LED infrared induction lamp human behavior pattern recognition method and system

Abstract

The application discloses a human behavior mode recognition method and system for LED infrared induction lamps, relates to the field of human behavior mode recognition for LED infrared induction lamps, and is used for more accurately recognizing human behavior modes in an induction area and comprises the following steps: acquiring infrared signals in the induction area; performing feature analysis on the infrared signals to obtain signal features of the infrared signals; the signal features include periodic fluctuation features related to life activities, spatial position change features of heat sources and heat distribution mode change features; generating a confidence index indicating the existence degree of personnel according to the signal features; and adjusting the lighting duration of the induction lamp according to the confidence index.

Description

Technical Field

[0001] This invention relates to the field of human behavior pattern recognition using LED infrared sensor lights, and more particularly to a method and system for human behavior pattern recognition using LED infrared sensor lights. Background Technology

[0002] In smart buildings, LED lighting systems with integrated infrared sensors are commonly installed in office areas to achieve refined energy management and improve the working experience. This system uses passive infrared sensors to capture changes in heat sources within the space. When it detects human infrared radiation, it determines that someone is active and turns on the lighting. When someone leaves, the sensor fails to detect a valid human heat source signal within a set delay time, and the system automatically turns off the lighting, thus saving electricity. This design works well in most cases, effectively avoiding the energy waste of traditional lighting systems that leave lights on even when people are gone.

[0003] However, in actual office environments, the system has revealed some problems. Firstly, besides people, there are other objects in an office that can continuously generate heat, such as high-performance laptops. When employees leave their workstations but don't turn off their running computers, the computers continue to emit heat, forming a stable, locally hot spot with a temperature significantly higher than the environment. Because the system's judgment logic is mainly based on the presence and intensity of heat source signals, it cannot effectively distinguish the nature of this heat source and easily misinterprets the stable infrared signals generated by the computer as "sitting people," thus continuously providing lighting for unoccupied areas and wasting electricity.

[0004] On the other hand, in winter or in environments with low air conditioning temperatures, employees may wear thick clothing and remain stationary for extended periods. Heavy clothing hinders the outward radiation of body heat, resulting in weak infrared signals emitted from their skin, sometimes approaching the level of ambient thermal noise. In this "extremely quiet" and "weak thermal signal" situation, the system may incorrectly identify employees actually working at their workstations as having "left" because the received signal is below the effective "human presence" threshold, and thus execute a lights-off operation. This not only disrupts the employee's concentration and affects work efficiency but also contradicts the "unobtrusive and comfortable" user experience that smart lighting systems aim to provide. Summary of the Invention

[0005] This application discloses a method and system for recognizing human behavior patterns using LED infrared sensor lights, aiming to solve the problem of misjudgment in existing smart building LED infrared sensor light systems when identifying stationary people or non-human heat sources, which leads to energy waste and a decline in user experience.

[0006] In a first aspect, this application discloses a method for recognizing human behavior patterns using an LED infrared sensor, comprising the following steps: Acquire infrared signals within the sensing area; Infrared signals are subjected to feature analysis to obtain their signal characteristics, which include periodic fluctuations related to life activities, spatial location variations of heat sources, and variations in heat distribution patterns. Based on the signal characteristics, a confidence index is generated to indicate the degree of presence of the personnel. Adjust the illumination duration of the sensor lights based on the confidence level index.

[0007] Optionally, based on signal characteristics, a confidence index is generated to indicate the degree of presence of the person instructing the signal, including: Based on the periodic fluctuation characteristics, a periodic fluctuation characteristic score is generated; A spatial dynamic score is generated based on the characteristics of spatial location changes and the characteristics of heat distribution pattern changes. A confidence index is generated based on the periodic fluctuation characteristic score and the spatial dynamic score.

[0008] Optionally, the method also includes: The infrared signal is subjected to low-frequency filtering to obtain a low-frequency signal; Temporal and spatial correlation analysis was performed on the low-frequency signal to obtain the correlation analysis results; Based on the correlation analysis results, determine the nature of the source of the low-frequency signal; Based on the nature of the source, the initial periodic fluctuation feature score is corrected to obtain the periodic fluctuation feature score.

[0009] Optionally, temporal and spatial correlation analysis is performed on the low-frequency signal to obtain the correlation analysis results, including: Time correlation evaluation of low-frequency signals is performed to obtain the cross-correlation coefficients and phase differences of low-frequency signals between different sensing units; and Spatial variability assessment of low-frequency signals is performed to obtain the intensity gradient of low-frequency signals between different sensing units.

[0010] Optionally, the method also includes: A preliminary analysis of the infrared signal was conducted to estimate the center location of the heat source and analyze the heat distribution, yielding preliminary analysis results. The visible light signal within the sensing area is acquired, and changes in visible light intensity are identified to obtain information on visible light signal changes. When the preliminary analysis results show that there is displacement at the center of the heat source or a change in the heat distribution pattern, the judgment result is obtained by determining whether the visible light signal change information shows a synchronous, significantly changing visible light spot movement, and whether its movement trajectory is consistent with the displacement trajectory of the heat source. Based on the judgment results, the initial spatial dynamic score is corrected to obtain the spatial dynamic score.

[0011] Optionally, the method also includes: The infrared signal is spatially divided into multiple sub-region signals; Heat source aggregation processing is performed on the signals of each sub-region to identify and separate the individual heat sources; For each independent heat source, we monitor minute changes in its spatial location and subtle changes in its heat distribution pattern to obtain the spatial location change characteristics and heat distribution pattern change characteristics of each independent heat source. Based on the spatial location variation characteristics and heat distribution pattern variation characteristics of each independent heat source, a dynamic feature score is generated for each individual heat source. The initial spatial dynamic score is obtained based on the dynamic feature scores of each entity.

[0012] Optionally, heat source aggregation processing is performed on the signal of each sub-region to identify and separate individual heat sources, including: Spatial distribution analysis of signals in sub-regions is performed to identify multiple potential heat source centers; Starting from the center of the potential heat source, signal intensity attenuation analysis is performed outward to estimate the radiation range of each potential heat source; For heat sources with overlapping radiation ranges, signal attribution is determined based on the signal intensity attenuation trend in the overlapping area. For heat sources that present fragmented signals, signal completion is performed based on the signal strength and distribution of the unobstructed portion. Based on the signal attribution judgment and signal completion results, each independent heat source is identified and separated.

