Lighting energy-saving control system with self-learning function in public indoor space

Abstract

The invention belongs to the field of indoor public place illumination control, and particularly discloses an illumination energy-saving control system with a self-learning function in a public indoor space, which is designed to capture illumination intensity change data of an indoor environment and personnel distribution image information of each indoor area. The collected illumination intensity change data and personnel distribution image information are sent to a main control unit; sending the illumination intensity change data and the personnel distribution image information to a cloud platform; based on a data analysis algorithm and an image recognition technology, the uploaded illumination intensity change data and personnel distribution image information are deeply compared and analyzed to recognize a real-time change mode of an illumination demand and personnel distribution, and based on an analysis result, a light adjustment instruction is generated and corresponding adjustment operation is executed, so that indoor light is adjusted in real time, and the user experience is improved. It is ensured that the illumination intensity meets user requirements, unnecessary energy waste is avoided, and therefore high efficiency and energy saving are achieved.

Description

Technical Field

[0001] The invention relates to the field of indoor public place lighting control, in particular to a lighting energy-saving control system with a self-learning function in a public indoor space. Background Art

[0002] Existing indoor electric light sources are designed according to the needs of lighting at night or during the day, and rely on manual opening and closing. The light source is not adjustable and cannot be adjusted in real time according to the light intensity of natural light sources such as the sun and the moon and the situation of indoor occupants. This is especially true for large indoor public places such as conference centers, exhibition halls, libraries, etc., resulting in huge waste of electricity.

[0003] Although the existing technology proposes to automatically adjust the switch status of lamps or lighting brightness by monitoring the light intensity of the current environment and detecting the human body through a camera, the existing way of detecting light intensity cannot accurately perceive complex lighting environments, especially in environments with mixed light sources (natural light and artificial light); the existing personnel detection technology (such as infrared sensors, cameras) cannot accurately identify the location and number of people, especially in high-density crowds or complex background environments; they often lack adaptive control algorithms and cannot be dynamically adjusted according to historical data and real-time data, resulting in the system's performance in different scenarios not being smart enough; the existing technology has a slow response speed, and the system may take a long time to respond after detecting environmental changes, affecting the user experience; at the same time, the existing technology does not propose that it can be applied to large indoor public places. The above reasons have led to its failure to be applied in reality.

[0004] In view of this, the present invention provides a lighting energy-saving control system with a self-learning function for a public indoor space to solve the above-mentioned problem. Summary of the invention

[0005] The purpose of the present invention is to provide a lighting energy-saving control system with a self-learning function for public indoor spaces, which can realize intelligent adjustment of indoor light brightness changes by integrating advanced automation control technology to achieve optimal energy-saving effects.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a lighting energy-saving control system with a self-learning function for a public indoor space, including a data acquisition module, a main control unit, a grid lighting control unit and a cloud platform, wherein each module is connected by wire or wirelessly;

[0008] The data acquisition module captures the light intensity change data of the indoor environment and the image information of the personnel distribution in each area of ​​the room, and sends the uploaded light intensity change data and personnel distribution image information to the main control unit;

[0009] The main control unit converts the light intensity change data and the personnel distribution image information into a unified standard and stores them safely, and sends the standard formatted light intensity change data and personnel distribution image information to the cloud platform;

[0010] Grid lighting control unit: divide the indoor space into grids according to the indoor space distribution characteristics and the installation positions of the lamps, and number each grid; the lighting equipment in each grid is independently controlled;

[0011] The cloud platform, based on data analysis algorithms and image recognition technology, conducts in-depth comparative analysis of the uploaded light intensity change data and personnel distribution image information to identify the changing patterns of lighting demand and personnel activities, and sends annotation signals to the main control unit based on the changing patterns. The main control unit generates control instructions based on the signals and sends them to the lighting control module.

[0012] As a preferred technical solution of the first aspect of the present invention, it also includes:

[0013] The lighting control module, after receiving the control command from the main control unit, interacts with the lighting control module through the communication interface and converts the command into a lighting control signal for controlling the light; the lighting control module executes the lighting control signal to realize intelligent adjustment of the indoor lighting system;

[0014] User interaction module, the user remotely or manually sends a lighting control signal to control the lighting, and realizes the lighting effect through the lighting control module.

[0015] As a preferred technical solution of the first aspect of the present invention, the logic for acquiring the light intensity change data is:

[0016] The photosensitive sensor is used to capture the light intensity information of the indoor environment and convert the light intensity information in the environment into an analog electrical signal. The analog electrical signal is sent to the main control unit through the signal transmission link and converted into digital light intensity information based on the main control unit.

