Multi-user lighting demand dynamic energy-saving matching method based on deep multi-modal behavior recognition

Abstract

This application relates to the field of intelligent lighting technology, and in particular to a dynamic energy-saving matching method for multi-user lighting needs based on deep multimodal behavior recognition. The method includes: acquiring a multimodal collaborative sensing information set; analyzing the coupling conflict characteristics between multi-user behavioral light needs and plant light needs based on the multimodal collaborative sensing information set to obtain a behavior-light coupling conflict information set; analyzing the energy consumption increase and plant growth interference risk caused by dynamic lighting based on the behavior-light coupling conflict information set to obtain a growth-energy consumption coupling assessment information set; analyzing the adjustment information for minimizing total energy consumption under the dual constraints of satisfying the plant light needs and the multi-user behavioral light needs based on the growth-energy consumption coupling assessment information set to obtain a lighting adjustment strategy; and adaptively adjusting lighting parameters based on the lighting adjustment strategy and outputting a dynamic energy-saving matching log for multi-user lighting needs. This ensures normal plant growth and meets the activity needs of multiple users.

Description

Technical Field

[0001] This application relates to the field of intelligent lighting technology, and in particular to a method for dynamic energy-saving matching of multi-user lighting needs based on deep multimodal behavior recognition. Background Technology

[0002] In public and semi-public spaces such as botanical gardens and exhibition greenhouses, artificial lighting is an important facility to ensure the healthy growth of plants and create an ornamental environment. Its energy consumption is the core component of operating costs. Current control strategies mainly involve fixed photoperiod lighting to meet the basic physiological needs of plants, or simple scene mode switching.

[0003] However, in complex scenarios where multiple users such as tourists and operators coexist, existing lighting control systems struggle to meet the needs of all parties and result in high energy consumption. Firstly, to meet the comfortable and high color rendering index lighting required by tourists, the system often continuously provides light intensity and spectrum exceeding the plant's growth requirements, leading to direct energy waste. Secondly, it cannot distinguish and dynamically respond to the inspection and maintenance needs of operators and the sightseeing needs of tourists, often operating in a global high-brightness mode. Thirdly, the rigid strategy fails to achieve fine-grained on-demand adjustment, resulting in a large amount of ineffective lighting time. This leads to a management dilemma in real-world scenarios with multiple roles and strong dynamics, where it is difficult to balance experience, growth, and energy consumption. Summary of the Invention

[0004] This application provides a dynamic energy-saving matching method for multi-user lighting needs based on deep multimodal behavior recognition to solve the above-mentioned problems. The method includes: acquiring a multimodal collaborative sensing information set; analyzing the coupling conflict characteristics between multi-user behavioral light needs and plant light needs based on the multimodal collaborative sensing information set to obtain a behavior-light coupling conflict information set; analyzing the energy consumption increase and plant growth interference risk caused by dynamic lighting based on the behavior-light coupling conflict information set to obtain a growth-energy consumption coupling assessment information set; analyzing the adjustment information for minimizing total energy consumption under the dual constraints of satisfying the plant light needs and the multi-user behavioral light needs based on the growth-energy consumption coupling assessment information set to obtain a lighting adjustment strategy; and adaptively adjusting lighting parameters based on the lighting adjustment strategy and outputting a dynamic energy-saving matching log for multi-user lighting needs.

[0005] Through the above technical solutions, the dynamic conflict between the light requirements of people and plants in mixed scenarios such as smart greenhouses is resolved intelligently, ensuring that the needs of both parties are met in a coordinated manner, improving the efficiency of space functions, and significantly reducing direct energy consumption and avoiding crop losses due to improper lighting under the dual constraints of ensuring plant growth and human visual comfort through optimized decision-making. This achieves deep and comprehensive energy saving, and the adaptive adjustment mechanism enhances the ability to respond to dynamic changes, reduces human intervention, and the generated logs provide data support for performance traceability and continuous strategy optimization, promoting closed-loop improvement and sustainable operation.

[0006] Optionally, the step of analyzing the coupling conflict characteristics between multi-user behavioral light requirements and plant light requirements based on the multimodal collaborative sensing information set to obtain a behavior-light coupling conflict information set includes: the multimodal collaborative sensing information set includes multi-user behavioral information, plant canopy light distribution information, and ambient light background information; based on the multi-user behavioral information and combined with the ambient light background information, analyzing the behavioral types of users in different spatial regions and their corresponding dynamic light requirements to obtain a multi-user behavioral dynamic light requirement information set; based on the plant canopy light distribution information and combined with the ambient light background information, analyzing the physiological light requirements of plants and their current actual light exposure status to obtain a plant dynamic light requirement information set; and based on the multi-user behavioral dynamic light requirement information set and combined with the plant dynamic light requirement information set, analyzing the differences and overlaps in light requirements between multi-users and plants in the spatial and temporal dimensions to obtain the behavior-light coupling conflict information set.

[0007] Optionally, the process of constructing the multi-user behavior dynamic light demand information set includes: based on the multi-user behavior information and combined with the ambient light background information, analyzing the interaction information between different user actions and plant layout and channel structure to obtain multi-user behavior intention recognition information; based on the multi-user behavior intention recognition information, analyzing the spectral preference of the lighting required to support different user behavior intentions relative to the background light to obtain a multi-user activity-spectral illuminance demand mapping table; based on the multi-user activity-spectral illuminance demand mapping table, analyzing the dominant activity type that needs to be prioritized in each lighting sub-area at the current moment and the corresponding dynamic light demand of the activity to obtain the multi-user behavior dynamic light demand information set.

[0008] Optionally, the process of constructing the multi-user behavior intent recognition information includes: based on the multi-user behavior information and combined with the ambient light background information, analyzing the coordination information between the user's behavior trajectory and spatial movement pattern and the plant layout and the channel structure under ambient light background to obtain user behavior spatial interaction features; based on the user behavior spatial interaction features, analyzing the influence of different regional lighting conditions on user dwell tendency and action precision to obtain user behavior trends; based on the user behavior trends, inferring the specific type and purpose of each user's current dominant activity to obtain the multi-user behavior intent recognition information.

[0009] Optionally, the process of constructing the plant dynamic light demand information set includes: based on the plant canopy light distribution information and combined with the ambient light background information, analyzing the actual light intensity and spectral composition of each region of the plant canopy to obtain a canopy light environment measured information set; based on the canopy light environment measured information set, and according to pre-stored plant species and growth stage information, analyzing the spectral intensity and light duration benchmarks required by different plants at each growth stage to obtain plant stratified and time-based physiological light demand benchmarks; based on the plant stratified and time-based physiological light demand benchmarks, analyzing the spectral ratio, spectral intensity, and spectral action period of supplemental artificial light to meet the needs of plant growth to obtain the plant dynamic light demand information set.

[0010] Optionally, based on the multi-user behavior dynamic light demand information set and combined with the plant dynamic light demand information set, the analysis of the differences and overlaps in light demands between multiple users and plants in spatial and temporal dimensions yields the behavior-light coupling conflict information set. This includes: based on the dominant activity type, the corresponding dynamic light demand, and combined with the spectral ratio and spectral intensity, analyzing the magnitude of the deviation and the possibility of complementarity between the spectral intensity required to meet the user's dominant activity and the spectral intensity required to meet the plant's physiological light demand within each of the lighting sub-regions, thus obtaining spatial light demand coupling information; based on the activity-corresponding dynamic light demand and combined with the spectral action period, analyzing the matching degree and contradictions between user activity light demand and plant light demand in terms of duration, activation sequence, and change frequency within a similar time period, thus obtaining temporal light demand coupling information; and based on the spatial light demand coupling information and combined with the temporal light demand coupling information, analyzing the feasibility level of simultaneously meeting the user's and plant's light demands, the required degree of compromise, and the potential impact of prioritizing one side, thus obtaining the behavior-light coupling conflict information set.

[0011] Optionally, the step of analyzing the energy consumption increase and plant growth disturbance risk caused by dynamic lighting based on the behavior-lighting coupling conflict information set to obtain a growth-energy consumption coupling assessment information set includes: based on the spatial light demand coupling information, analyzing the artificial light illumination power required to bridge the deviation level to obtain spatial conflict incremental energy consumption information; based on the temporal light demand coupling information, analyzing the extended artificial light illumination duration required to coordinate the matching degree and contradiction point to obtain temporal conflict incremental energy consumption information; based on the feasibility level and combined with the required compromise degree, analyzing the deviation degree of prioritizing user-led activities from the plant's stratified and time-based physiological light demand benchmark to obtain plant growth disturbance risk assessment information caused by lighting adjustments; and based on the spatial conflict incremental energy consumption information, the temporal conflict incremental energy consumption information, and combined with the plant growth disturbance risk assessment information, analyzing the energy consumption cost and plant growth risk of different priorities to obtain the growth-energy consumption coupling assessment information set.