[0013] Optionally, spatial distribution analysis can be performed on the sub-regional signals to identify multiple potential heat source centers, including: Multi-scale gradient analysis is performed on the sub-region signal to identify the boundaries of regions with high signal intensity change rates; Within the region boundary, multiple signal strength peak points are identified; Remove signal intensity peaks with intensity below a preset threshold and use the remaining signal intensity peaks as multiple potential heat source centers.

[0014] Optionally, multi-scale gradient analysis can be performed on the sub-region signal to identify the boundaries of regions with high signal intensity change rates, including: Gradient calculations are performed on the sub-region signal in multiple directions to obtain gradient data in multiple directions; Directional consistency evaluation is performed on gradient data in multiple directions to obtain directional consistency evaluation results, thereby identifying regions that exhibit gradient changes in multiple directions; Time series analysis was performed on the directional consistency assessment results to identify regions that persist and have stable gradient changes. Regions that are persistent and have stable gradient changes are considered as the boundaries of regions with high signal intensity change rates.

[0015] Secondly, this application also discloses an LED infrared sensor light human behavior pattern recognition system, the system comprising: The signal acquisition module is used to acquire infrared signals within the sensing area; The feature analysis module is used to perform feature analysis on infrared signals to obtain the signal characteristics of the infrared signals; the signal characteristics include periodic fluctuation characteristics related to life activities, spatial location variation characteristics of heat sources, and heat distribution pattern variation characteristics. The confidence generation module is used to generate confidence indices indicating the degree of presence of personnel based on signal characteristics. The lighting duration adjustment module is used to adjust the lighting duration of the sensor lamps based on the confidence index. Beneficial effects

[0016] This application provides a method for recognizing human behavior patterns in LED infrared sensor lights. By acquiring infrared signals within the sensing area and performing feature analysis, signal features are obtained, including periodic fluctuations related to life activities, spatial location changes of heat sources, and changes in heat distribution patterns. Based on these multi-dimensional signal features, this application can generate a confidence index indicating the degree of human presence and adjust the illumination duration of the sensor light accordingly. This method overcomes the limitations of existing technologies that rely solely on single or limited infrared signal features for judgment. It effectively solves the problems of misjudging non-human heat sources (such as running computers) as human presence and continuing to illuminate them, and misjudging the presence of people and turning off the lights when they are stationary or wearing heavy clothing due to weak infrared signals. By comprehensively analyzing features such as periodic fluctuations, spatial location changes, and heat distribution pattern changes, this application can more accurately identify human behavior patterns within the sensing area, significantly improving the recognition accuracy and robustness of the LED infrared sensor light system. This effectively avoids energy waste and enhances user comfort and work efficiency in smart buildings. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart of a method for recognizing human behavior patterns using an LED infrared sensor provided in an embodiment of the present invention; Figure 2This is a schematic diagram of another LED infrared sensor lamp human behavior pattern recognition method provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an LED infrared sensor light human behavior pattern recognition system provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] First, let's introduce the terminology used in this application.

[0021] "Infrared signal" refers to the infrared radiation emitted by objects within the sensing area, which is captured by an infrared sensor and converted into an electrical signal. These signals carry information such as the object's temperature and motion.

[0022] "Signal characteristics" are representative attributes extracted from infrared signals after processing and analysis, used to distinguish different types of heat sources and behavioral patterns. In this application, signal characteristics specifically include "periodic fluctuation characteristics," "spatial location variation characteristics of heat sources," and "heat distribution pattern variation characteristics."

[0023] "Periodic fluctuation characteristics" mainly refer to the slight changes in the intensity or pattern of infrared signals that are related to human life activities such as breathing and heartbeat, and have a certain periodicity.

[0024] "Spatial position change characteristics of heat source" describes the movement or displacement of heat source in the spatial position within the sensing area.

[0025] The “characteristics of heat distribution pattern change” reflects the dynamic changes in the range, shape or intensity of heat distribution around the heat source.

[0026] The "confidence index" is a quantitative indicator used to represent the probability or credibility of the system's judgment that a person exists within the sensing area. The higher the index, the greater the probability that a person is present.

[0027] "Lighting duration of the sensor light" refers to the length of time that the LED infrared sensor light maintains its illumination state after detecting the presence of a person.

[0028] The human behavior pattern recognition method proposed in this application performs multi-dimensional feature analysis on infrared signals and generates confidence indexes based on these features, thereby intelligently adjusting the illumination duration of the sensor lamp to achieve accurate recognition of human behavior patterns.

[0029] The following specific embodiments will provide a detailed introduction and explanation of the LED infrared sensor lamp human behavior pattern recognition method provided in this application.

[0030] Reference Figure 1 This invention provides a method for recognizing human behavior patterns using LED infrared sensors, comprising the following steps: S1, acquire the infrared signal within the sensing area.

[0031] Signal characteristics include periodic fluctuations related to life activities, spatial location variations of the heat source, and changes in heat distribution patterns. For example, to extract periodic fluctuation characteristics, time-domain analysis of infrared signals, such as Fourier transform or wavelet transform, can be performed to identify the presence of periodic components within a specific frequency range. These periodic components may correspond to physiological activities such as respiration or heartbeat. For spatial location variations of the heat source, the center position and trajectory of the heat source can be estimated by analyzing the signal strength and temporal relationships of different sensor units. For example, the centroid algorithm or triangulation method can be used to determine the real-time location of the heat source. Changes in heat distribution patterns can be obtained by analyzing the spatial distribution of signal intensity around the heat source. For example, parameters such as signal intensity gradient, area, or shape changes in the heat source region can be calculated to reflect the dynamic changes in heat distribution patterns.

[0032] S2. Perform feature analysis on the infrared signal to obtain the signal characteristics of the infrared signal.

[0033] Among them, the signal characteristics include periodic fluctuation characteristics related to life activities, spatial location variation characteristics of heat sources, and heat distribution pattern variation characteristics.

[0034] For example, to extract periodic fluctuation characteristics, time-domain analysis of infrared signals, such as Fourier transform or wavelet transform, can be performed to identify the presence of periodic components within a specific frequency range. These periodic components may correspond to physiological activities such as human respiration or heartbeat. For the spatial location variation characteristics of a heat source, the center position and trajectory of the heat source can be estimated by analyzing the signal strength and temporal relationship of different sensor units. For example, the centroid algorithm or triangulation method can be used to determine the real-time location of the heat source. For the variation characteristics of heat distribution patterns, this can be obtained by analyzing the spatial distribution of signal intensity around the heat source. For example, parameters such as the signal intensity gradient, area, or shape changes in the heat source region can be calculated to reflect the dynamic changes in the heat distribution pattern.

