Low-power-consumption illumination control method and system based on multi-source information and behavior analysis

Abstract

The invention relates to a low-power-consumption illumination control method and system based on multi-source information and behavior analysis, and the method comprises the steps: information collection, behavior analysis, theoretical illuminance measurement and calculation, illumination intensity distribution field calculation, control strategy generation and execution, and the like. By combining a convolutional neural network algorithm, a space conversion algorithm, a fuzzy control strategy and the like, precise control of each lamp in a scene is realized, the illumination intensity and the lamp angle are dynamically adjusted, the safety of pedestrians and vehicles and the eye comfort are ensured, meanwhile, the power consumption of the lamp in an area without an illumination demand is reduced, and an energy-saving effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent lighting, and in particular to an intelligent lighting control method and system based on multi-source information fusion. Background Art

[0002] In the existing lighting control systems, a single control method is usually adopted, such as simple timing control or adjustment based on light intensity. However, these methods have many deficiencies. For example, it is impossible to accurately control the lights according to the dynamic changes of scene modes, time, weather, as well as people and vehicle flows, resulting in poor lighting comfort, high power consumption and difficulty in meeting the application requirements of large scenes. In addition, in emergency scenarios, such as a smoke environment in a fire, the existing lighting systems lack an effective guiding function and cannot provide clear escape path indications for people and vehicles.

[0003] Therefore, there is an urgent need for a method and system that can perform intelligent lighting control by integrating multiple factors to meet the needs of modern lighting. Summary of the Invention

[0004] The technical problems solved by the present invention: To overcome the above-mentioned deficiencies of the prior art, the present invention provides a low-power lighting control method and system based on multi-source information and behavior analysis. By fusing multi-source information data and behavior data, combining theoretical illuminance calculation and comprehensive training, accurate control of the lights is achieved, ensuring the balance of safety and lighting comfort, effectively reducing power consumption, and adapting to the application of large scenes. At the same time, it has a lighting guiding function in emergency scenarios.

[0005] The solution adopted by the present invention to solve its technical problems is as follows:

[0006] A low-power lighting control method based on multi-source information and behavior analysis, comprising the following steps:

[0007] Step S1: Collect multi-source information data, including scene mode, time, weather, and the light intensity at each monitoring point, including dynamically collecting the light intensity at the intersection boundary of the theoretical light range of each lamp;

[0008] Step S2: Use monitoring videos, radars, and infrared sensors to collect the behavior data of pedestrians and vehicles in the scene;

[0009] Step S3: Use the collected scene mode, time, weather, and light intensity data at each monitoring point, and combine with the scene model to calculate the theoretical illuminance at each spatial observation point in the scene;

[0010] Step S4: Using a specific target recognition algorithm, analyze the behavioral data of personnel and vehicle traffic in real time through the data collected by surveillance videos, radars, and infrared sensors, identify the high-activity areas and low-activity areas of people and vehicles, as well as the position, speed, trajectory, and action information of people and vehicles within the areas;

[0011] Step S5: Input the theoretical illuminance distribution data of each observation point in the scene and the illuminance at the handover boundary of the theoretical illumination range of each lamp into the convolutional neural network algorithm module for real-time training, and output the illuminance distribution field of each spatial point in the scene;

[0012] Step S6: Using a spatial transformation algorithm, with the identified target as the focus, combine the target behavior data to perform multi-focus illuminance distribution field conversion on the illuminance distribution. For the converted illuminance distribution field, with the perspective light intensity requirements and comfort of pedestrians and vehicles at different positions, different time periods, different densities, different trajectories, and different speeds as the goals, use a fuzzy strategy to calculate the control strategies for the angles and illuminance of each lamp in the scene;

[0013] Step S7: Use the control strategies to adjust the angles and illuminance of each lamp in the scene, and repeat the above steps to dynamically and real-time control the lights, ensuring that pedestrians and vehicles in the space always obtain safe and comfortable illuminance during movement, while the lamps in areas without lighting requirements remain at the lowest power consumption or in the off state, ultimately achieving the optimal balance between lighting requirements and power consumption.

