Space-time perception method and system based on light intensity and color temperature dual-mode fusion and medium

Abstract

The application provides a spatiotemporal perception method, system and medium based on light intensity and color temperature bimodal fusion, simulates biological perception environment rhythm by means of bionics, and performs the following steps by observing and learning optical periodic characteristics of the environment: raw data preprocessing; detecting natural day cycle and calibrating internal clock; detecting key astronomical events and constructing phase axis; constructing local standard optical feature model based on autonomous learning mechanism; matching real-time data with the model to calculate the current local time; and estimating longitude and latitude information according to the length of day and the initial reference date. Thus, the local spatiotemporal data is calculated with extremely low cost and high precision. Not only can the dependence on external time and space basic hardware facilities be completely eliminated to realize fully autonomous spatiotemporal perception, but also the local model is constructed based on the autonomous learning mechanism, which is globally applicable. Furthermore, the light intensity and color temperature bimodal features are fused, which can effectively suppress the interference of environmental mutations and has high robustness.

Description

Technical Field

[0001] This invention relates to the field of optical sensing technology, specifically to a spatiotemporal sensing method, system, and medium based on the fusion of light intensity and color temperature dual modes. Background Technology

[0002] The core functionality of Internet of Things (IoT) applications relies on precise time synchronization and spatial positioning of electronic devices. Unfortunately, existing time synchronization technologies fail to fully meet this requirement due to the following shortcomings:

[0003] (1) The built-in real-time clock (RTC) chip is prone to cumulative errors due to the physical characteristics of the crystal oscillator being affected by temperature, and requires frequent calibration;

[0004] (2) Global positioning system (such as GPS, Beidou) modules suffer severe signal attenuation in enclosed environments such as indoors and tunnels, as well as in severe weather, resulting in a decrease in positioning accuracy; at the same time, their hardware costs are high, power consumption is high, and there is a risk of privacy and security leakage.

[0005] (3) Network Time Protocol (NTP) relies entirely on a stable network connection and cannot be used in remote areas, underground facilities or secure private networks.

[0006] (4) Existing solutions based on light sensors can only roughly distinguish between day and night states by judging light intensity thresholds, and cannot provide continuous time information. Furthermore, they are easily interfered with under complex weather conditions such as cloudy or rainy weather, which can lead to judgment failure. In addition, the solution completely lacks the ability to calculate geographical location, making it difficult to meet the requirements for accurate positioning.

[0007] Therefore, the industry urgently needs a brand-new spatiotemporal information sensing solution. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide a spatiotemporal sensing method, system and medium based on the fusion of light intensity and color temperature, which can calculate local spatiotemporal data with extremely low cost and high accuracy by sensing environmental rhythms.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0010] A spatiotemporal sensing method based on dual-modal fusion of light intensity and color temperature includes:

[0011] Step S1: Preprocess the acquired raw light intensity data sequence and raw color temperature data sequence to obtain the light intensity data sequence and color temperature data sequence;

[0012] Step S2: detecting the oscillator count difference between two consecutive sunrise events in the light intensity data sequence according to the light intensity change rate, and calibrating the effective count value of the oscillator according to the oscillator count difference;

[0013] Step S3: calculating the local noon time according to the light intensity data sequence, the color temperature data sequence, and the effective count value, and marking it as a reference phase point of the daily time phase axis;

[0014] Step S4: mapping the light intensity data sequence and the color temperature data sequence of consecutive M days to the standard phase axis with the reference phase point as the coordinate origin, and constructing a local standard optical feature model; wherein M is an integer greater than or equal to 2;

[0015] Step S5: calculating the instantaneous change rate of the original light intensity data sequence and the original color temperature data sequence collected in real time respectively, and constructing a feature vector at the current time; calculating the Euclidean distance between the feature vector and the model feature vector of the local standard optical feature model, and obtaining an optimal estimation phase, and then calculating the current local time;

[0016] Step S6: calculating the latitude and longitude information according to the system initial reference date, the calculated length of the day, and the time difference between the local noon time and the world standard time.