[0035] S3. Based on the signal characteristics, generate a confidence index indicating the degree of presence of the person in charge.

[0036] Specifically, a periodic fluctuation feature score can be generated based on the periodic fluctuation characteristics; a spatial dynamic score can be generated based on the spatial location change characteristics and the heat distribution pattern change characteristics; and a confidence index can be generated based on the periodic fluctuation feature score and the spatial dynamic score.

[0037] Specifically, the periodic fluctuation characteristic score refers to a numerical value obtained by quantifying the periodicity, amplitude, and stability of periodic fluctuation characteristics related to life activities in infrared signals, such as weak heat fluctuations caused by physiological activities like breathing and heartbeat, using specific algorithms (such as Fourier transform and wavelet analysis). This score reflects whether there is a source of activity with vital signs within the sensing area.

[0038] The spatial dynamic score can be understood as a numerical value obtained by comprehensively evaluating the dynamic attributes of the heat source within the sensing area, such as its trajectory, speed, direction, and changes in heat distribution range, shape, and intensity, based on the spatial location and heat distribution pattern changes of the heat source. This score aims to capture the motion and morphological changes of the heat source to distinguish between stationary objects and moving human bodies.

[0039] In practical applications, confidence scores are generated by comprehensively considering both periodic fluctuation characteristic scores and spatial dynamic scores. For example, weighted averages, fuzzy logic reasoning, or machine learning models can be used to fuse these two scores, resulting in a more comprehensive and accurate confidence score that indicates the degree of a person's presence.

[0040] S4. Adjust the illumination duration of the sensor lamps according to the confidence index.

[0041] A confidence threshold can be set. When the confidence index is higher than this threshold, the sensor light remains illuminated, and the illumination duration can be dynamically adjusted based on the confidence index value. For example, the higher the confidence index, the longer the illumination duration can be to avoid frequent switching on and off. When the confidence index is lower than this threshold, the system determines that the person may have left or there are no valid personnel present, thus starting a countdown to turn off the lights, and turning off the lights after the countdown ends. As another implementation method, algorithms such as fuzzy control or proportional-integral-derivative (PID) control can also be used to smoothly adjust the illumination duration based on real-time changes in the confidence index, providing a more user-friendly lighting experience.

[0042] The human behavior pattern recognition method proposed in this application acquires infrared signals within a sensing area and performs multi-dimensional feature analysis on these signals, including periodic fluctuation characteristics, spatial location change characteristics of heat sources, and heat distribution pattern change characteristics. These features can more comprehensively and meticulously reflect the presence and activity status of personnel within the sensing area. For example, periodic fluctuation characteristics can effectively identify human life activities in a static state, avoiding misjudging a sitting person as having left; spatial location change characteristics and heat distribution pattern change characteristics can distinguish between personnel movement and stable heating from non-human heat sources. Based on these signal features, the system can generate a confidence index indicating the degree of personnel presence. This index comprehensively considers multiple factors, thereby improving the accuracy of the judgment. Finally, the illumination duration of the sensor light is intelligently adjusted according to the confidence index, avoiding the misjudgment and missed judgment problems caused by traditional systems based on a single judgment criterion.

[0043] Compared to existing technologies, the core innovation of this application lies in the introduction of multi-dimensional infrared signal feature analysis and a lighting duration adjustment mechanism based on confidence index. Traditional systems often rely solely on the simple presence or absence or drastic changes in infrared signals to determine the presence of people. This is prone to misjudgment in complex scenarios such as non-human heat sources or stationary individuals. For example, existing systems cannot distinguish between a continuously heating laptop and a seated person, leading to energy waste; they also cannot effectively identify employees wearing heavy clothing or remaining stationary for extended periods, resulting in frequent light shutdowns and negatively impacting user experience.

[0044] This application effectively identifies human life activities in a static state by extracting periodic fluctuation features. Even when people are stationary for extended periods or wear heavy clothing resulting in weak heat signals, the system can still determine their presence through faint physiological activity signals, thus avoiding the problem of accidentally turning off lights. Simultaneously, by analyzing the spatial location and heat distribution pattern changes of heat sources, the system can more accurately identify the nature and dynamics of heat sources, effectively distinguishing between human movement and stable heating from non-human heat sources, thereby avoiding misinterpreting non-human heat sources as human presence and continuing to illuminate them. This multi-feature fusion analysis method significantly improves the system's accuracy and robustness in recognizing human behavior patterns. By generating confidence indices and adjusting lighting duration accordingly, this application achieves more refined and intelligent lighting control, effectively saving energy, greatly improving the user experience, and avoiding the inconvenience of frequent light switching in traditional systems.

[0045] In some embodiments described above, a periodic fluctuation feature score is generated based on periodic fluctuation characteristics. However, in practical applications, the original infrared signal may be affected by environmental noise, non-human heat sources, or occasional interference, resulting in impure extracted periodic fluctuation features or misjudgments, thus affecting the accuracy of the periodic fluctuation feature score. If these problems are not addressed, the sensor light may misjudge the presence of a person, unnecessarily extending the lighting duration, or failing to provide timely lighting when a person is actually present. Therefore, this application further proposes an optimization scheme that improves the accuracy of the periodic fluctuation feature score by preprocessing the infrared signal and determining its source nature.

[0046] like Figure 2 As shown, this application may also include the following steps: S101. Perform low-frequency filtering on the infrared signal to obtain a low-frequency signal.

[0047] Low-frequency filtering of infrared signals involves using digital or analog filters to remove high-frequency noise and transient interference, retaining only the low-frequency signal components relevant to human activity. The aim is to eliminate high-frequency interference that may be caused by rapid environmental changes or sensor noise, thus allowing subsequent analysis to focus more on potential human activity signals.

[0048] S102. Perform temporal and spatial correlation analysis on the low-frequency signal to obtain the correlation analysis results.

[0049] For example, the interrelationships of low-frequency signals at different points in time and between different sensing units can be evaluated. For instance, the temporal continuity and periodicity of signals, as well as the spatial coordination of signal strength or phase changes between different sensing units, can be analyzed. The aim is to identify signals with specific spatiotemporal patterns, which are typically associated with human movement or vital activities, and distinguish them from patterns of random noise or inanimate heat sources.