[0012] Optionally, the analysis of energy consumption costs and plant growth risks for different priorities to obtain the growth-energy consumption coupling assessment information set includes: based on the feasibility level, analyzing several lighting scheme options that can be implemented to simultaneously meet the plant light requirements and the multi-user behavior light requirements, obtaining a multi-scheme feasibility option set; based on the multi-scheme feasibility option set, combined with the spatial conflict incremental energy consumption information and the temporal conflict incremental energy consumption information, analyzing the estimated additional energy consumption of each lighting scheme option to achieve the corresponding lighting coverage, obtaining a multi-scheme incremental energy consumption index set; based on the multi-scheme feasibility option set, combined with the plant growth interference risk assessment information, analyzing the potential negative impact of light environment adjustments on plant growth rate and physiological indicators of each lighting scheme option, obtaining a multi-scheme plant growth risk index set; based on the multi-scheme incremental energy consumption index set, combined with the multi-scheme plant growth risk index set, analyzing the comprehensive advantages and disadvantages of several lighting scheme options in terms of energy consumption and growth risk within the required compromise range, obtaining the growth-energy consumption coupling assessment information set.

[0013] Optionally, the step of analyzing the adjustment information to minimize total energy consumption under the dual constraints of satisfying the plant light demand and the multi-user behavior light demand, based on the growth-energy consumption coupling assessment information set, to obtain the lighting adjustment strategy includes: analyzing the potential negative impact of the lighting scheme options on plant growth based on the multi-scheme plant growth risk index set, and screening all lighting scheme options with potential negative impact levels lower than a preset risk threshold to obtain a safe lighting scheme candidate set; based on the safe lighting scheme candidate set, combined with the multi-scheme incremental energy consumption index set, analyzing the additional energy consumption corresponding to each candidate lighting scheme option, and screening out the candidate scheme with the minimum additional energy consumption to obtain the lowest energy consumption safe scheme; based on the lowest energy consumption safe scheme, combined with the required compromise degree and the feasibility level, analyzing the feasibility of dynamically responding to user-led activities by fine-tuning local spectral intensity or temporarily extending the lighting duration of some areas while maintaining the core lighting parameters of the lowest energy consumption safe scheme to minimize energy consumption, to obtain the lighting adjustment strategy.

[0014] Optionally, the step of adaptively adjusting lighting parameters based on the lighting adjustment strategy and outputting a dynamic energy-saving matching log for multi-user lighting needs includes: generating direct control instructions for programmable lighting devices based on the core lighting parameters, the fine-tuning of local spectral intensity, and the temporary extension of lighting duration in certain areas, to drive the lighting lamps to perform corresponding adjustments, thereby obtaining strategy execution information; analyzing the impact of light changes on the plant canopy light distribution information and the multi-user behavior information based on the lighting adjustment strategy and the multi-modal collaborative sensing information set, thereby obtaining effect verification information; the effect verification information includes at least the trend of plant photosynthetic rate change and the trend of user behavior fluency change; and generating and outputting the dynamic energy-saving matching log for multi-user lighting needs by associating the current time stamp, the dominant activity type, and the plant stratified and time-division physiological light demand benchmark based on the strategy execution information and the effect verification information. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0017] Figure 2A flowchart of a multi-user lighting demand dynamic energy-saving matching method based on deep multimodal behavior recognition provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0020] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0021] In smart agricultural greenhouses, vertical greening leisure areas, and other environments where multiple users coexist with plants, existing lighting control has significant shortcomings. It is difficult to balance the needs of multiple parties and the energy consumption is too high. In order to ensure the viewing experience of tourists, the light intensity and spectrum of the lighting often exceed the needs of plant growth, resulting in energy waste. At the same time, it cannot distinguish between the needs of tourists and the needs of operators for inspection, and it often operates in a global high-brightness mode. Moreover, the control strategy is rigid and cannot be adjusted in a fine-grained manner as needed, resulting in a large amount of ineffective lighting time.

[0022] Based on this, this application provides a dynamic energy-saving matching method for multi-user lighting needs based on deep multimodal behavior recognition. This method intelligently resolves the conflict between the lighting needs of people and plants in mixed scenarios such as greenhouses, taking into account both plant growth and human visual comfort, reducing energy consumption, avoiding crop losses, adaptively adjusting to reduce human intervention, and using log data to support strategy optimization and sustainable operation.

[0023] Figure 1 This application provides an illustration of an application scenario. When multiple users are in a space containing plants, the method provided in this application can significantly reduce direct energy consumption and avoid crop losses due to improper lighting, thereby achieving deep integrated energy saving. The adaptive adjustment mechanism enhances the response capability to dynamic changes.

[0024] Specifically, the method provided in this application can be applied to any server, where the server interacts with the sensor hardware network to obtain a multimodal collaborative sensing information set provided by the sensor hardware network, intelligently resolves the dynamic conflict between the light needs of people and plants in mixed scenarios such as smart greenhouses, and outputs a dynamic energy-saving matching log of multi-user lighting needs to the lighting maintenance personnel, thereby achieving deep integrated energy saving.

[0025] For specific implementation details, please refer to the following examples.

[0026] Figure 2 This is a flowchart illustrating a method for dynamic energy-saving matching of multi-user lighting needs based on deep multimodal behavior recognition, provided in an embodiment of this application. The method of this embodiment can be applied to servers in the above scenarios. Figure 2 As shown, the method includes:

[0027] S201. Obtain a multimodal collaborative sensing information set. Based on the multimodal collaborative sensing information set, analyze the coupling conflict characteristics between multi-user behavior light demand and plant light demand to obtain a behavior-light coupling conflict information set.

[0028] The multimodal collaborative sensing information set can be a multi-dimensional sensing data set of multi-user behavioral states and plant growth states, with a sensor hardware network as the data source. Multi-user behavioral light requirements can be the needs of different users for light intensity, color temperature, and illumination duration during specific activities. Plant light requirements can be the light conditions required for plant growth and development. Coupling conflict features can be the contradictory attributes of multi-user behavioral light requirements and plant light requirements in terms of light parameters. The behavior-light coupling conflict information set can be a structured information set formed by integrating the above coupling conflict features.

[0029] Specifically, in the current process of multiple users coexisting with plants (such as smart greenhouse office areas, home gardening leisure areas, shared green plant meeting rooms, etc.), existing lighting methods have significant technical pain points. The light requirements of human activities and plant growth are fundamentally different and often change dynamically. Existing lighting control is either based only on simple human presence sensors or preset fixed scenes. Its fundamental defect is the lack of precise perception and quantitative identification of "demand conflicts".

[0030] S202. Based on the behavior-lighting coupling conflict information set, analyze the risk of energy consumption increase and plant growth disturbance caused by dynamic lighting, and obtain the growth-energy consumption coupling assessment information set.

[0031] Dynamic lighting can be a lighting method that adjusts lighting parameters in real time according to actual needs. Energy consumption increase can be the increase in energy consumption that may result from parameter adjustments during dynamic lighting. Plant growth disturbance risk can be the possibility that dynamic lighting parameters will have an adverse effect on plant growth when they do not meet the plant's light requirements. The growth-energy consumption coupling assessment information set can be a set of information formed after comprehensively assessing the energy consumption increase and the plant growth disturbance risk.

[0032] Specifically, after identifying light conflicts, adjusting lighting in a single dimension can easily lead to systemic risks. By elevating the energy-saving goal to maximizing resource efficiency and establishing an "energy consumption-growth" coupled assessment model, two major harms can be avoided: one-sided energy saving suppresses lighting, damaging the growth of high-economic-value crops and causing huge losses; one-sided demand satisfaction exacerbates energy waste, and optimal decision-making across the entire life cycle and all factors of production can be achieved.

[0033] S203. Based on the growth-energy consumption coupling assessment information set, analyze the adjustment information to minimize total energy consumption under the dual constraints of satisfying plant light demand and multi-user behavior light demand, and obtain the lighting light adjustment strategy.

[0034] Dual constraints can be restrictions that simultaneously satisfy the light requirements of multiple users and the light requirements of plants. Minimizing total energy consumption can be the goal of minimizing energy consumption under dual constraints. Lighting adjustment strategies can be lighting parameter adjustment schemes formulated to achieve the minimization of total energy consumption.