[0017] Optionally, the step S1 specifically comprises:

[0018] S1-1: performing a sliding average filtering process on the original light intensity data sequence and the original color temperature data sequence collected in real time respectively, to obtain a filtered light intensity data sequence and a filtered color temperature data sequence;

[0019] S1-2: calculating the change amount of adjacent sampling points corresponding to the current time of the filtered light intensity data sequence and the filtered color temperature data sequence respectively, if the change amount of adjacent sampling points exceeds a dynamic threshold, the sampling point at the current time is removed, and data repair is performed according to the previous sampling point or linear difference method to obtain a light intensity data sequence and a color temperature data sequence respectively;

[0020] wherein, the dynamic threshold of each is dynamically updated according to the L consecutive sampling points before the current time of the filtered light intensity data sequence and the filtered color temperature data sequence respectively, according to the following formula (1) ;

[0021] Formula (1): ;

[0022] ​Wherein, L is an integer greater than or equal to 2; the value range of α is 1-15, which is a sensitivity preset according to the actual application scene; j is a summation index variable, from j=k-L to j=k-1.

[0023] Optionally, the dynamic updating of each corresponding dynamic threshold , specifically comprising the following sub-steps:

[0024] Sub-step A, setting a history data sliding window with length L, L is an integer greater than or equal to 2;

[0025] Sub-step B, according to the history data sliding window, respectively obtaining L continuous sampling points before the current time from the filtered light intensity data sequence and the filtered color temperature data sequence in real time: ;

[0026] Sub-step C, according to the sampling points obtained in the sub-step B, calculating the light intensity data recent average change rate and the color temperature data recent average change rate respectively according to the following formula (1-1):

[0027] Formula (1-1): ;

[0028] Sub-step D, calculating the dynamic threshold of the light intensity data and the color temperature data respectively according to the following formula (1-2) , and updating each previous dynamic threshold ;

[0029] Formula (1-2): ;

[0030] Wherein, the value range of α is 1-15, which is a sensitivity preset according to the actual application scene.

[0031] Optionally, the step S2 specifically comprises:

[0032] S2-1: calculating the ideal count value of a natural day cycle according to the following formula (2) ;

[0033] Formula (2): ;

[0034] Wherein, is an ideal cycle; is the nominal frequency of the oscillator;

[0035] S2-2: detecting the oscillator count difference between the two consecutive sunrise events in the light intensity data sequence according to the light intensity change rate , and marking it as a natural day cycle;

[0036] ​S2-3: Calculate the calibration factor according to the following formula (3) ;

[0037] Formula (3): ;

[0038] S2-4: Calculate the effective count value of the oscillator according to the following formula (4) , and calibrate the original count value n of the oscillator to the effective count value ;

[0039] Formula (4): .

[0040] Optionally, the step S3 specifically comprises:

[0041] S3-1: Mark the oscillator count difference detected in step S2 as a natural day cycle;

[0042] S3-2: Identify the sunrise count value and / or sunset count value corresponding to the first and the next N natural day cycles according to the light intensity data sequence and the color temperature data sequence; N is an integer greater than or equal to 2;

[0043] S3-3: Analyze to obtain a predicted search count window and a count tolerance value according to the sunrise count value and / or sunset count value corresponding to the first and the next N natural day cycles;

[0044] S3-4: Calculate the light intensity change rate of adjacent light intensity data in the light intensity data set corresponding to the predicted search count window in the light intensity data sequence according to the following formula (5) ;

[0045] Formula (5): ;

[0046] wherein, is a preset time resolution of data collection in step S1; is the light intensity data at time t in the light intensity data sequence;

[0047] S3-5: Perform sliding smoothing filtering processing on the light intensity change rate to obtain a light intensity change rate smoothing curve ;

[0048] S3-6: Obtain all time candidate points satisfying the local maximum or local minimum value condition according to the light intensity change rate smoothing curve ; ;

[0049] S3-7: Calculate each time candidate point a corresponding sunrise event significance index or a sunset event significance index, and the sunrise candidate point with the highest sunrise event significance index and the sunset candidate point with the highest sunset event significance index are selected;

[0050] S3-8: According to the light intensity change rate smoothing curve The sunrise candidate point and the sunset candidate point are subjected to light intensity and color temperature multimodal cross verification according to the color temperature change rate smoothing curve of the same period, and sunrise time and sunset time are determined;

[0051] S3-9: The local noon time is calculated according to the sunrise time and the sunset time, and is marked as a reference phase point of the daily time phase axis.