[0050] Specifically, time correlation assessment can be performed on low-frequency signals to obtain the cross-correlation coefficients and phase differences between different sensing units; and Spatial variability assessment of low-frequency signals is performed to obtain the intensity gradient of low-frequency signals between different sensing units.

[0051] Among these, evaluating the time correlation of low-frequency signals involves analyzing the time correlation between low-frequency signals and different sensing units to determine the synchronicity or propagation characteristics of the signals. The cross-correlation coefficient quantifies the similarity between two signals at different time delays; a higher cross-correlation coefficient indicates a strong linear relationship or a common time pattern between the signals. Phase difference indicates the time difference in signal arrival between different sensing units, which is crucial for identifying the direction and speed of signal propagation. For example, when a human moves, the infrared signal it generates passes through different sensing units sequentially, causing a certain time lag or lead in the signals received by each unit, i.e., exhibiting a phase difference.

[0052] Furthermore, spatial variability assessment of low-frequency signals refers to determining the spatial characteristics of the signal by analyzing the differences in intensity distribution between different sensing units. Intensity gradient refers to the rate of change of signal intensity in space, reflecting the location, size, and trajectory of a heat source within the sensing area. For example, when a heat source (such as a human body) moves within the sensing area, the infrared signal intensity it generates on different sensing units will exhibit different distribution patterns and trends. By calculating these intensity gradients, the spatial dynamics of the heat source can be effectively depicted.

[0053] This application's solution combines temporal correlation assessment and spatial difference assessment of low-frequency signals, enabling more comprehensive and accurate correlation analysis. Specifically, temporal correlation assessment focuses on capturing the dynamic characteristics of the signal in the time dimension, such as periodicity, synchronicity, and propagation delay, which helps identify periodic fluctuations caused by vital activities (such as breathing and heartbeat). Spatial difference assessment, on the other hand, focuses on capturing the distribution characteristics of the signal in the spatial dimension, such as the location, movement trajectory, and heat distribution pattern of heat sources, which helps distinguish spatial changes caused by human movement or presence from environmental noise. It is precisely because of this combination of temporal and spatial analysis that the determination of the source nature of low-frequency signals becomes more accurate and reliable, effectively distinguishing genuine human activity signals from non-human activity signals (such as environmental interference like wind or swaying curtains).

[0054] S103. Based on the correlation analysis results, determine the source nature of the low-frequency signal.

[0055] Specifically, based on the results of temporal and spatial correlation analysis, it's possible to distinguish whether a signal originates from human activity, environmental interference (such as air conditioning airflow or curtain swaying), or other non-human heat sources (such as pets or equipment heating). For example, if a signal exhibits a continuous and regular spatial trajectory and shows periodicity in time, such as breathing or heartbeat, it might be judged to originate from a human body. The purpose is to provide a reliable basis for subsequent periodic fluctuation characteristic score correction.

[0056] S104. Based on the nature of the source, correct the initial periodic fluctuation feature score to obtain the periodic fluctuation feature score.

[0057] For example, if the signal is determined to originate from non-human interference, the score for that part can be reduced or set to zero; if the signal is determined to originate from a human body, the score weight can be maintained or appropriately increased. The purpose is to ensure that the periodic fluctuation characteristic score accurately reflects the actual presence of people within the sensing area, avoiding misjudgments.

[0058] This application's solution, by introducing low-frequency filtering, effectively suppresses high-frequency noise and transient interference in infrared signals, allowing subsequent analysis to focus more on valid signals related to human activity. Based on this, by performing temporal and spatial correlation analysis on the low-frequency signals, the system can deeply explore the signals' inherent patterns, identifying their continuity and coordination in the spatiotemporal dimensions, thus providing crucial evidence for distinguishing human signals from non-human interference. It is precisely because the system can accurately determine the source nature of the low-frequency signals that it can selectively correct the initial periodic fluctuation feature score, effectively eliminating or weakening spurious fluctuations caused by non-human factors, thereby ensuring the accuracy and reliability of the periodic fluctuation feature score.

[0059] Through the above technical solution, this application can significantly improve the accuracy of periodic fluctuation feature scores and effectively avoid misjudgments caused by environmental noise or non-human heat sources. This makes the sensor lighting system more accurate and robust in identifying the presence of people, thereby enabling more reasonable adjustment of lighting duration, avoiding unnecessary energy waste, and ensuring timely response when people need lighting, thus improving user experience and the system's intelligence level.

[0060] Traditional methods for recognizing human behavior patterns using LED infrared sensors rely solely on the spatial location and heat distribution patterns of infrared signals to generate spatial dynamic scores. However, this single-modal approach has limitations. For example, when non-human heat sources (such as pets, ambient temperature fluctuations, or moving machinery) move within the sensing area, their infrared signals may exhibit similar spatial dynamic characteristics to those of humans, leading to inaccurate spatial dynamic scores and consequently affecting the confidence level of human presence assessment. Failure to address these issues may result in inaccurate adjustment of the sensor's illumination duration, causing energy waste or a poor user experience. Therefore, this application proposes a method that incorporates visible light signals to correct spatial dynamic scores, thereby improving the accuracy of human behavior pattern recognition.

[0061] The above methods also include: S201. Perform preliminary analysis on the infrared signal to estimate the center location of the heat source and analyze the heat distribution to obtain preliminary analysis results.

[0062] Specifically, preliminary analysis of infrared signals refers to using raw infrared data acquired through an infrared sensor array and employing image processing or signal processing algorithms, such as centroid analysis, cluster analysis, or thermal mapping, to estimate the center location of the heat source within the sensing area and analyze the range and intensity distribution pattern of its heat spread, thereby obtaining preliminary analysis results. These preliminary analysis results aim to provide basic spatial dynamic information about the heat source.

[0063] S202. Acquire the visible light signal within the sensing area and identify changes in visible light intensity to obtain information on visible light signal changes.

[0064] Acquiring visible light signals within the sensing area can be understood as using a visible light sensor (such as a photoresistor, photodiode, or CMOS / CCD image sensor) integrated into the LED infrared sensing system to collect ambient light information in real time. Identifying changes in visible light intensity refers to processing the acquired visible light signals, such as through background subtraction, threshold segmentation, or motion detection algorithms, to detect and track the presence of visible light spots with significant intensity changes within the sensing area and record their movement trajectories, thereby obtaining information on visible light signal changes.