[0035] Specifically, after completing conflict identification and risk assessment, generating an execution plan is the core decision-making step. Existing rule-based reasoning or simple weighted methods cannot guarantee global optimality in complex, dynamic scenarios with hard constraints. This can lead to problems such as rigid strategies, locally optimal decisions, and difficulty in quantifying trade-offs. Introducing operations research optimization methods can formalize multiple constraints and objectives, search for optimal solutions, and is the core intelligent hub for achieving precise energy saving.

[0036] S204. Based on the lighting adjustment strategy, adaptively adjust the lighting parameters and output a dynamic energy-saving matching log of multi-user lighting needs.

[0037] Lighting parameters can be adjustable indicators of lighting light. A multi-user lighting demand dynamic energy-saving matching log can record the adaptive adjustment process of lighting parameters, demand matching status, and energy consumption data.

[0038] Specifically, in the process of dynamic optimization of intelligent lighting, it is crucial to achieve closed-loop decision-making, ensure execution effectiveness, and establish an optimization and iterative data cycle. Without it, the strategy and execution will be disconnected, and the control will fail due to delays and errors caused by manual operation. It will also make the process untraceable, unable to review, locate anomalies, or accumulate data for evolution. Adaptive control technology ensures that the strategy is implemented in milliseconds, and structured logs provide a basis for system optimization.

[0039] The method provided in this embodiment intelligently resolves the dynamic conflict between the light requirements of people and plants in mixed scenarios such as smart greenhouses, ensuring that the needs of both parties are met collaboratively, improving the efficiency of space functions, and significantly reducing direct energy consumption and avoiding crop losses due to improper lighting under the dual constraints of ensuring plant growth and human visual comfort through optimized decision-making. This achieves deep and comprehensive energy saving, enhances the ability to respond to dynamic changes through adaptive adjustment mechanisms, reduces human intervention, and provides data support for performance traceability and continuous strategy optimization through generated logs, promoting closed-loop improvement and sustainable operation.

[0040] In some embodiments, the multimodal collaborative sensing information set includes multi-user behavior information, plant canopy light distribution information, and ambient light background information. Based on the multi-user behavior information and combined with the ambient light background information, the behavior types of users in different spatial regions and their corresponding dynamic light demands are analyzed to obtain a multi-user behavior dynamic light demand information set. Based on the plant canopy light distribution information and combined with the ambient light background information, the physiological light demands of plants and their current actual light exposure status are analyzed to obtain a plant dynamic light demand information set. Based on the multi-user behavior dynamic light demand information set and combined with the plant dynamic light demand information set, the differences and overlaps in light demands between multi-users and plants in the spatial and temporal dimensions are analyzed to obtain a behavior-light coupling conflict information set.

[0041] Multi-user behavior information can include information on user actions, movement trajectories, and areas of stay within a space. Plant canopy light distribution information can characterize the light intensity and spectral composition of different parts of the plant canopy. Ambient light background information can include the intensity, spectral characteristics, and trends of natural or basic artificial lighting within the space. Multi-user behavior dynamic light demand information can be a set of information that clarifies the spectral preferences and illuminance requirements of different users in different activity scenarios. Plant dynamic light demand information can be a set of information on the spectral ratio, intensity, and duration of supplemental artificial light required by plants.

[0042] Specifically, in smart agricultural greenhouses, vertical greening leisure areas, and other environments where multiple users coexist with plants, relying solely on single information or directly adjusting lighting can lead to inhibited plant growth and a poor user experience due to the spatiotemporal differences between user behavioral light requirements (such as strong light for precise operations) and plant physiological light requirements (such as soft light for seedlings). Furthermore, redundant or insufficient lighting can result in energy waste and even operational safety hazards, failing to provide a valid basis for subsequent energy consumption and growth risk assessments. To address these issues, the system first relies on the synchronous acquisition and fusion of multi-source heterogeneous sensing data: a deployed depth camera network captures the user's skeletal keypoint sequence, combined with the thermal distribution of infrared sensors, and a pre-trained spatiotemporal graph convolutional network model, along with real-time acquired ambient light background information (such as light from wide-angle lenses). With the assistance of an illuminometer (measuring a background light of only 200 lux), the system identifies user-dominated behaviors within each sub-region (e.g., identifying "intensive reading" in region A and "group discussion" in region B) and maps corresponding dynamic light requirement parameters (e.g., "intensive reading" requires supplemental lighting of 300 lux and a color temperature of 4000K on top of the background light). Simultaneously, a suspended multispectral imager scans the plant canopy to acquire spectral images. Image processing algorithms analyze the actual photosynthetically active radiation and spectral composition received by each leaf layer (e.g., calculating insufficient red light proportion in the middle leaves). This data is then compared with the physiological requirements of the plant (e.g., lettuce) at its current growth stage in the knowledge base to generate a supplemental artificial light formula (e.g., supplementing with 100 μmol / m² light between 4 and 6 pm). 2 Based on this, the system projects the two types of dynamic demand sets onto a unified spatiotemporal grid and uses a spatial overlay and time series comparison algorithm for coupled analysis. For example, it can accurately calculate that in area A at 5 pm, there is a significant deviation in the spectral energy distribution between the 4000K cool white light required by people and the red-blue mixed light required by plants, and the demand periods of the two completely overlap. This allows the system to quantify the level of conflict and the room for compromise, and finally generate a structured and evaluable set of behavior-lighting coupled conflict information.

[0043] The method provided in this embodiment accurately captures the conflict points and complementary potential of the two types of light needs, avoiding the multiple harms of blind lighting. It not only provides comprehensive and reliable conflict data support for subsequent growth-energy consumption assessment, but also reduces ineffective lighting investment from the source, ensuring normal plant growth and user activity needs, realizing the efficient transformation of perceived data into decision-making basis, and laying a solid foundation for the implementation of the overall dynamic energy-saving matching method.

[0044] In some embodiments, based on multi-user behavior information and combined with ambient light background information, the interaction information between different user actions and plant layout and channel structure is analyzed to obtain multi-user behavior intention recognition information; based on multi-user behavior intention recognition information, the spectral preference of the lighting required to support different user behavior intentions relative to the background light is analyzed to obtain a multi-user activity-spectral illuminance demand mapping table; based on the multi-user activity-spectral illuminance demand mapping table, the dominant activity type that needs to be prioritized in each lighting sub-area at the current moment and the corresponding dynamic light demand of the activity are analyzed to obtain a multi-user behavior dynamic light demand information set.

[0045] Multi-user behavioral intent recognition information can be a set of information inferred from user behavior data and ambient light conditions, indicating the type, purpose, and lighting demand tendency of the user's current dominant activity. A multi-user activity-spectral illuminance demand mapping table can be a structured data table that establishes a one-to-one correspondence between different user behavioral intents and their corresponding spectral preferences and illuminance requirements. Dynamic lighting demand corresponding to an activity can be the dynamic lighting requirements relative to the ambient background light needed to support the user's dominant activity. The dominant activity type can be the type of activity that most users participate in or that a few users perform within a certain lighting sub-area but that has the highest priority for lighting demand during the same time period.

[0046] Specifically, when multiple users are in a space containing plants, if a dynamic light demand information set based on user behavior is not constructed, it will be impossible to accurately capture the differences in user behavior intentions and corresponding light requirements. This may lead to problems such as insufficient illuminance during precise operations and spectrum mismatch during leisure time. This not only affects the user experience but also causes energy waste due to unreasonable light parameters. Furthermore, it will exacerbate the conflict with the light requirements of plants, resulting in a lack of accurate data support for subsequent lighting adjustments and making it difficult to achieve a balance between supply and demand. To address the aforementioned issues: First, spatiotemporal alignment and fusion analysis are performed on the "multi-user behavior information" acquired by the depth camera and the "ambient light background information" acquired by the light sensor. For example, a trajectory clustering algorithm is used to identify a user who has formed a long-term, small-range circular lingering trajectory in front of a specific plant. Combined with the fine movement patterns of their hand key points, and the detection that the current ambient light in that area is relatively dim (e.g., below 50 lux), a pre-trained graph neural network model can be used. This model has learned the association between behavior-spatial context and intent, thereby inferring that the user's behavioral intent is "close-up observation of plant details." Subsequently, a pre-set "multi-user activity-spectral illuminance requirement mapping table" is queried. This table is constructed from visual ergonomics experimental data, with entries such as: for the "close-up observation of plant details" activity, to achieve optimal color reproduction and... For detailed identification, in addition to ambient light, diffuse light with a color rendering index Ra higher than 90, a color temperature of 4000K, and a key illumination illuminance of 300lx needs to be added. This requirement is then compared with the measured ambient light background of the area (e.g., current illuminance of 50lx and color temperature of 5500K) to obtain the spectral and illuminance parameters that need to be dynamically supplemented by artificial light (e.g., approximately 250lx of 4000K light needs to be added). Finally, within the virtually divided "lighting sub-regions," multiple simultaneously identified intentions are optimized and ranked using multi-objective methods, considering factors such as user priority and activity urgency, to determine the "dominant activity type" of the area at the current moment (e.g., "viewing" takes precedence over "reading"). The calculated corresponding dynamic light requirement parameters, along with the results from other areas, are integrated into a structured "multi-user behavior dynamic light requirement information set."

