, It is The importance weight value of the personnel movement feature associated with the marker.

5. The intelligent factory lighting control method according to claim 4, characterized in that: Step S26 includes the following steps: Step S261: performing important feature selection on the motion detail data set and the infrared portrait motion data, thereby obtaining motion angle feature data, motion speed feature data, and motion motion feature data; Step S262: using the FLANN algorithm to filter the moving angle feature data, the moving speed feature data, and the moving motion feature data by the feature ratio of adjacent lighting areas, thereby obtaining moving feature ratio data; Step S263: constructing a convolution model for the moving feature ratio data, thereby obtaining a convolution model; iteratively optimizing the convolution model using a RANSAC algorithm, thereby obtaining an optimized convolution model; Step S264: matching the movement features of people in adjacent lighting areas on the movement detail data set based on the optimized convolution model, thereby obtaining movement detail matching data.

6. The intelligent factory lighting control method according to claim 5, characterized in that: Step S3 includes the following steps: Step S31: performing linear correlation screening on the brightness attenuation spectrum data to obtain linear attenuation standard data; Step S32: performing linear correlation analysis on the brightness attenuation spectrum data according to the linear attenuation standard data, thereby obtaining linear brightness attenuation data; performing nonlinear correlation analysis of brightness interference on the brightness attenuation spectrum data according to the linear attenuation standard data, thereby obtaining nonlinear brightness attenuation data; Step S33: performing straight-line fitting on the linear brightness attenuation data and the nonlinear brightness attenuation data, thereby obtaining straight-line fitting attenuation data; Step S34: performing high-dimensional feature dimensionality reduction on the multi-dimensional linear attenuation data, thereby obtaining low-dimensional linear attenuation data; Step S35: constructing a time matrix for the low-dimensional linear attenuation data, thereby obtaining brightness attenuation matrix data; Step S36: performing dynamic analysis of illumination brightness on the image preprocessing data according to the brightness attenuation matrix data, thereby obtaining brightness multi-dimensional change data; Step S37: Perform time series correction modeling on the brightness multi-dimensional change data, thereby obtaining a time series brightness correction model.

7. The intelligent factory lighting control method according to claim 6, characterized in that: Step S36 includes the following steps: Step S361: performing HSV color space conversion on the image preprocessing data and extracting the brightness component to obtain brightness component data; Step S362: performing a brightness-distance attenuation cluster analysis on the brightness component data according to the brightness attenuation matrix data, thereby obtaining brightness-distance attenuation cluster data; performing a brightness-time attenuation cluster analysis on the brightness component data according to the brightness attenuation matrix data, thereby obtaining brightness-time attenuation cluster data; Step S363: performing brightness trend partitioning on the brightness component data according to the brightness-distance decay clustering data and the brightness-time decay clustering data, thereby obtaining brightness partitioning data; Step S364: performing gamma conversion on the brightness partition data, thereby obtaining gamma conversion partition data; Step S365: Perform dynamic analysis of lighting brightness on the gamma conversion partition data to obtain multi-dimensional brightness change data.

8. The intelligent factory lighting control method according to claim 7, characterized in that: Step S37 includes the following steps: Step S371: using the optical flow method to divide the brightness multi-dimensional change data into change gradients, thereby obtaining brightness change gradient data; Step S372: sorting the brightness change gradient data to obtain brightness gradient sequence data; Step S373: marking the time points of the brightness gradient sequence data according to the preset minimum brightness adaptation data, thereby obtaining brightness change time mark data; Step S374: Perform time series correction modeling on the brightness change time mark data to obtain a time series brightness correction model.

9. An intelligent factory lighting control system, characterized in that: Used to execute the intelligent plant lighting control method as claimed in claim 1, the intelligent plant lighting control system comprises: The attenuation spectrum generation module is used to collect the illumination brightness of the light source illumination area in the intelligent factory in multiple time periods, so as to obtain illumination brightness data; perform illumination area distribution analysis on the illumination brightness data, so as to obtain illumination area distribution data; perform image capture and preprocessing on the illumination area distribution data in multiple time periods, so as to obtain image preprocessing data; generate an attenuation spectrum of the brightness in the illumination area distribution data according to the image preprocessing data, so as to obtain brightness attenuation spectrum data; The lighting probability classification module is used to match the personnel movement characteristics of adjacent lighting areas with the lighting area distribution data, thereby obtaining a movement detail matching data set; and to construct a Bayesian probability classification model for the movement detail matching data set, thereby obtaining a lighting probability classification model; The time series correction modeling module is used to perform dynamic analysis of illumination brightness on the image preprocessing data according to the brightness attenuation spectrum data, so as to obtain multi-dimensional brightness change data; perform time series correction modeling on the multi-dimensional brightness change data, so as to obtain a time series brightness correction model; The regional brightness prediction module is used to predict the regional brightness changes when people flow based on the lighting probability classification model, so as to obtain the flow brightness prediction data; and to predict the regional brightness correction when natural light changes based on the time series brightness correction model, so as to obtain the natural light correction time series data.

Images