[0065] S203. When the preliminary analysis results show that there is displacement at the center of the heat source or a change in the heat distribution pattern, determine whether the visible light signal change information shows a synchronous, significantly changing visible light spot movement, and whether its movement trajectory is consistent with the displacement trajectory of the heat source, and obtain the judgment result.

[0066] In practical applications, when preliminary analysis results indicate displacement of the heat source's center or a change in its heat distribution pattern, the system further determines whether the visible light signal changes show synchronous, significantly varying intensity movements of visible light spots, and whether their trajectories match the heat source's displacement trajectory. This determination aims to verify, through multimodal information fusion, whether the infrared signal change is indeed caused by an entity with visible light characteristics (such as the human body). For example, if the heat source moves, but there are no synchronous, moving visible light spots in the visible light signal, it may indicate that the infrared signal change is not caused by the human body.

[0067] S204. Based on the judgment result, correct the initial spatial dynamic score to obtain the spatial dynamic score.

[0068] If the judgment result is "yes", that is, the movement of the heat source is synchronized with the movement of the visible light spot and the trajectory is consistent, then it is considered that the change in the infrared signal is caused by the human body, and the initial spatial dynamic score can be maintained or appropriately increased; if the judgment result is "no", then it is considered that the change in the infrared signal may not be caused by the human body, and the initial spatial dynamic score should be reduced or cleared to avoid misjudgment.

[0069] This application's solution effectively solves the misjudgment problem that may occur when relying solely on infrared signals to generate spatial dynamic scores by introducing visible light signals as an auxiliary judgment criterion. When the infrared sensor detects a change in the spatial position of a heat source or a change in the heat distribution pattern, the system no longer directly attributes it entirely to human activity. Instead, it further utilizes visible light signals acquired by the visible light sensor for cross-validation. Specifically, by comparing the displacement trajectory of the heat source with the movement trajectory of the visible light spot and judging their synchronicity and intensity significance, it can effectively distinguish between changes in infrared signals caused by human movement and changes in infrared signals caused by non-human heat sources (such as pets or environmental interference). This multimodal fusion judgment mechanism makes the generation of spatial dynamic scores more accurate, thereby improving the accuracy of the confidence index of the degree of human presence.

[0070] Through the above technical solution, this application can significantly improve the accuracy and robustness of LED infrared sensor lights in recognizing human behavior patterns in complex environments. By introducing visible light signals as an auxiliary verification method, it effectively avoids the misjudgments that traditional infrared sensor lights may produce when faced with non-human heat source movement or environmental interference, thereby reducing unnecessary lighting triggering and energy waste. In addition, this solution makes the generation of confidence indicators more reliable, and can more accurately reflect the actual presence of people in the sensing area, thus achieving more intelligent and energy-efficient adjustment of sensor light illumination duration.

[0071] In some preferred embodiments, a specific example is given below. Assume an LED infrared sensor system is configured to detect human activity in an office environment. When the infrared sensor array detects a moving heat source within the sensing area, and its spatial position change characteristics and heat distribution pattern change characteristics show a pattern similar to human movement, the system first generates an initial spatial dynamic score. At this point, the system does not immediately use this score for confidence index calculation; instead, it activates a visible light sensor for synchronous monitoring. If the visible light sensor simultaneously detects a visible light spot with a significantly changing intensity that matches the heat source's movement trajectory (e.g., the movement of a human shadow), the system confirms that the heat source is a human and maintains or increases the initial spatial dynamic score. Conversely, if the heat source moves, but the visible light sensor does not detect a synchronous visible light spot movement, or the detected visible light spot's movement trajectory is inconsistent with the heat source (e.g., a pet is moving, its infrared signal is detected, but its visible light characteristics are not obvious or do not match the expected human characteristics), the system determines that the change in infrared signal is not caused by a human, thereby reducing or resetting the initial spatial dynamic score to avoid misjudgment. In this way, even in scenarios with pets or environmental disturbances, the sensor lights can accurately determine whether someone is present, thus achieving precise lighting control.

[0072] In some embodiments described above, the initial spatial dynamic score is corrected by preliminarily analyzing infrared signals to estimate the center location of heat sources and analyze heat distribution, combined with visible light signal change information. However, in practical applications, multiple independent heat sources may exist within the sensing area, such as multiple people or people and pets present simultaneously, or a single heat source may exhibit discontinuous or fragmented signals due to obstruction. In such cases, simply estimating the overall center location and heat distribution of heat sources may fail to accurately capture the fine dynamic behavior of each independent heat source, thus affecting the accuracy of the initial spatial dynamic score and potentially leading to misjudgments of the presence of people. To address this, this application further proposes a more refined processing scheme. By dividing the infrared signal into spatial regions and aggregating heat sources, individual independent heat sources are identified and separated. Each independent heat source is then monitored and scored separately, resulting in a more accurate initial spatial dynamic score.

[0073] S301. Divide the infrared signal into spatial regions to obtain multiple sub-region signals.

[0074] Specifically, spatial region division of infrared signals refers to dividing the infrared signal data of the entire sensing area into multiple smaller, independent, or partially overlapping sub-regions based on a preset spatial grid or an adaptive algorithm based on signal intensity distribution. Each sub-region's signal represents the infrared radiation information within that region. For example, the sensing area can be divided into an M×N grid, with each grid cell being a sub-region.

[0075] S302. Perform heat source aggregation processing on the signal of each sub-region to identify and separate each independent heat source.

[0076] The process involves heat source aggregation for each sub-region signal. The aim is to accurately identify and separate each individual heat source emitted by a specific object (such as a human body or pet) from the complex infrared signal. This can include methods such as cluster analysis, connected component analysis, or model matching to distinguish the boundaries and centers of different heat sources.

[0077] S303. For each independent heat source, monitor the minute changes in its spatial location and the subtle changes in its heat distribution pattern to obtain the spatial location change characteristics and heat distribution pattern change characteristics of each independent heat source.

[0078] In practical applications, monitoring minute changes in the spatial location and heat distribution pattern of each independent heat source involves continuously tracking its movement along spatial coordinates and analyzing the changes in its heat distribution (e.g., temperature gradient, hot spot size, and shape) over time. This allows us to obtain the spatial location change characteristics and heat distribution pattern change characteristics of each independent heat source. For example, spatial location change characteristics can be obtained by calculating the change in the center of mass of the heat source, and heat distribution pattern change characteristics can be obtained by analyzing the changes in the second moment or entropy value of the heat source.