[0047] The method provided in this embodiment accurately identifies user behavioral intentions, establishes a correlation between activities and lighting needs, clarifies the lighting requirements of dominant activities in each area, and provides accurate data support for subsequent lighting control. This not only avoids the problem of lighting being out of sync with user needs and improves the lighting comfort of multiple users, but also reduces unnecessary energy consumption. At the same time, it lays the foundation for coordinating the light needs of users and plants and formulating scientific lighting strategies, promoting the coordinated development of lighting systems towards intelligence, humanization, and energy conservation.

[0048] In some embodiments, based on multi-user behavior information and combined with ambient light background information, the system analyzes the coordination information between user behavior trajectory and spatial movement pattern with plant layout and passage structure under ambient light background to obtain user behavior spatial interaction characteristics; based on user behavior spatial interaction characteristics, the system analyzes the influence of different regional lighting conditions on user dwell tendency and action precision to obtain user behavior trends; based on user behavior trends, the system infers the specific type and purpose of each user's current dominant activity to obtain multi-user behavior intent recognition information.

[0049] Plant layout can include the distribution location, planting density, and variety distribution of plants within a space. Passage structure can include the width, direction, and branching distribution of pathways within a space. User behavior spatial interaction characteristics can be the features presented by the coordination and adaptation between user behavior trajectories, movement patterns, and plant layout and passage structure. User dwell tendency can be the user's willingness to stay and preferred dwell time in different areas of the space due to differences in lighting conditions. Action fineness can be the level of detail and ease with which users complete actions such as reading, writing, and manipulating objects under specific lighting conditions. User behavior trends can be the user behavior preferences and activity development directions identified by analyzing the impact of lighting conditions on user dwell tendency and action fineness. The specific type and purpose of the dominant activity can be the categories of activities currently being undertaken by the user and their core needs.

[0050] Specifically, in spaces where multiple users coexist with plants, the lack of multi-user behavioral intent recognition means that relying solely on surface behavior cannot differentiate between different needs such as office work and leisure. This leads to a disconnect between lighting supply and actual demand, resulting in either mismatched spectrum and illuminance or energy waste. It also exacerbates the conflict with plant light requirements, interfering with plant growth and making subsequent lighting adjustments inaccurate, severely impacting the overall adaptability and energy efficiency of the lighting system. To address these issues: a network of deployed depth cameras captures key points of user skeletons and movement trajectories, combined with wearable sensors (such as inertial measurement units) to obtain fine-grained posture information, thus constructing multi-user behavioral data. Simultaneously, a distributed spectrometer and ambient light sensor array acquire real-time ambient light background information for each area, including the spectral power distribution and illuminance of natural light and existing supplemental lighting (e.g., the current photosynthetically active radiation in a certain area is only 150 μmol / m²). 2 / s), then, using spatiotemporal alignment and feature fusion technology, the user trajectory and posture sequence are mapped and analyzed with the plant layout (e.g., a hydroponic lettuce area with a plant spacing of 0.3 meters) and channel structure (e.g., a main working channel with a width of 1.2 meters) in a high-precision digital twin model. This extracts spatial interaction features of user behavior, such as calculating the average distance between the user trajectory and the plant ridge line, and the distribution of dwell points in the channel. Based on these features, a statistical regression model is used to analyze the relationship between light conditions (e.g., illuminance below a certain canopy drops to 50 lux) and micro-indicators of user behavior (e.g., the dwell time in this area increases from an average of 30 seconds to 90 seconds, and the action speed decreases). The correlation between the spatial interaction features and the quantified behavioral trends is quantified by analyzing the correlation between the 40% and the behavior trends. Finally, the fused spatial interaction features and quantified behavioral trends are input into a pre-trained lightweight graph neural network classifier or a rule-based decision tree. This model is associated with a predefined "scene-behavior-intent" knowledge graph, thereby inferring the specific type and purpose of each user's current dominant activity (e.g., user A moving slowly in a low-light lettuce area with frequent pauses and bending over to observe is identified as having a "growth status inspection" intent; user B moving quickly in a straight line in a bright passage is identified as having a "material handling" intent). The structured output is multi-user behavioral intent recognition information.

[0051] The method provided in this embodiment accurately uncovers the real needs behind user behavior, providing precise direction for subsequent lighting matching. This makes lighting adjustments more targeted, avoiding energy waste caused by blindly supplying light and reducing ineffective conflicts with plant light requirements. It ensures user comfort and plant growth needs, builds a bridge between perceived information and light requirements, and improves the efficiency and effectiveness of dynamic energy-saving matching.

[0052] In some embodiments, based on plant canopy light distribution information and combined with ambient light background information, the actual light intensity and spectral composition of each region of the plant canopy are analyzed to obtain a set of measured canopy light environment information. Based on the set of measured canopy light environment information, according to pre-stored plant species and growth stage information, the spectral intensity and light duration required by different plants at each growth stage are analyzed to obtain plant stratified and time-based physiological light demand benchmarks. Based on the plant stratified and time-based physiological light demand benchmarks, the spectral ratio, spectral intensity, and spectral action period of supplemental artificial light required to meet the needs of plant growth are analyzed to obtain a set of dynamic light demand information for plants.

[0053] The measured canopy light environment information set can be a structured dataset reflecting the actual light conditions in different areas of the plant canopy. Pre-stored plant species and growth stage information can be basic plant-related data pre-stored in the lighting control database. Plant stratified and time-specific physiological light requirement benchmarks can be light requirement standards developed for different plant species, different growth stages, and different canopy layers and time periods. Spectral ratios can be the proportions of different spectral components in the additional artificial light needed to meet plant physiological light requirements. Spectral intensity can be the light intensity of additional artificial light in the corresponding spectral band to compensate for insufficient natural or existing light. Spectral activity periods can be the time intervals between the on and off of supplementary artificial light.

[0054] Specifically, in spatial scenarios where multiple users and plants coexist (such as botanical gardens and plant factories), the lack of a behavior-light coupling conflict information set can lead to confusion in the light demands of users and plants. Spatially, discrepancies in spectral intensity demands between the two can result in excessive or insufficient light. Temporally, conflicting light demand sequences can lead to wasted lighting or inhibited plant growth. Furthermore, the feasibility and priority of supply and demand cannot be determined, ultimately resulting in a surge in energy consumption, damaged plant growth, and a deterioration in user experience. To address these issues, a distributed multispectral sensor array deployed in the upper, middle, and lower layers of the plant canopy can be used to collect real-time data on light intensity and spectral composition at each layer (e.g., measuring the red light intensity at the top of the canopy as 600 μmol / m²). 2 / s, while the bottom is only 150μmol / m 2 Simultaneously combining the dynamic background of natural light (such as current solar intensity and its spectral changes) captured by wide-angle ambient light sensors deployed around the periphery of the area, and employing background light stripping and data fusion techniques (methods), the effective illumination contributed by artificial light sources to each layer of the canopy after deducting the influence of ambient light is accurately calculated. This constructs a "canopy light environment measured information set" reflecting the true light-receiving profile. Next, a pre-constructed plant physiological optics knowledge base (technology) is invoked and queried. This knowledge base, based on botanical theory and experimental data, stores the ideal spectral ratios (e.g., blue light:red light 1:3) and light intensity thresholds (e.g., 300 μmol / m² for photosynthetic saturation) required by different plants (e.g., lettuce and tomato) at different growth stages (e.g., vegetative growth period and flowering period) at different heights within the canopy. 2Based on the light demand (e.g., 40% blue light deficit at the bottom), the measured information set is intelligently compared with the "plant stratified and time-specific physiological light demand benchmark" in the knowledge base through demand matching and difference analysis algorithms (methods). This allows for the precise diagnosis of light deficit (e.g., 40% blue light deficit at the bottom), spectral deviation (e.g., insufficient far-red light ratio), and time-specific gaps in plants in each region. Finally, based on these diagnostic results, the specific parameters for artificial intervention required to precisely fill these gaps are dynamically derived using spectral synthesis theory and photon flux density calculation models (methods). This generates parameters that include targeted "spectral ratios" (e.g., increasing the proportion of red light in supplemental light to a specific value to promote flowering) and compensatory "spectral intensity" (e.g., precisely setting the supplemental light intensity to compensate for bottom deficit by 150 μmol / m²). 2 The "plant dynamic light demand information set" (the level required for the / s difference) and the "spectral action period" (e.g., scheduling supplemental lighting in the afternoon when natural light is weaker) serve as precise instructions to drive subsequent intelligent supplemental lighting.