[0052] Optionally, the step S4 further comprises:

[0053] An arithmetic mean is calculated for all data points mapped to the same discrete phase point, and the calculated arithmetic mean is taken as the data point of the corresponding phase point.

[0054] Optionally, the step S4 further comprises:

[0055] The first-order difference values of the light intensity data and the color temperature data of each phase point in the local standard optical feature model are calculated and stored respectively.

[0056] Optionally, the optimal estimation phase obtained in the step S5, and the current local time calculated further comprise:

[0057] The optimal estimation phase is calculated according to the following formula (6);

[0058] Formula (6): ;

[0059] Wherein, is the Euclidean distance;

[0060] The current local time is calculated according to the following formula (7);

[0061] Formula (7): ;

[0062] Wherein, is the local noon time; is the optimal estimation phase.

[0063] The second technical solution provided by the application is:

[0064] A time-space perception system based on light intensity and color temperature bimodal fusion, comprising:

[0065] An optical sensing module configured to acquire an original light intensity data sequence and an original color temperature data sequence;

[0066] a data processing module configured to perform the steps of the above-mentioned spatiotemporal perception method based on fusion of light intensity and color temperature dual modalities;

[0067] an oscillator configured to provide a reference counting function for the system;

[0068] and an output module configured to output the current local time and latitude and longitude information calculated by the data processing and storage module.

[0069] The third technical solution provided by the application is:

[0070] A computer-readable storage medium having a computer program stored thereon, the computer program being capable of implementing the steps of the above-mentioned spatiotemporal perception method based on fusion of light intensity and color temperature dual modalities when executed by the above-mentioned data processing module.

[0071] The application has the advantages that: the application simulates biological perception of environmental rhythm by means of bionics; by observing and learning the optical periodic characteristics of the environment, the local time and latitude and longitude information can be autonomously calculated. Not only can the dependence on external time and space basic hardware facilities be completely eliminated to realize fully autonomous spatiotemporal perception, but also the local standard optical feature model can be constructed based on the autonomous learning mechanism, which is globally applicable. Furthermore, the model is constructed by using multi-day historical data and fusing light intensity and color temperature dual modalities, which can effectively suppress the interference of environmental mutations and has high robustness. In addition, the required hardware devices are all general low-cost devices, and the application also has the characteristics of extremely low power consumption and extremely low hardware cost. BRIEF DESCRIPTION OF DRAWINGS

[0072] Figure 1 A basic flowchart of a spatiotemporal perception method based on fusion of light intensity and color temperature dual modalities provided for an embodiment of the application;

[0073] Figure 2 A specific implementation flowchart of step S3 provided for a specific embodiment;

[0074] Figure 3 A variation curve diagram of light intensity and light intensity change rate in a natural day cycle provided for the light intensity and light intensity change rate in a natural day cycle described in a specific embodiment;

[0075] Figure 4 A variation curve diagram of color temperature and color temperature change rate in a natural day cycle provided for the color temperature and color temperature change rate in a natural day cycle described in a specific embodiment;

[0076] Figure 5 A diagram of light intensity model multi-day data alignment and averaging in the model autonomous learning construction process provided for a specific embodiment;

[0077] Figure 6 A schematic diagram of multi-day data alignment and averaging of a color temperature model in a model autonomous learning construction process is shown in the specific embodiment;

[0078] Figure 7 A structural schematic diagram of a time-space perception system based on light intensity and color temperature dual-mode fusion provided by an embodiment of the present application. DETAILED DESCRIPTION

[0079] To explain the technical principles, specific schemes, possible application scenarios, and the purposes and effects that can be achieved of the present application in detail, the exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. It should be understood that the embodiments described herein and shown in the drawings are only used to more clearly illustrate the technical solutions of the present application, and therefore only serve as examples and are intended to explain the present application, and should not be construed as limiting the present application. It should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein.