[0079] S304. Based on the spatial location change characteristics and heat distribution pattern change characteristics of each independent heat source, generate their respective dynamic feature scores.

[0080] This means that each identified individual heat source will receive an independent score based on its own dynamic behavior, reflecting the heat source's activity level or its likelihood of being present as a human. For example, heat sources that move quickly and exhibit significant changes in heat distribution patterns may receive higher dynamic characteristic scores.

[0081] S305. Based on the dynamic feature scores of each entity, the initial spatial dynamic score is obtained.

[0082] Specifically, this can be achieved by weighting, summing, or taking the maximum value of the dynamic characteristic scores of all independent heat sources, thereby comprehensively reflecting the dynamic activity of all independent heat sources within the entire sensing area and forming a more comprehensive and accurate initial spatial dynamic score.

[0083] The proposed solution meticulously divides the infrared signal within the sensing area into multiple sub-region signals and performs heat source aggregation processing on each sub-region signal, thereby identifying and separating individual heat sources within the sensing area. It is precisely because of the ability to distinguish and independently process multiple heat sources that the system can perform refined monitoring of the spatial location changes and heat distribution pattern changes of each individual heat source, generating separate dynamic feature scores. This mechanism of independent multi-heat source analysis effectively avoids the information loss and misjudgment caused by the confusion of multiple heat sources into a single overall heat source in traditional methods. Especially in scenarios with multiple people or complex heat source signals, it can more accurately capture the true dynamic behavior of each heat source. Therefore, by comprehensively considering the dynamic feature scores of all independent heat sources, the resulting initial spatial dynamic score can more comprehensively and accurately reflect the actual presence and activity patterns of people within the sensing area, thus providing a more reliable input for the subsequent generation of confidence indicators.

[0084] Through the above technical solution, this application effectively solves the problem that traditional methods struggle to accurately capture the dynamic behavior of all heat sources when multiple independent heat sources or complex heat source signals exist within the sensing area. By spatially dividing the infrared signal and aggregating the heat sources, precise identification and separation of each independent heat source are achieved. Based on this, each independent heat source is monitored individually, and a dynamic feature score is generated, significantly improving the accuracy and robustness of the initial spatial dynamic score. This refined processing method enables the system to more accurately determine the presence of people in complex environments, effectively avoiding misjudgments or missed judgments caused by heat source confusion or signal fragmentation, thereby improving the overall performance and user experience of the LED infrared sensor light human behavior pattern recognition method.

[0085] In some preferred embodiments, suppose an office area is equipped with LED infrared sensors, and two employees, A and B, are simultaneously moving around in different locations within that area. Traditional infrared sensors may only be able to detect a vague overall heat source movement, making it difficult to distinguish between a large-scale movement by a single person and small-scale movement by two people. According to the solution of this application, firstly, the infrared signal is divided into multiple sub-region signals. Then, these sub-region signals are processed by heat source aggregation, and the system successfully identifies and separates two independent heat sources, corresponding to employees A and B respectively. Subsequently, the system monitors employee A's heat source (e.g., its center position moves from (x1, y1) to (x1', y1'), and the heat distribution pattern changes from circular to elliptical) and employee B's heat source (e.g., its center position moves from (x2, y2) to (x2', y2'), and the heat distribution pattern remains relatively stable). Based on these independent monitoring results, dynamic characteristic scores are generated for employee A (e.g., higher, due to significant movement and changes in heat distribution) and employee B (e.g., lower, due to less activity). Finally, by combining the scores of these two independent dynamic features, a more accurate initial spatial dynamic score is obtained. This score can reflect the presence of two active persons in the area, thus making the adjustment of the illumination duration of the sensor lights more reasonable and avoiding premature extinguishing or unnecessary long-term illumination that may be caused by only identifying a vague heat source.

[0086] Specifically, the above-mentioned heat source aggregation processing of each sub-region signal is used to identify and separate each independent heat source, including: S401. Perform spatial distribution analysis on the sub-region signals to identify multiple potential heat source centers.

[0087] Specifically, multi-scale gradient analysis can be performed on the sub-region signal to identify the regional boundaries with high signal intensity change rates; within the regional boundaries, multiple signal intensity peak points are identified; signal intensity peak points with intensity below a preset threshold are removed, and the remaining signal intensity peak points are used as multiple potential heat source centers.

[0088] Multi-scale gradient analysis of sub-region signals aims to effectively identify the boundary between heat sources and the background by calculating the rate of change of signal intensity at different scales. For example, various gradient operators such as the Sobel operator, Prewitt operator, or Laplacian operator can be used, and calculations can be performed with convolution kernels of different sizes to capture boundary information of heat sources of different sizes. Multi-scale analysis improves the robustness of identifying boundaries of heat sources of different sizes and shapes.

[0089] Furthermore, within the identified region boundary, multiple signal intensity peaks can be determined by scanning the signal intensity distribution of the sub-regions. These peaks typically correspond to the core region of the heat source, i.e., the location where heat is most concentrated. For example, local maximum detection algorithms can be used to identify these peaks.

[0090] Furthermore, to eliminate environmental noise or weak, insignificant heat source interference, the identified signal strength peaks need to be screened. Specifically, signal strength peaks with intensities below a preset threshold are removed. This preset threshold can be set according to the actual application scenario and environmental noise level to ensure that only sufficiently significant heat source centers are retained. Ultimately, the screened signal strength peaks are identified as multiple potential heat source centers.

[0091] The proposed solution first employs multi-scale gradient analysis to accurately delineate the contours and boundaries of the heat source, avoiding the ambiguity or omissions that may occur with single-scale analysis. Based on this, signal intensity peaks are located within these clearly defined boundaries, effectively pinpointing the core location of the heat source. Finally, by setting an intensity threshold, weak signals lacking practical significance are filtered out, ensuring that the identified potential heat source centers are real and significant. This step-by-step and refined processing approach enables the system to accurately extract multiple independent heat source centers from complex infrared signals, laying a solid foundation for subsequent heat source aggregation processing.

[0092] S402. Starting from the center of the potential heat source, perform signal intensity attenuation analysis outward to estimate the radiation range of each potential heat source.

[0093] Specifically, the intensity of infrared signals typically decreases with increasing distance from the center of the heat source. By analyzing this attenuation pattern, a model can be built to predict the radiation boundary of the heat source, thereby determining the radiation range of each potential heat source. This analysis helps to distinguish the influence areas of different heat sources, even those that are spatially close.