[0055] The method provided in this embodiment can accurately clarify the spatial dimension spectral intensity deviation and complementarity possibility, and the temporal dimension light demand matching contradiction, clarify the feasibility of supply and demand and the degree of compromise, which not only provides a precise basis for subsequent supply and demand balancing and energy consumption reduction, avoiding energy waste caused by blind lighting, but also reduces the interference of improper lighting on plant growth, while ensuring that the user's behavioral light demand is reasonably met, and achieving synergistic optimization of the three.

[0056] In some embodiments, based on the dominant activity type and the corresponding dynamic light demand, combined with spectral ratio and spectral intensity, the magnitude of the deviation and the possibility of complementarity between the spectral intensity required to meet the user's dominant activity and the spectral intensity required to meet the plant's physiological light demand are analyzed in each lighting sub-region, thus obtaining spatial light demand coupling information; based on the activity-corresponding dynamic light demand, combined with the spectral action period, the matching degree and contradiction points of the user's activity light demand and the plant's light demand in terms of duration, activation sequence, and change frequency are analyzed within a similar time period, thus obtaining temporal light demand coupling information; based on spatial light demand coupling information, combined with temporal light demand coupling information, the feasibility level of simultaneously meeting the user's and plant's light demands, the required degree of compromise, and the potential impact of prioritizing one side are analyzed, thus obtaining a behavior-lighting coupling conflict information set.

[0057] A lighting sub-region can be an independent light control area defined based on spatial layout, plant distribution, and user activity range. The deviation magnitude can be the range of differences between the spectral intensity required for user-dominated activities and the spectral intensity required for plant physiological light needs within the lighting sub-region. Complementarity potential can be the degree to which overlapping bands in the spectra required by users and plants can be shared. Spatial light demand coupling information can be a set of information characterizing the conflict and complementarity states of user and plant light demands in terms of intensity and spectrum within the lighting sub-region. Duration can be the continuous duration of light required by user activities or plant light demands. Activation sequence can be the required time sequence for light activation. Change frequency can be the time interval for adjusting light parameters. Matching degree can be the degree of convergence between user and plant light demands in terms of duration, activation sequence, and change frequency within the same time period. Temporal light demand coupling information can be a set of information on the matching status and conflict points of user and plant light demands in the time dimension. Feasibility level can be a rating of the difficulty in simultaneously satisfying the light demands of both users and plants. Required compromise level can be the degree to which the other demand needs to be adjusted to simultaneously satisfy the needs of both parties or to prioritize one party. The potential impact of prioritizing one party could be the adverse consequences on the other party when only the light needs of the user or the plant are met.

[0058] Specifically, in spatial processes where multiple users and plants coexist (such as botanical gardens and plant factories), skipping this step will prevent the identification of spatial spectral intensity discrepancies between the two (e.g., the difference between users requiring 500 lx and plants requiring 300 lx). The temporal inconsistencies in activation timing will lead to blind lighting adjustments, either hindering plant growth or causing a decline in user experience. Furthermore, parameter conflicts will increase ineffective energy consumption, disrupting the balance between the two needs. To address these issues: First, through spatial demand coupling analysis, for each independently controllable lighting sub-area, the dynamic light demand mapped from user-dominated activities (e.g., fine-tuning operations require 600 lux, high color rendering white light) and the dynamic light demand information of plants are combined to define the spectral ratio and intensity for plants in that area (e.g., tomato seedlings require a photosynthetic photon flux density of 300 μmol / m²). 2The system juxtaposes and compares light sources with a red-to-blue light ratio of 3:1, specifically analyzing the magnitude of the deviation in illuminance / radiance intensity (e.g., the user's required illuminance is approximately twice that of the plant's baseline). It also assesses the possibility of complementary spectral compositions (e.g., whether the broad spectrum of the white light used in the operation can effectively cover the specific red and blue wavelengths required by the plant). Secondly, through temporal co-analysis, it aligns the user's activity light demand patterns (e.g., a group visit requiring high illuminance from 3 PM to 4 PM daily, lasting 1 hour) with the plant's spectral activity periods (e.g., continuous supplemental lighting for 3 hours from 2 PM to 5 PM on the same day), specifically evaluating the differences in duration (1 hour vs. 3 hours), activation timing (the visit falls entirely within the supplemental lighting period), and variation. The matching degree and contradictions in the frequency of lighting (visitors require instantaneous peak lighting, while plants need stable light intensity) are analyzed. Finally, based on the above spatiotemporal analysis results, a multi-objective conflict analysis method is used to assess the overall feasibility level of simultaneously meeting the dual requirements (e.g., synergy can be achieved in 70% of the area and within the time period). This is quantified into the degree of compromise required to achieve this feasibility (e.g., temporarily reducing the illuminance in the core visitor area by 20%, or adjusting the supplemental lighting for plants to be done in two phases before and after the visit). The potential impact of prioritizing user activities on plant growth is simulated (e.g., it may lead to a 15% shortage of plant light accumulation on that day). This generates a refined behavior-light coupling conflict information set that integrates conflict location, coordination costs, and risk prediction.

[0059] The method provided in this embodiment accurately analyzes the conflicts in light demand across spatial and temporal dimensions, clarifies the magnitude of deviations and points of conflict, and provides a targeted basis for subsequent lighting adjustments. This avoids interference with plant growth and poor user experience caused by blind adjustments, while also enabling targeted optimization of lighting power and duration, reducing ineffective energy consumption. This makes lighting adjustments more scientific and targeted, efficiently balancing the needs of multiple user behaviors, plant growth requirements, and energy-saving goals, achieving a dynamic balance among the three.

[0060] In some embodiments, based on spatial light demand coupling information, the artificial light illumination power required to bridge the deviation magnitude is analyzed to obtain spatial conflict incremental energy consumption information; based on temporal light demand coupling information, the extended artificial light illumination duration required to coordinate the matching degree and contradiction points is analyzed to obtain temporal conflict incremental energy consumption information; based on feasibility level and combined with the required compromise degree, the deviation of prioritizing user-led activities from the plant's stratified and time-based physiological light demand benchmark is analyzed to obtain plant growth disturbance risk assessment information caused by lighting adjustments; based on spatial conflict incremental energy consumption information, temporal conflict incremental energy consumption information, and combined with plant growth disturbance risk assessment information, the energy consumption cost and plant growth risk of different priorities are analyzed to obtain a growth-energy consumption coupling assessment information set.

[0061] Spatial conflict incremental energy consumption information can be related to the additional artificial lighting power required to bridge the discrepancy between the spectral intensity required for user-led activities and the spectral intensity required for plant physiological light needs. Temporal conflict incremental energy consumption information can be related to the additional artificial lighting duration required to coordinate the matching degree and contradictions between user activity light needs and plant light needs in terms of duration, activation sequence, and frequency of change. Plant growth disturbance risk assessment information can be related to the degree of deviation from the benchmark for stratified and time-based physiological light needs of plants when prioritizing user-led activities. Feasibility level can be used to characterize the feasibility of simultaneously meeting the light needs of multiple user behaviors and the physiological light needs of plants. Required compromise level can be used to characterize the degree of mutual concessions required to meet the light needs of both parties. Matching degree can be the degree of compatibility between user activity light needs and plant light needs in terms of duration, activation sequence, and frequency of change. Contradiction point can be the point of conflict between the two in the temporal dimension.