[0014] Preferably, in Step S3, the theoretical illuminance of each observation point in the scene is calculated using the following formula:

[0015]

[0016] Where:

[0017] Ei: The theoretical illuminance of the i-th point in the scene;

[0018] Ij: The luminous intensity of the j-th lamp;

[0019] θij: The angle between the light ray from the j-th lamp to the i-th point and the optical axis of the lamp;

[0020] Pij: The luminous intensity distribution coefficient considering the optical characteristics of the lamp (such as the light distribution curve), representing the proportion of the luminous intensity of the j-th lamp in the direction of the i-th point;

[0021] dij: The distance from the j-th lamp to the i-th point;

[0022] Cscene: The scene mode correction coefficient, which is set according to the different lighting requirements of different scenes (such as roads, squares, parking lots, etc.);

[0023] Ctime: The time correction coefficient, which is adjusted according to the natural light changes in different time periods of a day (such as daytime, evening, night).

[0024] Cweather: The weather correction coefficient, which corrects the influence of weather conditions (such as sunny, cloudy, rainy, foggy, etc.) on light.

[0025] n: The total number of lamps affecting the illuminance at the i-th point.

[0026] Preferably, the light intensity distribution field at each point in the space in step S5 is calculated using the following formula:

[0027] I(x, y) = f CNN (E(x,y), L boundary (x, y))

[0028] Where:

[0029] I(x,y): The light intensity distribution field at the position (x,y) in the scene;

[0030] f CNN : The convolutional neural network mapping function, which is defined by the weight parameters W and the bias term b of the network;

[0031] E(x,y): The theoretical illuminance distribution data at each point in the scene;

[0032] L boundary (x,y): The light intensity data at the intersection boundary of the theoretical light ranges of each lamp.

[0033] The convolutional neural network mapping function f CNN The execution process of each layer is as follows:

[0034]

[0035] Where:

[0036] I (l) (x,y): The feature map of the l-th layer at the position (x,y);

[0037] I (l+1) (x,y): The feature map of the (l + 1)-th layer at the position (x,y);

[0038] W k (l) : The weight parameter of the k-th convolutional kernel of the l-th layer;

[0039] *: The convolution operator;

[0040] b (l) : The bias term of the l-th layer;

[0041] σ: Activation function, such as ReLU, Sigmoid, etc.;

[0042] K: The number of convolutional kernels in the l-th layer.

[0043] Preferably, the spatial conversion algorithm in step S6 adopts the following formula:

[0044]

[0045] Where:

[0046] I transformed (x, y): The light intensity at the position (x, y) in the converted multi-focus light intensity distribution field;

[0047] K: The number of foci of the spatial target behavior data (such as targets in different activity areas or different behavior patterns);

[0048] W behavior (k): The weight of the k-th focus behavior data, determined according to factors such as the activity intensity and density of the target in this area;

[0049] I(x, y): The light intensity at the position (x, y) in the original light intensity distribution field;

[0050] μ view (x, y, k): The perspective light intensity membership function value of the k-th focus at the position (x, y), reflecting the degree of conformity of the light intensity at this position to the k-th focus target.

[0051] Preferably, the control strategy in step S6 is calculated using the following formula:

[0052]

[0053] Where,

[0054] C(x, y): The lamp control strategy (including angle and light intensity) at the position (x, y);

[0055] c: Possible control strategy options;

[0056] argmax: Determine among all possible control strategies c, find the control strategy that makes the objective function reach the maximum value;

[0057] μ comfort (I transformed (x, y), k): The light comfort membership function value of the k-th focus at the position (x, y), indicating the degree of conformity of the light intensity at this position to the comfort of the k-th focus target.

[0058] The light comfort membership function is calculated using the following formula:

[0059]

[0060] Wherein:

[0061] μ comfort (I): Membership degree of comfort corresponding to the light intensity I;

[0062] Ic: Central value of the comfortable light intensity, i.e., the light intensity to which the human eye is most adapted;

[0063] k: Steepness parameter, controlling the steepness of the curve. The larger the value, the steeper the curve; this value is adjusted according to actual needs; a larger k value indicates more sensitivity to fluctuations in light intensity and a more rapid change in comfort; a smaller k value indicates a higher tolerance for fluctuations in light intensity and a gentler change in comfort.