[0080] The present application simulates biological perception of environmental rhythm by means of bionics; by observing and learning the optical periodic characteristics of the environment, the local time and latitude and longitude information can be autonomously calculated; and the local time and space data can be calculated with extremely low cost and high precision.

[0081] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in combination with the drawings and specific embodiments of the specification.

[0082] Figure 1 A basic flowchart of a time-space perception method based on light intensity and color temperature dual-mode fusion provided by an embodiment of the present application.

[0083] As shown in Figure 1 , the present application provides a time-space perception method based on light intensity and color temperature dual-mode fusion, which includes the following steps S1-S6.

[0084] In step S1, the original light intensity data sequence and the original color temperature data sequence obtained by collection are preprocessed to obtain a light intensity data sequence and a color temperature data sequence.

[0085] In this embodiment, by preprocessing the original environmental optical data, noise can be suppressed and outliers can be removed, the data quality and reliability can be improved, and the accuracy of subsequent analysis can be ensured.

[0086] In some specific embodiments, the original light intensity data sequence Iraw(t) and the original color temperature data sequence Craw(t) of the environment are synchronously collected by the light intensity sensor and the color temperature sensor, respectively, at a time resolution (i.e. sampling interval ) of data collection set in the system in advance.

[0087] In some embodiments, the pre-processing means taken for the collected environmental optical raw data, i.e. the step S1, specifically comprises the following sliding filter processing of S1-1 and the outlier rejection and repair processing of S1-2.

[0088] S1-1 sliding filter processing: the sliding average filter processing is respectively performed on the real-time collected raw light intensity data sequence Iraw(t) and raw color temperature data sequence Craw(t), to obtain the filtered light intensity data sequence and filtered color temperature data sequence.

[0089] Specifically, the length-N sliding window is adopted to smooth the raw data sequence; for any time point , the value of the filtered raw data is calculated according to the following formula: ; wherein, is the raw light intensity data in the raw light intensity data sequence Iraw(t) or the raw color temperature data in the raw color temperature data sequence Craw(t); N is an integer greater than or equal to 2.

[0090] S1-2 outlier rejection and repair processing: the adjacent sampling point variation corresponding to the current time point in the filtered light intensity data sequence and the filtered color temperature data sequence is respectively calculated, if the adjacent sampling point variation exceeds the dynamic threshold , the sampling point at the current time point is rejected, and the data is repaired according to the previous sampling point or the linear difference method.

[0091] First, a dynamic threshold is respectively set for “light intensity” and “color temperature”, which can be dynamically updated based on the recent average variation rate.

[0092] As a specific example, the dynamic update of the dynamic threshold corresponding to “light intensity” and “color temperature” is realized through the following sub-steps A to D:

[0093] Sub-step A, a length-L historical data sliding window is set, L is an integer greater than or equal to 2, and the unit is “number”.

[0094] Sub-step B, according to the historical data sliding window, L consecutive sampling points before the current time point are respectively obtained from the filtered light intensity data sequence and the filtered color temperature data sequence in real time: .

[0095] Assuming the preset sampling interval =1 minute, set the window length L=60 sampling points, then use the historical data sliding window to collect optical data every minute, the collected optical data is the past 1 hour of light intensity / color temperature data corresponding to the historical data sliding window, which contains the current time Before the L consecutive pre-processed sampling points: .

[0096] Sub-step C, for the light intensity / color temperature sampling points obtained by the window in sub-step B, the average value of the absolute value of the change amount of all adjacent sampling points in the window (i.e. the amplitude of the instantaneous change rate) is calculated according to formula (1-1) respectively, which is used as the recent average change rate of light intensity / light intensity data in the window ;

[0097] Formula (1-1): .

[0098] Sub-step D, the dynamic threshold of light intensity / color temperature data is calculated according to formula (1-2) respectively , that is, the dynamic threshold is set to α times of the corresponding average change rate; and the dynamic threshold of each preceding dynamic threshold is updated;

[0099] Formula (1-2): ;

[0100] Wherein, the optional value range of α is 1-15, which is the sensitivity preset according to the actual application scene; the preferred value range is 3-10.