[0062] Specifically, in complex spaces where multiple users and plants coexist, skipping this step and blindly adjusting lighting can lead to uncontrolled energy consumption (such as excessively increasing power or extending lighting time), deviating from the plant's physiological light requirements (such as insufficient light duration), causing stunted plant growth, and failing to meet the light needs of users' activities (such as excessively dim lighting), resulting in visual fatigue and decreased activity efficiency, thus violating the core objectives of dynamic energy saving and meeting dual needs. To address these issues, multi-level quantitative modeling and coupling analysis techniques are used to transform the conflict into a measurable cost indicator. For spatial conflicts, based on the spatial light demand coupling information from the previous step, a light intensity deviation power conversion model is used to calculate the incremental LED driving power required to bridge the illuminance deviation in a specific area (e.g., calculating the need to add a 18W rated power lighting module), thereby generating incremental energy consumption information for spatial conflicts. To address time-series conflicts, based on the coupled information of time-series light demand, and using time-slot coordination and energy consumption mapping algorithms, the system analyzes the additional artificial lighting duration required for aligned or staggered lighting (e.g., to accommodate a 2-hour evening visit, the plant supplemental lighting dark period needs to be interrupted and extended), along with its corresponding energy consumption (e.g., calculating that this would increase the total lighting duration by 2 hours and energy consumption by approximately 0.5 kWh). This generates incremental energy consumption information for time-series conflicts. Simultaneously, it calls upon pre-stored plant growth models and light response databases, combining feasibility levels and the required degree of compromise, and employs a growth deviation risk assessment algorithm to simulate and predict under specific compromise schemes (e.g., ...). By reducing the daily cumulative light duration from the preset 14 hours to 10 hours for three consecutive days, the deviation of the plant's stratified and time-specific physiological light requirement benchmark was assessed. This was then used to evaluate the potential loss probability and degree in key physiological indicators such as dry matter accumulation and flowering synchronization, and to output quantitative plant growth disturbance risk assessment information. Finally, through a multi-objective cost fusion engine, the above-mentioned incremental energy consumption data and growth risk index were normalized and correlated to generate a clear growth-energy consumption coupling assessment information set. This set reveals the comprehensive cost of each potential solution in a structured data form (such as the "energy consumption-risk" cost spectrum under different plans).

[0063] The method provided in this embodiment accurately captures the energy consumption patterns and risk characteristics of dynamic lighting adjustments, providing a reliable basis for subsequent lighting strategy formulation, avoiding various problems caused by blind adjustments, and allowing lighting to flexibly select priority targets or optimize compromise solutions. While taking into account user experience and plant maintenance, it minimizes total energy consumption and improves the adaptability and operating efficiency of the lighting system.

[0064] In some embodiments, based on feasibility levels, several feasible lighting scheme options are analyzed to simultaneously meet the light requirements of plants and the light requirements of multiple user behaviors, resulting in a set of multiple feasible options. Based on the set of multiple feasible options, combined with spatial conflict incremental energy consumption information and temporal conflict incremental energy consumption information, the estimated additional energy consumption of each lighting scheme option to achieve the corresponding lighting coverage is analyzed, resulting in a set of multiple incremental energy consumption indicators. Based on the set of multiple feasible options, combined with plant growth disturbance risk assessment information, the potential negative impact of light environment adjustments on plant growth rate and physiological indicators of each lighting scheme option is analyzed, resulting in a set of multiple plant growth risk indicators. Based on the set of multiple incremental energy consumption indicators, combined with the set of multiple plant growth risk indicators, the comprehensive ranking of the advantages and disadvantages of several lighting scheme options in terms of energy consumption and growth risk within the required compromise range is analyzed, resulting in a set of growth-energy consumption coupling assessment information.

[0065] The multi-option feasibility set can be a collection of several lighting schemes that are technically feasible and cost-controllable, selected based on feasibility levels. The multi-option incremental energy consumption index set can be the set of estimated additional energy consumption for each lighting scheme option to achieve lighting coverage. The multi-option plant growth risk index set can be the set of the potential negative impacts of light environment adjustments on plant growth for each lighting scheme option. Energy cost can be the estimated additional energy consumption cost during the implementation of each lighting scheme option. Plant growth risk can be the potential negative impacts of light environment adjustments on plant growth rate and physiological indicators for each lighting scheme option.

[0066] Specifically, in the process of dynamically matching the lighting needs of multiple users with the light needs of plants, if only scattered energy consumption and risk information is available, it is impossible to clarify the trade-off relationship between different lighting schemes. Blindly choosing a scheme may lead to either excessive risks to plant growth (such as a significant decrease in growth rate) or uncontrolled energy consumption that violates energy-saving goals. Furthermore, the scheme may fail to be implemented due to a lack of feasibility level and degree of compromise, which seriously affects the scientific nature and effectiveness of subsequent lighting strategy formulation. To address the aforementioned issues: First, a multi-objective optimization algorithm is invoked, using the feasibility level obtained from the behavior-lighting coupling conflict information set as the search boundary. This automatically generates a series of lighting scheme options differentiated in spectral ratio, intensity, and timing. For example, Scheme A prioritizes user reading activities by slightly increasing the intensity of 400-500nm blue light; Scheme B prioritizes plant photosynthesis by maintaining a high proportion of 600-700nm red light; and Scheme C seeks a compromise by using a mixed spectrum but appropriately reducing total illuminance. These options collectively constitute a multi-scheme feasibility option set. Subsequently, two evaluations are performed in parallel: First, an energy consumption prediction model is used to correlate the specific parameters of each scheme with spatial conflict incremental energy consumption information from the prior analysis (e.g., the power increase required to compensate for the spectral intensity deviation between user and plant needs in a certain area, e.g., 0.2kW) and temporal conflict incremental energy consumption information (e.g., the additional artificial lighting required to coordinate user evening activities with plant twilight supplemental lighting). The system performs fusion calculations on lighting duration (e.g., 30 minutes) to accurately predict the increased energy consumption of each scheme compared to the baseline scenario, forming a multi-scheme incremental energy consumption index set. On the other hand, it uses a plant growth response model to simulate the physiological responses that each lighting scheme may cause, such as changes in plant photosynthetic rate and abnormal internode elongation, based on the risk types defined by the plant growth disturbance risk assessment information (e.g., spectral shift, photoperiod interruption). It quantifies the degree of potential negative impact (e.g., the estimated inhibition rate of the target plant's daily growth, e.g., 5%), forming a multi-scheme plant growth risk index set. Finally, it adopts a decision evaluation method based on compromise planning. Within the required compromise degree limited by the behavior-lighting coupling conflict information set (e.g., allowing a maximum decrease of 10% in plant light demand satisfaction), it assigns dynamic weights to the two objectives of energy consumption and growth risk, comprehensively scores and ranks all feasible schemes, and outputs a growth-energy consumption coupling assessment information set with a clear structure and containing a sequence of advantages and disadvantages of each scheme.

[0067] The method provided in this embodiment clearly presents the relationship between energy consumption and risk for each solution, effectively avoiding the drawbacks of single-objective decision-making. Under the dual constraints of demand, the optimal solution is accurately selected, ensuring that the risk of plant growth is controllable while minimizing energy consumption. This provides solid decision support for subsequent lighting adjustment strategies and improves the feasibility and adaptability of the overall solution.

[0068] In some embodiments, based on a multi-scheme plant growth risk index set, the potential negative impact of lighting scheme options on plant growth is analyzed, and all lighting scheme options with potential negative impact levels below a preset risk threshold are selected to obtain a candidate set of safe lighting schemes. Based on the candidate set of safe lighting schemes, combined with a multi-scheme incremental energy consumption index set, the additional energy consumption corresponding to each candidate lighting scheme option is analyzed, and the candidate scheme with the smallest additional energy consumption is selected to obtain the lowest energy consumption safe scheme. Based on the lowest energy consumption safe scheme, combined with the required compromise degree and feasibility level, the feasibility of dynamically responding to user-led activities by fine-tuning local spectral intensity or temporarily extending the lighting duration of some areas while maintaining the core lighting parameters of the lowest energy consumption safe scheme to minimize energy consumption is analyzed to obtain a lighting adjustment strategy.

[0069] The potential negative impact level can be a quantified result of the adverse effects of the lighting scheme option on plant growth rate and physiological indicators after implementation. The preset risk threshold can be a pre-set critical value used to determine whether a lighting scheme is safe for plant growth. The safe lighting scheme candidate set can be a set obtained by screening lighting scheme options with a potential negative impact level below the preset risk threshold. Additional energy consumption can be the additional electrical energy required by the lighting scheme option compared to the baseline lighting state to meet the light needs of plants and users. The lowest energy consumption safe scheme can be the candidate scheme with the lowest additional energy consumption selected from the safe lighting scheme candidate set. Core lighting parameters can be the key lighting parameters in the lowest energy consumption safe scheme that determine energy consumption and plant growth safety. Fine-tuning local spectral intensity can be a small adjustment of the spectral intensity of a certain lighting sub-region while keeping the core lighting parameters unchanged. Temporarily extending the lighting duration of a certain area can be a short-term extension of the lighting duration of a specific lighting sub-region without significantly increasing the total energy consumption. Dynamically responding to user-driven activities can be a flexible adjustment of lighting parameters based on changes in the user's real-time behavioral light needs.