[0064] Preferably, in step S8, in the event of a smoke emergency such as a fire, the light guiding function is activated, and through a specific light mode, clear escape path indications are provided for personnel and vehicles.

[0065] Preferably, the light state is adjusted according to the distance between the path point and the dangerous area:

[0066]

[0067] Wherein:

[0068] Ln represents the light state at the path point (x n , y n ). Green flashing indicates a safe path, and red flashing indicates a dangerous area.

[0069] (x n , y n ) is a light guiding path point.

[0070] (x d , y d ) is the central position of the dangerous area.

[0071] R d is the influence radius of the dangerous area.

[0072] ΔR is the safety buffer distance, used to ensure a certain distance between the path point and the dangerous area.

[0073] The present invention also provides a low-power lighting control system based on multi-source information and behavior analysis, including:

[0074] Multi-source information acquisition module: By arranging light intensity sensors in the scene, it acquires the light intensity at the intersection boundary of the predetermined monitoring points and the theoretical light range of the lamps; obtains weather information through a dedicated interface; obtains time information through a clock module; and obtains scene mode information through a scene recognition module.

[0075] Behavior data acquisition module: Through monitoring cameras, radars, and infrared sensors, it acquires the behavior data of people and vehicles in the scene in real time.

[0076] Theoretical light intensity calculation module: According to the acquired data and the scene model, it calculates the theoretical illuminance of each observation point in the scene using a formula.

[0077] Behavior data analysis module: It analyzes the behavior data using object recognition algorithms to identify high-activity areas, low-activity areas, and the specific positions, speeds, trajectories, and action information of people and vehicles.

[0078] Light intensity distribution field generation module: It inputs the theoretical illuminance distribution data of the observation points and the light intensity at the intersection boundary of the theoretical light range of the lamps into a convolutional neural network for real-time training and outputs a light intensity distribution field.

[0079] Control strategy generation module: Focusing on the identified targets and combining behavior data, it performs multi-focus light intensity distribution field conversion on the light intensity distribution, and uses a fuzzy strategy to calculate the control strategies for the angles and light intensities of each lamp based on the light intensity requirements and comfort objectives from the perspectives of pedestrians and vehicles.

[0080] Control execution module: According to the calculated control strategies, it sends control signals to the drive devices of each lamp through an intelligent lighting control system to achieve precise regulation of the lamps.

[0081] Preferably, the system further includes an emergency scenario processing module: By installing emergency situation monitoring devices such as smoke sensors and connecting them to the lighting control system, in case of an emergency, it plans the best escape route based on the lighting space layout deployed by the lighting system and the location of the safety exits, and provides clear instructions for people and vehicles by controlling the state of the lights.

[0082] Advantages of the present invention compared with the prior art:

[0083] (1) High-precision lighting control: By integrating multi-source information data and behavior data, combining theoretical illuminance calculation and comprehensive training, it can achieve high-precision control of the lights and meet the lighting requirements in different scenarios.

[0084] (2) Good light comfort: By comprehensively considering the positions, trajectories, and action information of people and vehicles, it determines a control strategy that balances light intensity and comfort, improving light comfort.

[0085] (3) Low power consumption and low cost: By adopting a lightweight deployment method and combining with a dynamic adjustment mechanism, it can reduce power consumption and cost while ensuring the lighting effect, and is particularly suitable for large-scale applications.

[0086] (4) Additional functions: It has the function of guiding lights in emergency scenarios, and can provide effective escape path instructions for personnel and vehicles in smoky environments such as fires, improving the safety and practicality of the system. Description of the Drawings

[0087] Figure 1 is a schematic structural diagram of an embodiment of the low-power lighting control system based on multi-source information and behavior analysis in the present invention.