[0094] S403. For heat sources with overlapping radiation ranges, the signal attribution is determined based on the signal intensity attenuation trend in the overlapping area.

[0095] In practical applications, for heat sources with overlapping radiation ranges, determining signal attribution based on the signal intensity attenuation trend in the overlapping region is crucial to addressing the signal aliasing problem when multiple heat sources are close to each other. When the radiation ranges of two or more heat sources overlap, the signal intensity in the overlapping region is the result of the combined effect of all heat sources. By analyzing the attenuation trend of the signal intensity within this region relative to the centers of each potential heat source, it is possible to determine which heat source the overlapping signal primarily originates from, or to proportionally distribute it among the heat sources, thereby achieving precise separation of heat source signals.

[0096] S404. For heat sources that present fragmented signals, complete the signal based on the signal strength and distribution of the unobstructed portion.

[0097] Its purpose is to address situations where heat source signals are incomplete due to obstruction, sensor blind spots, or other interference. Fragmented signals may not fully reflect all the information about a heat source. By analyzing the existing, unobstructed signal portions and combining them with the typical radiation patterns and intensity distribution characteristics of the heat source, the missing signal portions can be reasonably inferred and supplemented, thereby restoring the complete information of the heat source and improving the accuracy of identification.

[0098] S405. Based on the signal attribution judgment and signal completion results, identify and separate each independent heat source.

[0099] This step integrates the results of the aforementioned analysis and processing to form a clear and complete identification of all independent heat sources within the sensing area. Through precise signal attribution and effective signal completion, it can be ensured that each identified heat source has a clear spatial location and complete signal characteristics, providing reliable basic data for subsequent human behavior pattern recognition.

[0100] This application's solution effectively addresses the challenge of accurately identifying and separating independent heat sources in complex scenarios where multiple heat sources interfere with each other or signals are incomplete. This is achieved through refined spatial distribution analysis, signal intensity attenuation analysis, signal attribution determination, and signal completion. Specifically, spatial distribution analysis provides initial location of heat sources, while signal intensity attenuation analysis further defines their influence range. For overlapping heat sources, attenuation trend analysis prevents misidentification of multiple heat sources as one or incorrect assignment of a single heat source's signal to multiple sources. For fragmented heat sources, the signal completion mechanism ensures the recovery of complete heat source characteristics even with missing information, preventing missed identification or misjudgment. The synergistic effect of these steps enables the system to accurately extract complete information from complex infrared signals, laying a solid foundation for subsequent feature analysis and behavioral pattern recognition.

[0101] Through the above technical solution, this application can significantly improve the accuracy and robustness of LED infrared sensor lights in recognizing human behavior patterns in complex environments. Specifically, through refined heat source aggregation processing, multiple human heat sources that are close to each other or partially obscured within the sensing area can be effectively distinguished and identified, avoiding misjudgments or missed judgments caused by signal aliasing or incompleteness. This enables the sensor lights to more accurately determine the presence, number, and activity patterns of people, thereby achieving smarter and more energy-efficient lighting control, improving user experience, and reducing energy consumption.

[0102] The above-mentioned multi-scale gradient analysis of sub-region signals to identify the boundaries of regions with high signal intensity change rates includes: S501. Perform gradient calculations in multiple directions on the sub-region signal to obtain gradient data in multiple directions.

[0103] Specifically, multi-directional gradient calculation of a sub-region signal involves applying different gradient operators (such as the Sobel operator, Prewitt operator, or Roberts operator) to perform differential operations on the infrared signal intensity in multiple directions, including horizontal, vertical, and diagonal directions, thereby quantifying the rate of change of signal intensity in different directions. This yields a set of gradient data reflecting the trend of signal intensity variation in various directions. This gradient data can preliminarily reveal potential edges or boundaries within the sub-region signal.

[0104] S502. Perform directional consistency assessment on gradient data in multiple directions to obtain directional consistency assessment results, so as to identify regions that show gradient changes in multiple directions.

[0105] This step aims to determine whether the gradients calculated in different directions exhibit a consistent trend. For example, this can be done by comparing the angles between gradient vectors in different directions or the correlation of their components. When gradient data from multiple directions exhibit a high degree of spatial consistency, it indicates that the signal intensity changes in that region are stable and significant, rather than random noise or transient fluctuations. This assessment yields a directional consistency evaluation result, thereby identifying regions that exhibit gradient changes in multiple directions; these regions typically correspond to the actual boundaries of the heat source.

[0106] S503. Perform time series analysis on the directional consistency assessment results to identify regions that persist and have stable gradient changes.

[0107] Specifically, evaluation results obtained within consecutive time frames can be tracked and compared to determine whether gradient changes in a certain region are persistent and have stable intensity and direction. For example, a time window can be set; if a region is identified as having consistent gradient changes across multiple consecutive time frames, it is considered persistent. Through time series analysis, unstable signals caused by environmental interference or brief movements can be effectively filtered out, thereby identifying regions that are persistent and have stable gradient changes.

[0108] S504. Regions that are persistent and have stable gradient changes are designated as regions with high signal intensity change rates as boundaries.

[0109] These boundaries represent clear demarcation lines between heat sources and the background, or between different heat sources, providing a reliable spatial basis for the subsequent accurate identification of potential heat source centers.

[0110] This application's scheme, through multi-scale gradient analysis of sub-region signals combined with directional consistency assessment and time series analysis, can effectively and robustly identify the boundaries of regions with high signal intensity change rates. First, gradient calculations in multiple directions comprehensively capture the spatial variation information of the signal. Second, the directional consistency assessment mechanism filters out regions exhibiting stable gradient changes in multiple directions, thereby enhancing the ability to identify true boundaries and suppressing the influence of random noise. Finally, time series analysis further ensures the persistence and stability of the identified boundaries, avoiding misjudging instantaneous fluctuations as heat source boundaries. It is precisely due to the synergistic effect of these steps that the identification of heat source boundaries becomes more accurate and reliable.

[0111] The above technical solution overcomes the limitations of traditional methods in identifying heat source boundaries in complex environments, such as susceptibility to noise interference, unclear boundaries, and sensitivity to instantaneous changes. This solution ensures higher accuracy and reliability of the identified region boundaries through the comprehensiveness of multi-directional gradient calculation, the robustness of directional consistency assessment, and the stability of time series analysis. This provides a solid foundation for subsequently determining signal intensity peak points within these boundaries and identifying multiple potential heat source centers, significantly improving the accuracy and stability of heat source identification.