[0070] Specifically, in lighting processes where multiple users and plants coexist, the absence of this step directly leads to lighting schemes either interfering with plant physiological indicators due to a lack of screening for plant growth risks, or wasting energy due to a lack of optimized energy consumption. Furthermore, the schemes cannot dynamically adapt to changes in user activity, resulting in untimely fulfillment of light demands, poor user experience, and disrupting the balance between the two needs. To address these issues: First, a preset threshold screening technique is employed. For example, a preset risk threshold of 5 is set for a comprehensive indicator reflecting the degree of plant light stress (such as the "light stress index"). This threshold is used to automatically compare all lighting scheme options in the set of plant growth risk indicators, quickly eliminating all schemes with potentially excessive negative impacts (e.g., an index higher than 5), thus generating a candidate set of safe lighting schemes. Next, an optimization comparison and ranking technique is used to read the additional energy consumption corresponding to each scheme from this safe candidate set (e.g., scheme A requires an additional 3.2 kWh / cycle, scheme B requires 2.8 kWh / cycle). Through direct numerical comparison or ascending sorting algorithms, the specific scheme with the lowest energy consumption (e.g., scheme B) is accurately identified and established as the lowest energy consumption safe scheme. The core lighting parameters of this scheme constitute the static optimal framework of the strategy. Finally, dynamic elasticity assessment and rule embedding technology are introduced. Combining the feasibility level of the scheme with the predetermined degree of compromise (such as allowing the daily cumulative light intensity of plants to fluctuate by ±10% in the short term in response to user activities), the possibility of making local and temporary adjustments to specific sub-regions while keeping the core parameters unchanged is analyzed. For example, when a user is detected performing fine operations on a workbench, a fine-tuning rule is evaluated and generated: temporarily increase the white light intensity in the spectrum of that area by 15% for a duration not exceeding 20 minutes. Finally, such evaluated and feasible dynamic fine-tuning rules are integrated with the basic framework of the lowest energy consumption safety scheme to output a complete lighting adjustment strategy that combines static optimization and dynamic adaptability.

[0071] The method provided in this embodiment, through risk screening and energy consumption optimization, avoids excessive interference of lighting on plant growth, ensures stable plant growth, minimizes total energy consumption, and reduces energy costs. At the same time, the local parameter fine-tuning design can dynamically respond to user needs, improve the adaptability of the light environment and the smoothness of user behavior, and ensure the efficient and accurate operation of the lighting system under dual constraints.

[0072] In some embodiments, direct control instructions for programmable lighting devices are generated based on core lighting parameters, fine-tuning of local spectral intensity, and temporary extension of lighting duration in certain areas to drive the lighting lamps to perform corresponding adjustments, thereby obtaining strategy execution information. Based on the lighting adjustment strategy and combined with a multimodal collaborative sensing information set, the impact of light changes on plant canopy light distribution information and multi-user behavior information is analyzed to obtain effect verification information. The effect verification information includes at least the trend of plant photosynthetic rate change and the trend of user behavior fluency change. Based on the strategy execution information and effect verification information, the current time stamp, dominant activity type, and plant stratified time-division physiological light demand benchmark are associated to generate and output a dynamic energy-saving matching log of multi-user lighting demand.

[0073] Programmable lighting equipment can be lighting devices that support receiving digital control commands and can flexibly adjust parameters such as spectrum, intensity, and duration. Direct control commands can be digital commands generated based on lighting parameter adjustment requirements and can be directly recognized and executed by the programmable lighting equipment. Strategy execution information can be data reflecting the parameter adjustment execution status after the lighting equipment receives and executes control commands. Effect verification information can be verification data reflecting the impact of light environment adjustment on plant growth and user behavior. The trend of plant photosynthetic rate change can be the trend of plant photosynthetic efficiency changing over time after light adjustment. The trend of user behavior smoothness change can be the change in the continuity and smoothness of user actions during activities after light adjustment. Timestamps can be specific time information marking the execution of lighting strategies and the generation of logs.

[0074] Specifically, in spaces where multiple users and plants coexist, without adaptive adjustment of lighting parameters and log output steps, the lighting strategy is merely a theoretical solution that cannot be implemented to meet dual lighting needs. Furthermore, deviations in plant growth and user experience after light adjustment are difficult to detect in a timely manner. Without log traceability, subsequent strategy optimization will be impossible, which in the long run can easily lead to problems such as energy waste, damaged plant growth, and interference with user behavior, thus disrupting the balance between energy conservation and meeting needs. To address the aforementioned issues: The process begins by translating the strategy into executable device instructions. After receiving the lighting adjustment strategy, the control module employs parameter compilation and instruction encapsulation techniques. For example, core lighting parameters in the strategy (such as setting the basic spectral ratio to R:G:B=1:2:1 and the basic illuminance to 300 lx) are compiled into a specific format of control protocol data packets. Simultaneously, for dynamic adjustments in the strategy, such as fine-tuning local spectral intensity (e.g., instantly increasing the blue light component intensity by 20% in a user's precision work area) or temporarily extending the lighting duration of certain areas (e.g., extending the supplemental lighting duration in area A by 15 minutes to cope with sudden harvesting tasks), corresponding dimming curves and timed trigger instructions are generated. These instructions are sent to the corresponding programmable lighting devices (such as LED lamp arrays with PWM dimming and independent spectral channel control) via wired or wireless communication interfaces (e.g., DALI, Zigbee), driving them to precisely execute adjustments to the spectrum, intensity, and switching timing, and providing real-time status feedback to form strategy execution information. Following this, an effect verification loop is initiated: by calling the relevant departments... The system utilizes an environmental sensor network (such as spectrometers and PAR sensors on the canopy) and behavior sensing units (such as high-definition cameras and infrared sensors) to collect new multimodal collaborative sensing information after the strategy is executed. It employs change detection and correlation analysis methods, for example, comparing canopy light distribution maps before and after the adjustment to calculate the spatial uniformity change of photosynthetically active radiation (PAR), and combining this with a plant photosynthesis-light response model to deduce the overall trend of plant photosynthetic rate. Simultaneously, it analyzes user behavior video streams, using posture recognition and trajectory analysis algorithms to quantify the time, accuracy, and number of pauses for users to complete specific actions (such as detection and handling), assessing the trend of user behavior fluency. Finally, it uses information fusion and structured template technology to integrate the above strategy execution information and effect verification information, along with the current timestamp (e.g., 2023-10-27 14:30:05), the type of dominant activity triggering this adjustment (e.g., "team visit"), and the relevant plant light requirement benchmarks (e.g., "the target for lettuce canopy light intensity during the growing season is a daily light integral (DLI) of 17 mol / m²"). 2 By associating and integrating context information such as " / d", a unified and complete JSON or XML format log file is automatically generated, which is the dynamic energy-saving matching log of multi-user lighting demand. The log file is then stored in the database or pushed to the management interface to complete the current control cycle.

[0075] The method provided in this embodiment precisely drives the implementation strategy of lighting equipment, ensuring that light parameters meet dual requirements and avoiding a disconnect between theory and practice. Real-time effect verification can quickly avoid adjustment deviations, ensuring plant photosynthetic efficiency and smooth user behavior. Log retention provides data support for subsequent strategy iterations, helping to continuously optimize the lighting solution. Ultimately, it achieves synergy in minimizing energy consumption, ensuring plant growth safety, and improving user experience, forming a closed loop of "strategy-execution-optimization".

[0076] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A dynamic energy-saving matching method for multi-user lighting needs based on deep multimodal behavior recognition, characterized in that, include: A multimodal collaborative sensing information set is acquired. Based on the multimodal collaborative sensing information set, the coupling conflict characteristics between multi-user behavioral light demand and plant light demand are analyzed to obtain a behavior-light coupling conflict information set. Based on the behavior-lighting coupling conflict information set, the energy consumption increase and plant growth disturbance risk caused by dynamic lighting are analyzed to obtain the growth-energy consumption coupling assessment information set. Based on the growth-energy consumption coupling evaluation information set, the adjustment information for minimizing total energy consumption under the dual constraints of satisfying the plant light demand and the multi-user behavior light demand is analyzed, and the lighting light adjustment strategy is obtained. Based on the aforementioned lighting adjustment strategy, lighting parameters are adaptively adjusted, and a dynamic energy-saving matching log of multi-user lighting needs is output.