[0017] Compare and analyze the light intensity information with the preset light intensity threshold issued by the cloud platform. When the light intensity information exceeds or falls below the preset light intensity threshold, it is marked in the grid. Combined with the content marked during the personnel distribution detection, the main control unit will quickly generate the corresponding control instructions and send them to the lighting control module through the communication interface.

[0018] After receiving the control signal from the main control unit, the lighting control module performs corresponding lighting adjustment operations according to the instruction content.

[0019] As a preferred technical solution of the first aspect of the present invention, the logic for acquiring the image information of personnel distribution in each area of ​​the room is as follows:

[0020] Preload the indoor public space map in advance, and record the spatial layout characteristics of each area based on the indoor public space map;

[0021] On the indoor public space map, mark the installation locations of all lamps, and reasonably divide the grid areas on the indoor public space map according to the distribution of lamps and spatial layout characteristics. Each grid is effectively monitored and independently controlled;

[0022] Based on the spatial layout characteristics, high-definition cameras are reasonably installed to analyze and count the number of people in each grid area in real time, and different states are classified and processed according to the statistical results;

[0023] If a grid is unoccupied for a long time, the corresponding position of the grid area on the indoor public space map is identified and the idle state is marked by comparing the layout features of the actual area with the map;

[0024] If the number of people in the grid exceeds the preset threshold, the corresponding position of the grid area on the indoor public space map is identified and marked as a warning state by comparing the layout features of the actual area with the map;

[0025] When the number of people is between 0 and the threshold, no comparison is performed and the current state is kept unmarked.

[0026] As a preferred technical solution of the first aspect of the present invention, the data analysis algorithm includes an adaptive control algorithm, and the adaptive control algorithm includes the following steps:

[0027] Collect historical data, including indoor and outdoor light intensity, indoor human flow dynamics, lighting system power supply voltage and switch status information;

[0028] Preprocess the historical data to obtain the characteristics of light intensity changes and human flow activity changes;

[0029] An adaptive control algorithm is constructed, and the characteristics of light intensity change and human flow activity change are used as input factors of the adaptive control algorithm. The output factor of the adaptive control algorithm is the predicted lighting brightness.

[0030] Through an iterative training process, the algorithm gradually identifies and internalizes the underlying rules and patterns in the data; the algorithm learns and predicts changes in lighting needs under different modes to achieve lighting adjustment tasks.

[0031] As a preferred technical solution of the first aspect of the present invention, the data analysis algorithm includes real-time monitoring and anomaly detection, including the following steps:

[0032] Compare the current lighting parameters with the lighting parameter range under the normal operation mode. If the lighting parameters are within the lighting parameter range, the current lighting system is normal.

[0033] If the lighting parameters are not within the lighting parameter range, the current lighting system is abnormal;

[0034] The abnormal cause of the lighting system is determined based on the data analysis algorithm, and countermeasures are taken based on the abnormal cause to restore the normal operation of the lighting system.

[0035] As a preferred technical solution of the first aspect of the present invention, the data analysis algorithm includes a feedback mechanism:

[0036] Collect the brightness monitoring data and energy consumption data of the lighting area fed back by users in actual applications; determine the key parameters and control strategies that need to be optimized through continuous iteration and optimization of algorithms;

[0037] The optimized key parameters and control strategies are deployed to the actual lighting system for verification, and the brightness monitoring data and energy consumption data of the lighting area are monitored until the changes in the brightness monitoring data and energy consumption data of the lighting area meet the convergence conditions.

[0038] In a second aspect, the present invention provides a lighting energy-saving control method with a self-learning function in a public indoor space, based on the implementation of the first aspect, comprising the following steps:

[0039] Capture the light intensity change data of indoor and outdoor environments and the image information of personnel distribution in various areas of the room, and send the uploaded light intensity change data and personnel distribution image information to the main control unit;

[0040] The light intensity change data and the personnel distribution image information are processed through data conversion, formatted into a unified standard and stored securely, and the standard formatted light intensity change data and personnel distribution image information are sent to the cloud platform;

[0041] Based on data analysis algorithms and image recognition technology, an in-depth comparative analysis is performed on the uploaded light intensity change data and personnel distribution image information to identify the changing patterns of lighting requirements and personnel activities, and precise control instructions are sent to the main control unit based on the changing patterns.