[0088] Figure 2 is a schematic diagram of the low-power lighting control method based on multi-source information and behavior analysis in the present invention. Detailed Embodiments

[0089] To make the purpose, technical solutions and advantages of the implementation of the present invention clearer, the technical solutions in the embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in the embodiments of the present invention. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, of the embodiments of the present invention. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention. The technical solutions of the present invention will be further specifically described below with reference to the drawings, and an embodiment of the present invention is given.

[0090] An underground parking lot of a large commercial complex, which has a multi-layer structure, and each layer is divided into multiple areas, including an entrance area, an exit area, a parking area, a passage area, and an emergency evacuation area. The parking lot needs to control the lights according to different scene modes (such as day, night, peak hours, special events, etc.), time, weather, and the dynamic changes of people and vehicle flows, and at the same time provide effective light guidance in emergency scenarios.

[0091] The entire system is as Figure 1 shown and includes the following parts.

[0092] Multi-source information acquisition module: By arranging light intensity sensors in the parking lot, the light intensity at the predetermined monitoring points and the intersection boundaries of the theoretical light ranges of the lamps is collected; weather information is obtained through the weather forecast interface; time information is obtained through the clock module; and scene mode information is obtained through the scene recognition module.

[0093] Behavior data collection module: collects behavior data of people and vehicles in the scene in real time through monitoring cameras, radars and infrared sensors.

[0094] Theoretical light intensity calculation module: Based on the collected data and parking lot model, information such as entrances and exits, distribution of parking spaces, and natural light entrances is obtained from the model, and the theoretical illuminance of each pre-selected spatial observation point in the parking lot that is conducive to calculating the light distribution field is calculated using the formula.

[0095] Behavior data analysis module: Use target recognition algorithms to analyze behavioral data to identify high-activity areas, low-activity areas in the parking lot, and the specific location, speed, trajectory and movement information of pedestrians and vehicles.

[0096] Light intensity distribution field generation module: The theoretical illuminance distribution data of each observation point and the light intensity at the intersection of the theoretical illumination range of the lamp are input into the convolutional neural network for real-time training, and the light intensity distribution field is output.

[0097] Control strategy generation module: With the identified pedestrians and vehicles in the parking lot as the focus, the light intensity distribution is converted into a multi-focus light intensity distribution field, and the control strategy of each lamp angle and light intensity is calculated using the fuzzy strategy in combination with the visual intensity and comfort targets of pedestrians and vehicles.

[0098] Control execution module: According to the calculated control strategy, the intelligent lighting control system sends control signals to the driving devices of each lamp to achieve precise control of each lamp in the parking lot.

[0099] Emergency scenario processing module: By installing emergency monitoring equipment such as smoke sensors and connecting them to the lighting control system, in an emergency, the best escape route is planned based on the lighting space layout deployed by the lighting system and the location of the safe exit, and clear instructions are provided to people and vehicles by controlling the status of the lights.

[0100] After the system is installed and deployed in the parking lot, it operates through the following steps to achieve low-power control of space lighting.

[0101] Step S1, after surveying the parking lot, according to the spatial characteristics of the parking lot, lighting fixtures are installed at specific locations, as well as various sensors for monitoring the environment, light intensity, temperature, smoke, personnel, vehicles and other information in the parking lot, to collect multi-source information data;

[0102] Step S2, installing high-definition cameras, millimeter-wave radars, and pedestrian infrared detection sensors at the entrance, exit, passages, and parking areas of the parking lot;

[0103] Step S3, using the environmental characteristics of the parking lot, calculating the theoretical light intensity of each selected observation point in the parking lot;

[0104] Step S4, using the YOLO algorithm improved by deep learning, combined with the point cloud data of the suspected target detected by the millimeter wave radar, to identify the position, movement, speed, trajectory of the pedestrian, and the speed, trajectory, headlight brightness, etc. of the vehicle;

[0105] Step S5, inputting the light intensity value at the intersection boundary of the lighting range of the lamps arranged in the parking lot and the theoretical light intensity distribution of the selected observation point into the convolutional neural network algorithm for training, calculating the light intensity distribution field of each spatial position in the parking lot, and dynamically establishing a light intensity distribution model in the parking lot;