[0112] like Figure 3 As shown, this embodiment of the invention also provides an LED infrared sensor light human behavior pattern recognition system. The system includes: The signal acquisition module is used to acquire infrared signals within the sensing area; The feature analysis module is used to perform feature analysis on infrared signals to obtain the signal characteristics of the infrared signals; the signal characteristics include periodic fluctuation characteristics related to life activities, spatial location variation characteristics of heat sources, and heat distribution pattern variation characteristics. The confidence generation module is used to generate confidence indices indicating the degree of presence of personnel based on signal characteristics. The lighting duration adjustment module is used to adjust the lighting duration of the sensor lamps based on the confidence index.

[0113] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by a computer program instructing related hardware. This program can be stored in the computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be an internal storage unit of the task execution device (including a data sending end and / or a data receiving end) of any of the foregoing embodiments, such as the hard disk or memory of the task execution device. The computer-readable storage medium can also be an external storage device of the terminal device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the terminal device. Further, the computer-readable storage medium can include both the internal storage unit of the task execution device and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the task execution device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0114] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0115] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, essentially, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0116] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application.

Claims

1. A method for recognizing human behavior patterns using an LED infrared sensor, characterized in that, include: Acquire infrared signals within the sensing area; The infrared signal is subjected to feature analysis to obtain the signal characteristics of the infrared signal; the signal characteristics include periodic fluctuation characteristics related to life activities, spatial location variation characteristics of heat sources, and heat distribution pattern variation characteristics; Based on the signal characteristics, a confidence index is generated to indicate the degree of presence of the person. Adjust the illumination duration of the sensor lamps based on the confidence index.

2. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 1, characterized in that, The step of generating a confidence index indicating the degree of presence of a person based on the signal characteristics includes: Based on the periodic fluctuation characteristics, a periodic fluctuation characteristic score is generated; Based on the spatial location change characteristics and heat distribution pattern change characteristics, a spatial dynamic score is generated; The confidence index is generated based on the periodic fluctuation characteristic score and the spatial dynamic score.

3. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 2, characterized in that, The method further includes: The infrared signal is subjected to low-frequency filtering to obtain a low-frequency signal; Perform temporal and spatial correlation analysis on the low-frequency signal to obtain the correlation analysis results; Based on the correlation analysis results, determine the source nature of the low-frequency signal; Based on the source nature, the initial periodic fluctuation feature score is corrected to obtain the periodic fluctuation feature score.

4. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 3, characterized in that, A temporal and spatial correlation analysis is performed on the low-frequency signal to obtain the correlation analysis results, including: The low-frequency signal is subjected to time correlation evaluation to obtain the cross-correlation coefficient and phase difference between different sensing units; and Spatial difference assessment is performed on the low-frequency signal to obtain the intensity gradient of the low-frequency signal between different sensing units.

5. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 2, characterized in that, The method further includes: The infrared signal is preliminarily analyzed to estimate the center location of the heat source and analyze the heat distribution, and preliminary analysis results are obtained. The visible light signal within the sensing area is acquired, and changes in visible light intensity are identified to obtain information on visible light signal changes. When the preliminary analysis results show that there is displacement at the center of the heat source or a change in the heat distribution pattern, it is determined whether the visible light signal change information shows a synchronous, significantly changing visible light spot movement, and whether its movement trajectory is consistent with the displacement trajectory of the heat source, and the determination result is obtained. Based on the judgment result, the initial spatial dynamic score is corrected to obtain the spatial dynamic score.

6. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 5, characterized in that, The method further includes: The infrared signal is spatially divided to obtain multiple sub-region signals; Heat source aggregation processing is performed on the signals of each sub-region to identify and separate the individual heat sources; For each independent heat source, we monitor minute changes in its spatial location and subtle changes in its heat distribution pattern to obtain the spatial location change characteristics and heat distribution pattern change characteristics of each independent heat source. Based on the spatial location change characteristics of each independent heat source and the heat distribution pattern change characteristics, a dynamic feature score is generated for each of them. The initial spatial dynamic score is obtained based on the dynamic feature scores of each entity.

7. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 6, characterized in that, The process of aggregating heat sources for each sub-region signal to identify and separate individual heat sources includes: Spatial distribution analysis of the signals in the sub-regions is performed to identify multiple potential heat source centers; Starting from the center of the potential heat source, signal intensity attenuation analysis is performed outward to estimate the radiation range of each potential heat source; For heat sources with overlapping radiation ranges, signal attribution is determined based on the signal intensity attenuation trend in the overlapping area. For heat sources that present fragmented signals, signal completion is performed based on the signal strength and distribution of the unobstructed portion. Based on the signal attribution judgment and signal completion results, each independent heat source is identified and separated.

8. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 7, characterized in that, The spatial distribution analysis of the signals in the sub-regions to identify multiple potential heat source centers includes: Multi-scale gradient analysis is performed on the signals in the sub-regions to identify the boundaries of regions with high signal intensity change rates; Within the boundary of the region, multiple signal strength peak points are identified; Remove signal intensity peaks with an intensity lower than a preset threshold, and use the remaining signal intensity peaks as the centers of the plurality of potential heat sources.

9. The method for recognizing human behavior patterns using an LED infrared sensor lamp according to claim 8, characterized in that, The step of performing multi-scale gradient analysis on the sub-region signal to identify the boundaries of regions with high signal intensity change rates includes: Gradient calculations are performed on the sub-region signal in multiple directions to obtain gradient data in multiple directions; The gradient data in the multiple directions are evaluated for directional consistency to obtain directional consistency evaluation results, thereby identifying regions that exhibit gradient changes in multiple directions. Time series analysis was performed on the directional consistency assessment results to identify regions that persist and have stable gradient changes. The region that persists and has a stable gradient change is taken as the boundary of the region with a high rate of change in signal intensity.

10. A human behavior pattern recognition system using LED infrared sensors, characterized in that, The system includes: The signal acquisition module is used to acquire infrared signals within the sensing area; The feature analysis module is used to perform feature analysis on the infrared signal to obtain the signal features of the infrared signal; the signal features include periodic fluctuation features related to life activities, spatial location variation features of heat sources, and heat distribution pattern variation features; The confidence generation module is used to generate a confidence index indicating the degree of presence of a person based on the signal characteristics. The lighting duration adjustment module is used to adjust the lighting duration of the sensor lamp according to the confidence index.

Images