[0042] In a third aspect, the present invention provides an electronic device comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0043] The processor executes the first aspect by calling a computer program stored in the memory.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the first aspect.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] The multiple lighting modes and flexible control methods provided by the present invention fully consider the balance between energy conservation and environmental protection and user experience in design. By accurately controlling the operating status and light intensity of the lighting equipment, while meeting the user's demand for light brightness, unnecessary energy consumption is minimized to achieve dual optimization of energy conservation and emission reduction and user experience. In addition, it also supports remote management and control functions. Managers can remotely access the system interface to monitor the system in real time, adjust parameters, and maintain and manage the system. The convenience and flexibility of the system are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a structural framework diagram of the lighting energy-saving control system of the present invention;

[0048] Figure 2 It is a flow chart of indoor personnel situation detection of the present invention;

[0049] Figure 3 is a flow chart of the lighting energy-saving control method of the present invention;

[0050] Figure 4 is a schematic diagram of grid lighting and detection;

[0051] Figure 5 It is a structural diagram of the adaptive algorithm. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] In the description of the present invention, it should be noted that the terms "vertical", "up", "down", "horizontal", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0054] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] This embodiment integrates photosensitive sensors and camera recognition technology, which can sense the intensity of indoor and outdoor light and the distribution of people indoors in real time. Based on these precise data, the system can intelligently adjust the power supply voltage and start and stop status of lighting equipment in each area, and achieve coordinated and optimized control of the overall lighting fixtures. While meeting the basic needs of users for light brightness, it minimizes unnecessary lighting energy consumption and achieves the goal of efficient use of light energy and energy conservation and emission reduction.

[0056] Example 1: Figure 1 As shown, this embodiment provides a lighting energy-saving control system with a self-learning function for a public indoor space, including a data acquisition module, a main control unit, a lighting control module, a user interaction module and a cloud platform, and each module is connected by wire or wirelessly.

[0057] The data acquisition module captures the light intensity change data of the indoor environment and the image information of the personnel distribution in each area of ​​the room, and sends the uploaded light intensity change data and personnel distribution image information to the main control unit;

[0058] It should be noted that photosensitive sensors are used to accurately capture changes in light intensity in the indoor environment. Changes in light intensity include natural light sources and artificial lighting electric light sources. At the same time, combined with camera recognition technology, high-definition image recognition and positioning of the distribution of people in various areas of the room are performed.

[0059] The main control unit processes the light intensity change data and the personnel distribution image information through data conversion, formats them into a unified standard and stores them safely, and sends the standard formatted light intensity change data and personnel distribution image information to the cloud platform.

[0060] It should be noted that the collected light intensity change data and personnel distribution image information are uploaded to the main control unit with ESP32 as the core in real time through an efficient data transmission protocol. The ESP32 main control unit has powerful data processing capabilities and achieves seamless connection with the cloud platform through Wi-Fi and Bluetooth communication functions. As the center of data processing and decision-making, the cloud platform uses advanced data analysis algorithms to analyze factors such as ambient light and personnel activities in real time, and generates optimized control instructions accordingly. These instructions are quickly conveyed to each lamp through the ESP32 sensor, realizing precise adjustment of light intensity.

[0061] The cloud platform, based on data analysis algorithms and image recognition technology, conducts in-depth comparative analysis of the uploaded light intensity change data and personnel distribution image information to identify the changing patterns of lighting requirements and personnel activities, and sends precise grid identification to the main control unit based on the changing patterns.

[0062] It should be noted that: based on intelligent control methods, it automatically adapts to the lighting needs of different scenarios, reduces manual intervention, improves response speed and accuracy, and thus improves user experience. By optimizing lighting strategies, unnecessary lighting waste is reduced, achieving significant energy-saving and emission reduction effects.

[0063] Specifically, within a grid, when the light intensity is greater than the preset maximum threshold, the lamps are turned off regardless of whether there is anyone in the grid; when the light intensity is less than the preset maximum threshold, if there is no one in the grid, the lamps are turned off, and if there is someone in the grid, the lamps are turned on and adjusted to maximum brightness; when the light intensity is between the preset maximum threshold and the minimum threshold, if there is no one in the area, the lamps are turned off, and if the number of people in the area is greater than the preset threshold, the lamps are turned on and adjusted to maximum brightness; if the number of people in the area is between 0 and the maximum threshold, the lamps are not adjusted. At this time, the lighting brightness fully meets user needs.

[0064] Build an adaptive control algorithm to identify the changing patterns of light intensity and personnel distribution, and obtain potential problems of lighting equipment based on the changing patterns;

[0065] Specifically, the adaptive control algorithm based on machine learning analyzes the collected data to identify the changing patterns of light intensity and personnel distribution. Based on these analysis and identification results, the algorithm continuously learns and iterates to optimize the control model. In addition, the algorithm also has anomaly detection and fault diagnosis functions, which can timely discover and deal with potential problems of lighting equipment based on real-time and historical data, and improve the stability and reliability of system operation.