[0101] That is, the outlier rejection threshold of light intensity / color temperature will be dynamically updated in real time according to the corresponding filtered light intensity / color temperature data average change rate. The calculation formula of the dynamic threshold can be summarized as formula (1) as follows.

[0102] Formula (1): ;

[0103] Wherein, L is an integer greater than or equal to 2; the value range of α is 1-15, which is the sensitivity preset according to the actual application scene; and j is the summation index variable, from j=k-L to j=k-1.

[0104] Secondly, according to the dynamic threshold corresponding to "light intensity" and "color temperature", the outlier rejection and repair are performed:

[0105] If the current time Light intensity / color temperature change Then determine the current time. sampling points If a value is an outlier, remove it; after removal, use the value from the previous time step. Valid sampling points Alternatively, linear interpolation can be used for data repair.

[0106] Through the above preprocessing, a denoised, high-quality light intensity data sequence will be obtained. With color temperature data sequence This output will provide a high-quality data foundation for subsequent data processing such as event detection, model building, and learning, improving the credibility and validity of the results.

[0107] It should be noted that the threshold used for outlier removal in this embodiment is a dynamic threshold that is dynamically updated based on the recent average rate of change. Compared to existing conventional outlier removal thresholds that use fixed values ​​or local / global statistics (such as mean, variance, etc.), this embodiment has the following significant advantages:

[0108] Environmental optical signals exhibit significant non-stationary characteristics, specifically manifested in the dramatic fluctuations in their rate of change depending on weather conditions (sunny, cloudy, rainy) and time of day (dawn, noon, dusk). Therefore, using existing outlier removal threshold settings for outlier removal will yield highly unreliable results. This embodiment employs a dynamically updated threshold. By removing outliers, the system can adapt to the dynamic characteristics of the ambient optical signal in real time, improve sensitivity during stable periods, avoid misjudgment during periods of drastic change, and has significantly better robustness than the existing fixed threshold method.

[0109] In step S2, the oscillator count difference between two consecutive sunrise events in the light intensity data sequence is detected based on the light intensity change rate, and the effective count value of the oscillator is calibrated based on the oscillator count difference.

[0110] In some specific embodiments, this step specifically includes the following S2-1 to S2-4:

[0111] S2-1: Set the nominal frequency of the internal high-stability oscillator to [value]. The ideal count value for a natural day cycle (i.e., 24 hours) is... Calculated according to the following formula (2);

[0112] Formula (2): ;

[0113] in, is the ideal period, in seconds, that is, the ideal and accurate time of 24 hours, 24 hours = 24 * 3600 seconds = 86400 seconds; is the nominal frequency of the internal oscillator.

[0114] Here, the ideal count value calculated is based on the nominal frequency The theoretical ideal value calculated is the basis for subsequent comparison and verification of the internal oscillator.

[0115] S2-2: Detect the oscillator count difference between the two consecutive sunrise events in the light intensity data sequence according to the light intensity change rate , and mark it as a natural day cycle; here, the core of the "detection" process is to capture the significant feature points of the light intensity change rate (i.e., the first derivative of the light intensity-time curve).

[0116] Here, the oscillator count difference to be detected is the real period count value measured by the system by observing actual astronomical events; due to the hardware error of the oscillator, there will be an error between it and the ideal count value .

[0117] In some embodiments, in combination with Figure 3 , the detection process specifically includes the following steps:

[0118] 1. Data preprocessing and change rate calculation:

[0119] The denoised, high-quality light intensity data sequence obtained in step S1 , the light intensity change rate is calculated in real time:

[0120] where, is the sampling interval; the light intensity change rate reflects the degree of light intensity change.

[0121] 2. Set a dynamic detection threshold and search window:

[0122] To improve the robustness and efficiency of detection, the system will set a predicted search window. For example, after detecting the first sunrise event, the system will predict that the next sunrise event will occur approximately at the position of "the oscillator count of the first sunrise event + ". Then, a reasonable time range is set as a search window around this predicted point, so as to avoid interference.