2. The method according to claim 1, characterized in that, Based on the multimodal collaborative sensing information set, the coupling conflict characteristics between multi-user behavioral light demand and plant light demand are analyzed to obtain a behavior-light coupling conflict information set, including: The multimodal collaborative sensing information set includes multi-user behavior information, plant canopy light distribution information, and ambient light background information; Based on the multi-user behavior information and combined with the ambient light background information, the behavior types of users in different spatial areas and their corresponding dynamic light requirements are analyzed to obtain a multi-user behavior dynamic light requirement information set. Based on the plant canopy light distribution information and the ambient light background information, the physiological light requirements and current actual light exposure of the plant are analyzed to obtain a dynamic light requirement information set of the plant. Based on the multi-user behavior dynamic light demand information set and combined with the plant dynamic light demand information set, the differences and overlaps in light demand between multi-users and plants in spatial and temporal dimensions are analyzed to obtain the behavior-light coupling conflict information set.

3. The method according to claim 2, characterized in that, The process of constructing the multi-user behavior dynamic optical demand information set includes: Based on the multi-user behavior information and combined with the ambient light background information, the interaction information between different user actions and plant layout and channel structure is analyzed to obtain multi-user behavior intent recognition information. Based on the multi-user behavior intent recognition information, the spectral preference of the illumination required to support different user behavior intents relative to the background light is analyzed to obtain a multi-user activity-spectral illuminance requirement mapping table. Based on the multi-user activity-spectral illuminance demand mapping table, the dominant activity types and corresponding dynamic light demands that each lighting sub-region needs to satisfy at the current moment are analyzed to obtain the multi-user behavior dynamic light demand information set.

4. The method according to claim 3, characterized in that, The process of constructing the multi-user behavior intent recognition information includes: Based on the multi-user behavior information and combined with the ambient light background information, the user's behavior trajectory and spatial movement pattern under ambient light background are analyzed in conjunction with the plant layout and the channel structure to obtain the user's spatial interaction characteristics. Based on the aforementioned user behavior spatial interaction characteristics, the influence of different regional lighting conditions on user dwell tendency and action precision is analyzed to obtain user behavior trends. Based on the user behavior trends, the specific type and purpose of each user's current dominant activity are inferred, thus obtaining the multi-user behavior intent recognition information.

5. The method according to claim 4, characterized in that, The process of constructing the plant dynamic light demand information set includes: Based on the plant canopy light distribution information and the ambient light background information, the actual light intensity and spectral composition of each region of the plant canopy are analyzed to obtain a set of measured information on the canopy light environment. Based on the measured information set of canopy light environment, and according to the pre-stored information on plant species and growth stages, the spectral intensity and light duration required by different plants at each growth stage are analyzed to obtain the physiological light requirement benchmarks of plant stratification and time. Based on the plant's stratified and time-based physiological light requirements benchmark, the spectral ratio, spectral intensity, and spectral duration of supplemental artificial light required to meet plant growth are analyzed to obtain the plant's dynamic light requirement information set.

6. The method according to claim 5, characterized in that, The method, based on the multi-user behavior dynamic light demand information set and combined with the plant dynamic light demand information set, analyzes the differences and overlaps in light demands between multi-users and plants in spatial and temporal dimensions to obtain the behavior-light coupling conflict information set, including: Based on the dominant activity type and the corresponding dynamic light demand of the activity, combined with the spectral ratio and the spectral intensity, the deviation magnitude and complementarity possibility between the spectral intensity required to meet the user's dominant activity and the spectral intensity required to meet the plant's physiological light demand are analyzed in each of the lighting sub-regions to obtain spatial light demand coupling information. Based on the dynamic light demand corresponding to the activity, and combined with the spectral action period, the matching degree and contradictions between the user's activity light demand and the plant's light demand in terms of duration, activation sequence and change frequency are analyzed within the same time period, so as to obtain the time-series light demand coupling information. Based on the spatial light demand coupling information and the temporal light demand coupling information, the feasibility level, the required degree of compromise, and the potential impact of prioritizing one side's light demand are analyzed to obtain the behavior-light coupling conflict information set.

7. The method according to claim 6, characterized in that, Based on the behavior-lighting coupling conflict information set, the risk of energy consumption increase and plant growth disturbance caused by dynamic lighting is analyzed to obtain a growth-energy consumption coupling assessment information set, including: Based on the spatial light demand coupling information, the artificial light illumination power required to bridge the deviation magnitude is analyzed to obtain the spatial conflict incremental energy consumption information. Based on the temporal light demand coupling information, the artificial light illumination time required to coordinate the matching degree and the contradiction point is analyzed to obtain the temporal conflict incremental energy consumption information. Based on the feasibility level and the required compromise level, the deviation of prioritizing user-led activities from the benchmark of plant stratified and time-based physiological light requirements is analyzed to obtain risk assessment information on plant growth disturbance caused by lighting adjustments. Based on the spatial conflict incremental energy consumption information and the temporal conflict incremental energy consumption information, combined with the plant growth disturbance risk assessment information, the energy consumption cost and plant growth risk of different priority satisfaction are analyzed to obtain the growth-energy consumption coupling assessment information set.

8. The method according to claim 7, characterized in that, The analysis of energy consumption costs and plant growth risks for different priority satisfactions yields the growth-energy consumption coupling assessment information set, including: Based on the feasibility level, several lighting scheme options that can be implemented to simultaneously meet the plant light requirements and the multi-user behavior light requirements are analyzed, resulting in a set of multiple feasible options. Based on the set of multiple feasible options, combined with the spatial conflict incremental energy consumption information and the temporal conflict incremental energy consumption information, the estimated additional energy consumption of each lighting option to achieve the corresponding lighting coverage is analyzed to obtain a set of multiple incremental energy consumption indicators. Based on the set of multiple feasible options and combined with the plant growth disturbance risk assessment information, the potential negative impact of light environment adjustment on plant growth rate and physiological indicators of each lighting option is analyzed to obtain a set of plant growth risk indicators for multiple options. Based on the multi-scheme incremental energy consumption index set and the multi-scheme plant growth risk index set, the comprehensive advantages and disadvantages of several lighting scheme options in terms of energy consumption and growth risk are analyzed within the required compromise range, thus obtaining the growth-energy consumption coupled evaluation information set.

9. The method according to claim 8, characterized in that, Based on the growth-energy consumption coupling assessment information set, the adjustment information for minimizing total energy consumption under the dual constraints of satisfying the plant's light requirements and the multi-user behavior light requirements is analyzed to obtain an illumination light adjustment strategy, including: Based on the multi-scheme plant growth risk index set, the potential negative impact of the lighting scheme options on plant growth is analyzed, and all lighting scheme options with potential negative impact levels lower than a preset risk threshold are selected to obtain a candidate set of safe lighting schemes. Based on the candidate set of safe lighting schemes, and combined with the multi-scheme incremental energy consumption index set, the additional energy consumption corresponding to each candidate lighting scheme option is analyzed, and the candidate scheme with the smallest additional energy consumption is selected to obtain the lowest energy consumption safe scheme. Based on the minimum energy consumption safety scheme, and considering the required compromise level and the feasibility level, the feasibility of dynamically responding to user-led activities by fine-tuning local spectral intensity or temporarily extending the lighting duration of some areas while maintaining the core lighting parameters of the minimum energy consumption safety scheme to minimize energy consumption is analyzed, thus obtaining the lighting adjustment strategy.

10. The method according to claim 9, characterized in that, The process of adaptively adjusting lighting parameters based on the lighting adjustment strategy and outputting a dynamic energy-saving matching log of multi-user lighting needs includes: Based on the core lighting parameters, the fine-tuned local spectral intensity, and the temporary extension of the lighting duration in a certain area, direct control instructions for the programmable lighting device are generated to drive the lighting lamp to perform corresponding adjustments, thereby obtaining strategy execution information; Based on the aforementioned illumination adjustment strategy and combined with the multimodal collaborative sensing information set, the impact of light changes on the plant canopy light distribution information and the multi-user behavior information is analyzed to obtain effect verification information. The effect verification information includes at least the trend of plant photosynthetic rate change and the trend of user behavior fluency change; Based on the strategy execution information and the effect verification information, and by associating the current timestamp, the dominant activity type, and the plant stratified time-based physiological light demand benchmark, a dynamic energy-saving matching log for multi-user lighting demand is generated and output.

Images