[0066] After receiving the control command, the lighting control module interacts with the lighting control module through the communication interface and converts the command into a lighting control signal for controlling the light; the lighting control module executes the lighting control signal to realize intelligent adjustment of the indoor lighting system.

[0067] User interaction module, the user remotely or manually sends a lighting control signal to control the lighting, and realizes the lighting effect through the lighting control module.

[0068] This embodiment uses camera recognition technology to accurately monitor the distribution and number of people in each area of ​​the room. Through grid-based area division, the indoor space is subdivided into multiple independently controllable areas, and photosensitive devices are installed in each area. This refined area division and lighting control strategy enables the system to be independently adjusted according to the actual needs of different areas, improving the accuracy and efficiency of lighting control.

[0069] Clearly understand the current situation of public indoor spaces being generally equipped with surveillance systems, and through technical upgrades, directly utilize or enhance the functions of existing cameras to enable them to have personnel detection and analysis capabilities. This not only significantly reduces the economic cost of system deployment, but also maximizes the use of existing surveillance resources, avoids duplication and waste of resources, and embodies a new paradigm of intelligent lighting management that is efficient, economical, and environmentally friendly.

[0070] An adaptive control algorithm is introduced to analyze and learn the collected data, and key data such as light intensity and personnel activities in public indoor spaces are collected in real time as input to the adaptive control algorithm. Machine learning technology is used to extract patterns from historical data and combine them with real-time data to achieve fine-grained control of lighting parameters, including switch status, brightness, color temperature, etc. By adjusting these parameters, the system can minimize energy consumption while meeting lighting needs. At the same time, the system can automatically adjust the lighting strategy according to the lighting needs of different scenes (such as conference rooms, corridors, rest areas, etc.) to continuously optimize the control strategy, automatically identify and adapt to the changing patterns of light intensity and personnel distribution, thereby improving control accuracy and energy-saving effects; compared with traditional lighting systems, this embodiment can significantly reduce energy consumption. This energy-saving effect not only helps to reduce operating costs, but also helps to promote green and sustainable development.

[0071] The lighting control system of this embodiment is flexibly designed and can adapt to a variety of indoor scenes and lighting requirements. For large commercial spaces or public buildings, traditional manual or centralized lighting control methods are often inefficient and difficult to meet personalized needs. Remote monitoring and centralized management are implemented according to actual needs, which greatly improves management efficiency and reduces maintenance costs. Managers can view the lighting status of each area in real time through the system interface and perform remote control to ensure that the lighting environment of the entire space is always in the best state.

[0072] Example 2: The parts not described in detail in this example are as shown in Example 1, please refer to Figure 2 This embodiment provides a detailed structural diagram of indoor population distribution monitoring:

[0073] An algorithm is created in the cloud platform, the algorithm comprising:

[0074] At the cloud platform level, the indoor public space map is preloaded in advance, and the spatial layout characteristics of each area are recorded based on the indoor public space map. In view of the dynamic changes in spatial layout characteristics, a regular or triggered map update mechanism is established to ensure the timeliness and accuracy of the data.

[0075] On the indoor public space map, mark the installation locations of all lamps, and reasonably divide the grid areas on the indoor public space map according to the distribution of lamps and spatial layout characteristics. Each grid is effectively monitored and independently controlled, such as Figure 4 shown.

[0076] High-definition cameras are reasonably installed based on the spatial layout characteristics so that all divided grid areas can be clearly covered by the cameras without blind spots. The video streams collected by the cameras are uploaded to the cloud platform in real time to provide basic data for subsequent data processing and analysis.

[0077] First, set a reasonable threshold for the number of people in each area based on the characteristics of the venue (such as functional areas, crowd density, etc.). Perform real-time analysis on the uploaded video data and count the number of people in each grid area.

[0078] According to the statistical results, different states are classified and processed:

[0079] If a grid is unoccupied for a long time, the corresponding position of the grid area on the map is identified and the idle state is marked by comparing the layout features of the actual area with the map;

[0080] If the number of people in the grid exceeds the preset threshold, the warning status of the corresponding position of the grid area on the map is identified by comparing the layout features of the actual area with the map;

[0081] When the number of people is between 0 and the threshold, no comparison is performed and the current state is kept unmarked.

[0082] Example 3: The parts not described in detail in this example are as shown in Example 1, please refer to Figure 3 , this embodiment provides a detailed structural diagram of an indoor light detection module:

[0083] Using multi-band photosensors, which are capable of detecting multiple wavelength ranges, the intensity of natural light and artificial light can be measured separately. Installing multiple photosensors at different locations in the indoor space (based on a grid distribution) can capture changes in light intensity in different areas.