[0123] 3. Event identification and oscillator count recording:

[0124] Within the search window, the algorithm looks for a positive maximum of the light intensity rate of change . At sunrise, the sun rising causes a rapid increase in the light intensity, and the light intensity rate of change will present a significant positive peak. When the value of the light intensity rate of change exceeds a preset significance threshold, the system determines that a sunrise event occurs, and immediately records the accumulated counter value of the internal oscillator at this moment, denoted as .

[0125] 4. Period calculation:

[0126] After successfully detecting two consecutive sunrise events, the system obtains two oscillator counter values: and . The difference between the two is the actual oscillator counter difference value of a natural day period : ;

[0127] As a specific example:

[0128] Preset conditions: the nominal frequency of the internal oscillator = 32768 Hz; the ideal day period counter value = 24 x 3600 x 32768 = 2,831,155,200.

[0129] First day: the system detects that the light intensity rate of change presents a significant positive peak exceeding the threshold at the oscillator counter n = 1,000,000,000, and determines it as a sunrise event. Record = 1,000,000,000.

[0130] Second day: the system detects another sunrise event meeting the conditions at the oscillator counter n = 3,831,155,200. Record = 3,831,155,200.

[0131] Period calculation: the actual natural day period counter is:

[0132] = 3,831,155,200 1,000,000,000 = 2,831,155,200;

[0133] In this example, the measured is exactly equal to the ideal day period counter value , indicating that the internal oscillator is highly accurate. However, in reality, due to hardware errors such as crystal drift, the measured oscillator counter difference value with the ideal day cycle count value There is certainly a deviation, so the calibration factor calculated by the following step S2-3 is needed to correct the system error.

[0134] S2-3: Calculate the calibration factor according to the following formula (3)

[0135] Formula (3):

[0136] S2-4: Calculate the effective count value of the oscillator according to the following formula (4) , and calibrate the original count value n of the oscillator to the effective count value

[0137] Formula (4):

[0138] It can be understood that the purpose of the present application is to create a self-contained timing system that does not rely on any external reference (such as GPS, NTP, etc.). The only timing basis of this system is the internal oscillator (crystal oscillator). However, all physical oscillators have manufacturing tolerances, temperature drifts and aging effects, which cause their vibration frequency to deviate from the nominal value. That is, the real physical time elapses exactly 24 hours, while the natural day cycle count recorded by the error oscillator is not equal to the ideal day cycle count value ( = 24 × 3600 × ). Therefore, if the natural day cycle count is directly used as a timing reference, the system's time calculation will have a cumulative error, which will eventually be completely distorted after a day's accumulation. Therefore, the present embodiment first "detects" the real error of the internal hardware clock (oscillator); then calibrates it to ensure that the internal oscillator can provide a high-precision timing basis.

[0139] From the above, the present embodiment can achieve the following two points through the above step S2:

[0140] 1. Find the basic rhythm: detect the local continuous sunrise event according to the pre-processed ambient optical signal, determine the basic period (24 hours) of the local earth rotation;

[0141] 2. Calibrate the hardware clock: use the natural day cycle, which is an eternal and unchanging physical reference, to calibrate the frequency drift of the internal oscillator, and provide a high-precision timing basis for all subsequent time calculations.

[0142] In this way, without the need for external basis equipment to provide a reference, that is, under the premise of "no external reference", the precise time synchronization and spatial positioning of electronic equipment can be achieved.​​​​​

[0143] At step S3, the local noon time is calculated according to the light intensity data sequence, the color temperature data sequence, and the effective count value, and is marked as a reference phase point of the daily time phase axis.

[0144] In some embodiments, as shown in FIG. 3, the step S3 specifically includes the following S3-1 to S3-9. Figure 2

[0145] S3-1: The oscillator count difference detected in step S2 is marked as a natural day cycle. S3-2: According to the light intensity data sequence and the color temperature data sequence, the sunrise count value and / or the sunset count value corresponding to the first and the next N natural day cycles are identified; N is an integer greater than or equal to 2.