[0106] Step S6, based on the identified perspectives of pedestrians in the parking lot and vehicles entering the parking lot, calculate the field distribution conversion of the illumination distribution field at multiple perspectives and multiple locations under the required perspective, and calculate the real-time dynamic control strategy of each lamp in the parking lot based on the actual impact of the illumination intensity distribution on the eyes of pedestrians and drivers, with the lower limit of meeting the 10-meter visibility requirement and the upper limit of preventing glare as for pedestrians and vehicles;

[0107] Step S7, using the control strategy to dynamically adjust the brightness and angle of each lamp, and repeat the above steps to dynamically ensure the observation needs of pedestrians and drivers in the parking lot, and turn off lamps in some positions or reduce the brightness of lamps at the edge of the target, effectively reducing the power consumption of lamps in the parking lot.

[0108] In the emergency scene mode, the intelligent guidance function is activated. When the smoke sensor installed in the parking lot detects a fire on a certain floor of the parking lot, the emergency scene guidance module is immediately activated. The system uses a specific light mode to indicate the direction of the safe exit in a green flashing mode, and uses the lights near the fire area in a red flashing mode to warn people and vehicles to stay away from the dangerous area, providing people and vehicles with clear escape path instructions and guiding them to evacuate quickly and safely.

[0109] Parts of the present invention that are not described in detail are common knowledge in the art.

[0110] Through the above embodiments, the purpose of the present invention is fully and effectively achieved. Those skilled in the art can understand that the present invention includes but is not limited to the contents described in the drawings and the above specific embodiments. Although the present invention has been described with respect to the most practical and preferred embodiments currently considered, it should be understood that the present invention is not limited to the disclosed embodiments, and any modification that does not deviate from the functional and structural principles of the present invention will be included in the scope of the claims.

Claims

1. A low-power lighting control method based on multi-source information and behavior analysis, characterized in that, It includes the following steps: Step S1: Collect multi-source information data, including scene mode, time, weather, and the light intensity at each monitoring point, including dynamically collecting the light intensity at the intersection boundary of the theoretical light ranges of each lamp; Step S2: Use surveillance videos, radars, and infrared sensors to collect the behavior data of pedestrians and vehicles in the scene; Step S3: Use the collected scene mode, time, weather, and light intensity data at the monitoring points, and combine with the scene model to calculate the theoretical illuminance at each spatial observation point in the scene; Step S4: Use a specific target recognition algorithm to analyze the behavior data of personnel and vehicle flows in real time through the data collected by surveillance videos, radars, and infrared sensors, and identify the high-activity areas, low-activity areas of people and vehicles, as well as the position, speed, trajectory, and action information of people and vehicles within the areas; Step S5: Input the theoretical illuminance distribution data of each observation point in the scene and the light intensity at the intersection boundary of the theoretical light ranges of each lamp into the convolutional neural network algorithm module for real-time training, and output the light intensity distribution field of each spatial point in the scene; Step S6: Use the spatial transformation algorithm, with the identified target as the focus, combine with the target behavior data to perform multi-focus light intensity distribution field conversion on the light intensity distribution, and for the converted light intensity distribution field, with the perspective light intensity requirements and comfort of pedestrians and vehicles at different positions, different time periods, different densities, different trajectories, and different speeds as the goal, use the fuzzy strategy to calculate the control strategies for the angles and light intensities of each lamp in the scene; Step S7: Use the control strategies to adjust the angles and light intensities of each lamp in the scene, and repeat the above steps to dynamically and real-time control the lights, ensuring that pedestrians and vehicles in the space always obtain a safe and comfortable light intensity during movement, while the lamps in the areas without light demand maintain the lowest power consumption or are in the off state, ultimately achieving the optimal balance between light demand and power consumption.