[0084] The photosensor is used to capture the light intensity information of the indoor environment and convert it into an analog electrical signal. The analog electrical signal is sent to the ESP32 main control unit through the signal transmission link. Its built-in analog-to-digital converter (ADC) module performs a high-precision and high-efficiency analog-to-digital signal conversion process, converting the analog signal into a digital signal format and storing it in a non-volatile memory for subsequent data processing and analysis.

[0085] Using advanced data processing and machine learning technology, preset light intensity threshold parameters are pre-set and optimized. And compared with the light data captured by the sensor, processed and uploaded by the main control unit. Once it is detected that the light intensity of a specific indoor area exceeds or falls below the preset light intensity threshold, it will be marked on the preset grid map: if the light intensity in the grid is less than the preset minimum threshold, it will be marked as "turn on the light"; if the light intensity in the grid is greater than the preset maximum threshold, it will be marked as "turn off the light"; if the light intensity in the grid is between the minimum threshold and the maximum threshold, it will be marked according to the distribution of people in the room. If the distribution of people is marked as "idle state", it will be marked as "turn off the light"; if the distribution of people is marked as "warning state", it will be marked as "turn on the light"; if there is no marking for the distribution of people, the lighting equipment status at this time will remain unchanged and no marking will be made. According to the marking, the microcontroller will quickly generate the corresponding control instructions and send them to the lighting control module through the communication interface.

[0086] After receiving the control signal from the main control unit, the lighting control module performs corresponding lighting adjustment operations according to the instruction content.

[0087] According to the results of the algorithm analysis of Embodiment 2 and Embodiment 3, the main control unit generates corresponding control instructions:

[0088] The "turn off lights" label signal generates a command to turn off the lights;

[0089] The "turn on the light" label signal generates a command to turn on the light and adjust it to the highest brightness;

[0090] If there is no marked signal, the lamp will not be adjusted.

[0091] After receiving the control instruction from the main control unit, the lighting control module performs the corresponding lighting adjustment operation according to the instruction content.

[0092] During the implementation process, the system should continuously collect user feedback and operation data to evaluate the lighting control effect, and continuously optimize the algorithm and control strategy accordingly. Regular system maintenance should be carried out, including updating maps, adjusting camera positions, optimizing algorithm parameters, etc., to ensure the long-term stable operation of the system and meet actual needs.

[0093] Example 4: The parts not described in detail in this example are as shown in Example 1, please refer to Figure 5 This embodiment provides an adaptive control algorithm that analyzes historical data, learns and predicts different modes. It includes the following steps:

[0094] Extensive historical data collection covers indoor and outdoor light intensity under different time periods and weather conditions, indoor human flow dynamics, lighting system power supply voltage and its switch status. Subsequently, these data are uploaded to the algorithm system built into the cloud platform, and in-depth analysis is performed with the help of advanced adaptive control algorithms: 1. In the data preprocessing stage, the system will eliminate outliers, for example, identify and exclude maximum or minimum values ​​in indoor photosensor data that may be caused by errors. 2. Extract characteristic variables that are of practical significance for subsequent analysis from the original data set, including but not limited to the changing trends of indoor and outdoor light intensity in different time spans (days, weeks, months, years), and the regular patterns of indoor human flow activities. At the same time, the system also integrates past experience data on manual adjustment of lighting brightness in special situations and uploads it to the algorithm.

[0095] Self-learning algorithms can analyze historical data, learn and predict different patterns. The first step of the system is to collect historical data extensively, covering indoor and outdoor light intensity under different time periods and weather conditions, indoor traffic dynamics, lighting system power supply voltage and its switch status. Subsequently, these data are uploaded to the algorithm system built into the cloud platform, and in-depth analysis is carried out with the help of advanced self-learning algorithms: 1. In the data preprocessing stage, the system will remove outliers. For example, statistical methods (such as Z-score or IQR) can be used to detect and remove outliers to identify and exclude the maximum or minimum values ​​in the indoor photosensor data that may be caused by errors. For missing data, interpolation methods (such as linear interpolation, polynomial interpolation) or methods such as mean and median can be used to fill in. Finally, all data are standardized to make data from different sources comparable. Commonly used methods include Min-Max standardization and Z-score standardization.