[0146] S3-2: According to the light intensity data sequence and the color temperature data sequence, the sunrise count value and / or the sunset count value corresponding to the first and the next N natural day cycles are identified; N is an integer greater than or equal to 2. Specifically, first, the first day event detection is performed. The first day refers to the first complete working day cycle after the system is powered on (generally, the device will restore the factory settings when it leaves the factory, clear all existing data of the device, and reach the installation site and then start using), and formally enters the effective event detection. Specifically, it refers to the time window of about 24 hours from the time when the system successfully detects and confirms the first "oscillator count difference" in step S2 to the occurrence of the next sunrise event. The event refers to the sunrise / sunset event. It is obtained by performing full-time analysis on the optical data in the first natural day cycle (the count interval determined in step S2) in the light intensity data sequence and the color temperature data sequence. The analysis is based on: only when the light intensity change rate is extremely significant, and the light intensity change rate and the color temperature change rate show the extreme value characteristics of high synchronization, it is confirmed as a reliable astronomical event. Based on this, at least one set of (sunrise, sunset) events or at least one most reliable event (sunrise or sunset) can be successfully identified on the first day through analysis. Here, the time stamp of the sunrise / sunset event is recorded in the form of oscillator count, marked as the sunrise count value and the sunset count value, as the initial anchor point for subsequent operations.

[0147] Specifically, first, the first day event detection is performed. The first day refers to the first complete working day cycle after the system is powered on (generally, the device will restore the factory settings when it leaves the factory, clear all existing data of the device, and reach the installation site and then start using), and formally enters the effective event detection. Specifically, it refers to the time window of about 24 hours from the time when the system successfully detects and confirms the first "oscillator count difference" in step S2 to the occurrence of the next sunrise event. The event refers to the sunrise / sunset event. It is obtained by performing full-time analysis on the optical data in the first natural day cycle (the count interval determined in step S2) in the light intensity data sequence and the color temperature data sequence. The analysis is based on: only when the light intensity change rate is extremely significant, and the light intensity change rate and the color temperature change rate show the extreme value characteristics of high synchronization, it is confirmed as a reliable astronomical event. Based on this, at least one set of (sunrise, sunset) events or at least one most reliable event (sunrise or sunset) can be successfully identified on the first day through analysis. Here, the time stamp of the sunrise / sunset event is recorded in the form of oscillator count, marked as the sunrise count value and the sunset count value, as the initial anchor point for subsequent operations.

[0148] Secondly, the next day and subsequent event detection is performed. The same analysis method as the first day event detection is used to analyze the sunrise / sunset count values corresponding to the next N natural day cycles after the first day. N is an integer greater than or equal to 2.

[0149] ​​​​​​​​​S3-3: According to the sunrise / sunset count value corresponding to the first and the next N consecutive natural day cycles, a predicted search count window and a count tolerance value are analyzed. The predicted search count window refers to a count interval (which can also be understood as a time period) selected in the optical data sequence in order to more accurately predict the sunrise / sunset event of the nth (n is an integer greater than or equal to 1) day; the count tolerance value refers to a count tolerance based on the stability of the day cycle (which can be optionally set to 30 minutes).

[0150] In this embodiment, it is defined that represents a predicted search count window for predicting the sunrise event of the nth day; represents a predicted search count window for predicting the sunset event of the nth day. Wherein, respectively represent the sunrise count value and the sunset count value corresponding to the sunrise event and the sunset event of the previous day; is the oscillator count difference value, i.e., a natural day cycle; ΔN is the count tolerance value.

[0151] Through the predicted search count window, the count value corresponding to the sunset / sunrise event of the nth day can be predicted. Specifically, the event count of the previous day is used to predict the predicted search count window (i.e., the possible count interval) of the current day. For example, it is believed that the sunrise event of the nth day will occur within the count interval of the predicted search count window . As the system continues to run, the multi-day event timestamps (i.e., multi-day event count values) are accumulated, and the count interval of the sunrise / sunset event of any day can be more accurately predicted, thereby gradually narrowing the search window ΔN and improving the detection efficiency and anti-interference ability.

[0152] In some specific embodiments, this embodiment can predict the dynamic update and convergence of the search window based on an adaptive learning mechanism. The essence is to use the continuously accumulated historical data to perform a closed-loop optimization process of "learning-prediction-verification-correction".