2. The low-power lighting control method for multi-source information and behavior analysis according to claim 1, wherein In step S3, the calculation of the theoretical light intensity at each observation point in the scene adopts the following formula: Where: Ei: The theoretical illuminance at the i-th point in the scene; Ij: The luminous intensity of the j-th lamp; θij: The angle between the light ray from the j-th lamp to the i-th point and the optical axis of the lamp; Pij: The luminous intensity distribution coefficient considering the optical characteristics of the lamp (such as the light distribution curve), representing the proportion of the luminous intensity of the j-th lamp in the direction of the i-th point; dij: The distance from the j-th lamp to the i-th point; Cscene: The scene mode correction coefficient, which is set according to the different light requirements of different scenes (such as roads, squares, parking lots, etc.); Ctime: The time correction coefficient, which is adjusted according to the natural light changes at different time periods of a day (such as daytime, evening, night); Cweather: The weather correction coefficient, which corrects the influence of weather conditions (such as sunny, cloudy, rainy, foggy, etc.) on the light; n: The total number of lamps affecting the illuminance at the i-th point.

3. The low-power lighting control method for multi-source information and behavior analysis according to claim 1, characterized in that, The control strategy in step S6 is calculated using the following formula: Where, C(x,y): The control strategy (including angle and light intensity) of the lamp at position (x,y); c: Possible control strategy options; argmax: Determine the control strategy that maximizes the objective function among all possible control strategies c. μ comfort (I transformed (x, y), k): The value of the illumination comfort membership function for the k-th focus at the position (x, y), representing the degree of compliance of the illumination intensity at that position with the comfort of the k-th focus target.

4. The low-power lighting control method for multi-source information and behavior analysis according to claim 3, characterized in that The membership function of lighting comfort is calculated using the following formula: Where: μ comfort (I): Membership degree of comfort corresponding to light intensity I Ic: The central value of the comfortable lighting intensity, i.e., the lighting intensity most adaptable to the human eye. k: The steepness parameter that controls the steepness of the curve. The larger the value, the steeper the curve.

5. The low-power lighting control method for multi-source information and behavior analysis according to claim 1, characterized in that In step S8, in the event of a smoke emergency scenario such as a fire, the lighting guidance function is activated, and through a specific lighting mode, clear escape path indications are provided for personnel and vehicles.

6. A low-power lighting control system based on multi-source information and behavior analysis, characterized in that, Including: Multi-source information acquisition module: By arranging lighting intensity sensors in the scene, the lighting intensity at the intersection boundary of the predetermined monitoring points and the theoretical lighting range of the lamps is collected. Obtain weather information through a dedicated interface; obtain time information through a clock module; Obtain scene mode information through a scene recognition module. Behavior data acquisition module: Through monitoring cameras, radars, and infrared sensors, the behavior data of personnel and vehicles in the scene are collected in real time. Theoretical lighting intensity calculation module: According to the collected data and the scene model, the theoretical illuminance of each observation point in the scene is calculated using a formula. Behavior data analysis module: Use target recognition algorithms to analyze the behavior data, and identify high-activity areas, low-activity areas, and the specific positions, speeds, trajectories, and action information of personnel and vehicles. Lighting intensity distribution field generation module: Input the theoretical illuminance distribution data of each observation point and the lighting intensity at the intersection boundary of the theoretical lighting range of the lamps into a convolutional neural network for real-time training, and output the lighting intensity distribution field. Control strategy generation module: Focusing on the spatial target behavior data, perform multi-focus lighting intensity distribution field conversion on the lighting intensity distribution, and combine the perspective light intensity and comfort target of pedestrians and vehicles, and use fuzzy strategies to calculate the control strategies for the angles and lighting intensities of each lamp. Control execution module: According to the calculated control strategy, send control signals to the driving devices of each lamp through an intelligent lighting control system to achieve precise regulation of the lamps.

7. The low-power lighting control system based on multi-source information and behavior analysis according to claim 6, characterized in that, It also includes an emergency scenario processing module. By installing emergency situation monitoring devices such as smoke sensors and connecting them to the lighting control system, in case of an emergency, based on the lighting space layout deployed by the lighting system and the location of the safety exits, the best escape path is planned, and clear indications are provided for personnel and vehicles by controlling the state of the lights.

Images