[0096] 2. Extract feature variables that are of practical significance for subsequent analysis from the original data set, for example: analyze the relationship between features and target variables through the correlation matrix, and select features with higher correlation. Use machine learning models (such as random forests, XGBoost) to perform feature importance analysis and select features that contribute most to model prediction. Use dimensionality reduction techniques such as PCA (principal component analysis) to reduce feature dimensions and improve model training efficiency and generalization ability. Including but not limited to the changing trends of indoor and outdoor light intensity in different time spans (days, weeks, months, years), and the regular patterns of indoor human flow activities. At the same time, the system also integrates past experience data on manual adjustment of lighting brightness in special situations and uploads it to the algorithm.

[0097] 3. Use this information to build an accurate algorithm model.

[0098] <1> Choose the appropriate model: Linear regression: suitable for simple linear relationships. Decision tree: suitable for nonlinear relationships. Random forest: suitable for complex nonlinear relationships. Support vector machine (SVM): suitable for high-dimensional data, with good classification and regression capabilities. Neural network: suitable for complex nonlinear relationships, with strong fitting capabilities.

[0099] <2> Training the model: 1. Dataset division: Divide the data set into training set, validation set and test set, usually in a ratio of 70%:15%:15%. 2. Model training: Use the training set to train the model, and use the cross-validation method to evaluate the generalization ability and prediction accuracy of the model. 3. Hyperparameter tuning: Use grid search, random search and other methods to optimize the model's hyperparameters and improve the model's performance.

[0100] <3> Optimize the model: Use hyperparameter tuning: Grid search: Search for the best hyperparameter combination through exhaustive search. Random search: Search for the best hyperparameter combination through random sampling. Bayesian optimization: Use Bayesian optimization method to efficiently search for the best hyperparameter combination.

[0101] By stacking multiple models, the prediction accuracy of the model can be improved.

[0102] <4> Real-time prediction: Collect data such as light intensity, personnel distribution, and equipment status in each area in real time. Preprocess the real-time data, such as standardization and feature extraction. Input the processed data into the trained model for real-time prediction.

[0103] <5> According to the prediction results, corresponding control instructions are generated, such as increasing or decreasing the lighting brightness, turning off or turning on lamps, etc. The control instructions are sent to the lighting control module through the communication interface to realize the intelligent adjustment of the indoor lighting system.

[0104] Through training, the algorithm can gradually identify and internalize the potential rules and patterns in the data. For example, during sunny daytime hours, if there is sufficient light outside and few people indoors, the lighting demand will naturally decrease; on the contrary, when it is rainy or night falls, even if the indoor density of people is not high, it may be necessary to increase the lighting intensity to meet basic needs; or at fixed times of the day, it is necessary to uniformly turn off or turn on the lights. Relying on the algorithm's efficient learning and accurate prediction capabilities, the above lighting adjustment tasks can be automated, thereby significantly improving the system's intelligence level and energy efficiency management efficiency.

[0105] Build an adaptive control algorithm and use this information to build an accurate algorithm model. Through an iterative training process, the algorithm can gradually identify and internalize the potential rules and patterns in the data. For example, during sunny daytime hours, if there is sufficient light outside and the indoor flow of people is sparse, the lighting demand will naturally decrease; on the contrary, when it is rainy or night falls, even if the indoor flow density is not high, it may be necessary to increase the lighting intensity to meet basic needs; or at fixed times of the day, it is necessary to uniformly turn off or turn on the lights. Relying on the algorithm's efficient learning and accurate prediction capabilities, the above lighting adjustment tasks can be automated, thereby significantly improving the system's intelligence level and energy efficiency management efficiency.

[0106] Adaptive control algorithms can monitor the operating status of the lighting system in real time and identify abnormal situations.

[0107] The adaptive control algorithm extracts features useful for abnormality detection, such as the changing trend of lighting brightness and the stability of equipment status, from the received normal historical data (including lighting brightness in each area, equipment status, such as switch status, power supply voltage, etc., and possible other related parameters, such as indoor crowd density, outdoor light intensity, etc.), and establishes a benchmark model for normal operation.

[0108] The operating status of the lighting system is also monitored in real time based on the uploaded data and compared with the normal operating mode. When the data detected deviates significantly from the normal operating mode, the algorithm will identify it as an abnormality. For example, if the lighting brightness in a certain area suddenly decreases or increases, and this change is not within the normal adjustment range, the algorithm will regard it as an abnormal situation.

[0109] And based on the abnormal situation, determine the cause of the abnormality, such as equipment failure, human error, etc.

[0110] If an abnormal situation is determined, the adaptive control algorithm will immediately take countermeasures to restore the normal operation of the lighting system. For example, if a lighting device failure is detected, resulting in reduced brightness, the algorithm can automatically adjust the brightness of other devices to compensate for the impact of the faulty device;

[0111] If the fault cannot be repaired immediately or requires human intervention, the adaptive control algorithm will notify the manager to perform maintenance by sending an alarm, such as sending the cause of the abnormal information judgment to the manager's mobile device or monitoring system interface.

[0112] The adaptive control algorithm also includes a feedback mechanism: 1. Collect feedback data: (1) Establish a user feedback platform or interface to facilitate users to submit their opinions and suggestions on lighting effects at any time. (2) Use sensors to monitor the brightness of the lighting area in real time and compare the actual lighting effect with the preset target value. (3) Collect statistics on the energy consumption data of the lighting system, including power consumption, power factor, etc., analyze the energy-saving effect of the lighting system, and compare it with industry standards or historical data. 2. Determine the key parameters and control strategies that need to be optimized based on the collected and analyzed data 3. Adjust the parameters and model structure of the adaptive control algorithm. Introduce new control strategies or optimization algorithms to improve the control accuracy and response speed of the system. 4. Deploy the optimized algorithm to the actual lighting system for verification. Monitor the system's operating status and user feedback to ensure the effectiveness of the algorithm in actual applications. 5. Continuously iterate and optimize the algorithm based on feedback and evaluation results in actual applications.

[0113] Embodiment 5: For the parts not described in detail in this embodiment, please refer to the description of Embodiment 1. This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0114] The processor executes the above-mentioned edge computing underwater target detection method by calling the computer program stored in the memory.

[0115] Embodiment 6: This embodiment provides a computer-readable storage medium storing instructions. When the instructions are executed on a computer, the computer executes the above-mentioned edge computing method for underwater target detection.

[0116] It should also be understood by those skilled in the art that the various exemplary devices, modules, units and steps described in conjunction with this patent disclosure can be implemented as electronic hardware, computer software or a combination of the two. In order to clearly illustrate this interchangeability of hardware and software, a general description has been given of the functions of various illustrative systems, modules, units and steps. Whether such functions are implemented as software or hardware depends on the specific application and the design constraints imposed on the entire system. Those skilled in the art can implement the functions in various ways for each specific application, but such implementation decisions should not be interpreted as causing a departure from the scope disclosed in the embodiments of the present invention.

[0117] The various exemplary devices, modules, units, and steps described in conjunction with the disclosure herein may be implemented or performed using the following components designed to perform such functions: a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination of these components. A general purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP, and / or any other such configuration.

[0118] The steps of the method or algorithm described in conjunction with the disclosure herein may be directly contained in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0119] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0120] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A lighting energy-saving control system with self-learning function for public indoor spaces, characterized by: It includes data acquisition module, main control unit, grid lighting control unit and cloud platform, and each module is connected by wire or wireless; The data acquisition module captures the light intensity change data of the indoor environment and the image information of the personnel distribution in each area of ​​the room, and sends the uploaded light intensity change data and personnel distribution image information to the main control unit; The main control unit converts the light intensity change data and the personnel distribution image information into a unified standard and stores them safely, and sends the standard formatted light intensity change data and personnel distribution image information to the cloud platform; Grid lighting control unit: divide the indoor space into grids according to the indoor space distribution characteristics and the installation positions of the lamps, and number each grid; the lighting equipment in each grid is independently controlled; The cloud platform, based on data analysis algorithms and image recognition technology, conducts in-depth comparative analysis of the uploaded light intensity change data and personnel distribution image information to identify the changing patterns of lighting demand and personnel activities, and sends annotation signals to the main control unit based on the changing patterns. The main control unit generates control instructions based on the signals and sends them to the lighting control module.

2. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 1 is characterized in that: Also includes: Lighting control module: After receiving the control command from the cloud platform, the main control unit interacts with the lighting control module through the communication interface and converts the command into a lighting control signal for controlling the lights; the lighting control module executes the lighting control signal to realize intelligent adjustment of the indoor lighting system; User interaction module, the user remotely or manually sends a lighting control signal to control the lighting, and realizes the lighting effect through the lighting control module.

3. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 1 is characterized in that: The logic for obtaining the image information of personnel distribution in each area of ​​the room: Preload the indoor public space map on the cloud platform, and record the spatial layout characteristics of each area based on the indoor public space map; On the indoor public space map, mark the installation locations of all lamps, and reasonably divide the grid areas on the indoor public space map according to the distribution of lamps and spatial layout characteristics. Each grid is effectively monitored and independently controlled; Based on the spatial layout characteristics, high-definition cameras are reasonably installed to analyze and count the number of people in each grid area in real time. According to the statistical results, different states are classified and processed: If a grid is unoccupied for a long time, the corresponding position of the grid area on the indoor public space map is identified and the idle state is marked by comparing the layout features of the actual area with the map; If the number of people in the grid exceeds the preset threshold, the corresponding position of the grid area on the indoor public space map is identified and marked as a warning state by comparing the layout features of the actual area with the map; When the number of people is between 0 and the threshold, no comparison is performed and the current state is kept unmarked.

4. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 1 is characterized in that: The logic for obtaining the light intensity change data: Install multiple light sensors at different locations in the indoor space according to the grid distribution to capture the changes in light intensity corresponding to the combined light intensity of natural light and artificial light in different areas; Based on the captured light intensity information of the indoor environment, the light intensity information is converted into an analog electrical signal, which is sent to the main control unit through a signal transmission link and converted into digital light intensity information by the main control unit. Compare and analyze the digitized light intensity information with the maximum and minimum light intensity thresholds preset on the cloud platform, and mark different analysis results: When the light intensity exceeds the preset maximum light intensity threshold, it is marked as "turn off the light"; When the light intensity is lower than the preset minimum light intensity threshold and the personnel distribution is not marked as "non-idle state", it is marked as "turn on the light"; When the light intensity is between the maximum threshold and the minimum threshold, check whether there is a label for the personnel distribution. If the personnel distribution is labeled as "idle state", label "turn off the light"; if the personnel distribution is labeled as "warning state", label "turn on the light"; if there is no label for the personnel distribution, keep the lighting device status unchanged and do not label it. The annotation is sent to the main control unit, which will quickly generate corresponding control instructions: control instructions include light off instructions, light on instructions, and light adjustment instructions; The "turn off lights" signal generates a command to turn off the lamps; The "turn on light" signal generates a command to turn on the light and adjust it to the highest brightness; After receiving the control instruction from the main control unit, the light control module performs the corresponding light adjustment operation according to the content of the control instruction.

5. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 4 is characterized in that: The data analysis algorithm includes an adaptive control algorithm, and the adaptive control algorithm includes the following steps: Collect historical data, including indoor light intensity, indoor human flow dynamics, lighting system power supply voltage and switch status information; Preprocess the historical data to obtain the characteristics of light intensity changes and human flow activity changes; An adaptive control algorithm is constructed, and the characteristics of light intensity change and human flow activity change are used as input factors of the adaptive control algorithm. The output factor of the adaptive control algorithm is the predicted lighting brightness. Through an iterative training process, the algorithm gradually identifies and internalizes the underlying rules and patterns in the data; the algorithm learns and predicts changes in lighting needs under different modes to achieve lighting adjustment tasks.

6. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 5 is characterized in that: The data analysis algorithm includes real-time monitoring and anomaly detection, including the following steps: Compare the current lighting parameters with the lighting parameter range under the normal operation mode. If the lighting parameters are within the lighting parameter range, the current lighting system is normal. If the lighting parameters are not within the lighting parameter range, the current lighting system is abnormal; The abnormal cause of the lighting system is determined based on the data analysis algorithm, and countermeasures are taken based on the abnormal cause to restore the normal operation of the lighting system.

7. The lighting energy-saving control system with self-learning function for public indoor spaces according to claim 6 is characterized in that: The data analysis algorithm includes a feedback mechanism: Collect the brightness monitoring data and energy consumption data of the lighting area fed back by users in actual applications; determine the key parameters and control strategies that need to be optimized through continuous iteration and optimization of algorithms; The optimized key parameters and control strategies are deployed to the actual lighting system for verification, and the brightness monitoring data and energy consumption data of the lighting area are monitored until the changes in the brightness monitoring data and energy consumption data of the lighting area meet the convergence conditions.

8. A lighting energy-saving control method with a self-learning function for a public indoor space, based on the implementation of a lighting energy-saving control system with a self-learning function for a public indoor space according to any one of claims 1 to 7, characterized in that: The following steps are involved: Capture the light intensity change data of the indoor environment and the image information of the personnel distribution in each area of ​​the room, and send the uploaded light intensity change data and personnel distribution image information to the main control unit; The light intensity change data and the personnel distribution image information are processed through data conversion, formatted into a unified standard and stored securely, and the standard formatted light intensity change data and personnel distribution image information are sent to the cloud platform; Based on data analysis algorithms and image recognition technology, an in-depth comparative analysis is performed on the uploaded light intensity change data and personnel distribution image information to identify the changing patterns of lighting requirements and personnel activities, and precise control instructions are sent to the main control unit based on the changing patterns.

9. An electronic device comprising: include: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor executes the lighting energy-saving control system with self-learning function for public indoor spaces as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

10. A computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the lighting energy-saving control system with self-learning function for a public indoor space as claimed in any one of claims 1 to 7.

Images