[0153] In short, it is an adaptive learning process that continuously collects historical data of successful events, uses statistical methods to continuously reduce the uncertainty of prediction, thereby gradually optimizing (narrowing) the search window, and may further correct the prediction model parameters, and finally realizes the continuous improvement of detection efficiency and reliability.

[0154] Specifically, the core mechanism lies in gradually narrowing the count tolerance value ΔN. The initial state is high uncertainty and wide window. This is because at the initial stage of system operation (for example, after successfully detecting only 1-2 events), the historical data is scarce, and the regularity of the local event occurrence time is not well understood. At this time, the system will use a larger initial count tolerance value ​(e.g. equivalent to ± 60 minutes), a wider prediction search window is set (e.g. ). With continuous learning and convergence, the window is narrowed down with data accumulation. This is because the system keeps running, and successfully detects events on day 3, 4, 5,... and accumulates a sequence of event timestamps (e.g. ...); the system discovers, through analysis of this sequence, that after the hardware oscillator clock is calibrated, the events occur highly stably on the oscillator count scale, i.e. the deviation of the actual occurrence point of the daily event from the predicted center point (N ) usually fluctuates randomly within a bounded and small range; based on this statistical law, the system can gradually and adaptively reduce the count tolerance value ΔN. For example, from ± 60 minutes to ± 30 minutes, and further to ± 10 minutes.

[0155] The above steps S3-1 to S3-3 can also be understood as adaptive setting of the prediction search count window based on the oscillator count. The purpose is to predict the time period in which the current event is likely to occur using historical event information, thereby greatly reducing the search range. This greatly improves the event detection efficiency while suppressing the misleading of noise interference in non-target time periods.

[0156] The following operations are data processing performed on the light intensity data set corresponding to the prediction search count window in the light intensity data sequence.

[0157] S3-4: Calculate the light intensity change rate of the adjacent two light intensity data in the light intensity data set corresponding to the prediction search count window in the light intensity data sequence according to the following formula (5) ;

[0158] Formula (5): ;

[0159] wherein, is a preset system constant, representing the time resolution (sampling interval) of data acquisition in the aforementioned step S1, which can also be understood as the time interval of adjacent data in the light intensity / color temperature data sequence. For example, if the system is set to collect optical data once every minute, then = 60 seconds. In formula (5), acts as the denominator to convert the difference value of light intensity into the change rate per unit time.

[0160] is the light intensity data at time t in the light intensity data sequence; is the previous light intensity data of . Here, it refers specifically to the light intensity data within the prediction search count window, so The set of light intensity change rates of each adjacent two light intensity data in the light intensity data set corresponding to the prediction search count window is specifically predicted. Here, the light intensity values with higher credibility after being purified are used to calculate the change rate instead of the original data, which can significantly improve the quality of the change rate signal, avoid the sharp fluctuation of the change rate curve caused by data noise, and ensure the reliability of the extreme value points (sunrise / sunset) detected subsequently.

[0161] S3-5: performing a sliding smoothing filtering process on the set of light intensity change rates to obtain a light intensity change rate smoothing curve. Thus, the high-frequency noise in the set of light intensity change rates is suppressed.

[0162] S3-6: finding all the count values corresponding to the light intensity change rates satisfying the local maximum or local minimum condition in the light intensity change rate smoothing curve, marking them as time candidate points, and constituting a set of time candidate points. Among them, the local maximum corresponds to a sunrise candidate point, and the local minimum corresponds to a sunset candidate point.

[0163] As a specific example, if one of the light intensity change rates in the light intensity change rate smoothing curve presents a significant local minimum near the oscillator count value / time point, the time point is selected as a time candidate point.

[0164] S3-7: calculating the sunrise event significance index or sunset event significance index corresponding to each time candidate point respectively, and selecting the sunrise candidate point with the highest sunrise event significance index and the sunset candidate point with the highest sunset event significance index.

[0165] As a specific example, if the time candidate point corresponds to a local maximum, it is a sunrise candidate point, and its corresponding sunrise event significance is calculated by the following formula:

[0166]

[0167] If the time candidate point corresponds to a local minimum, it is a sunset candidate point, and its corresponding sunset event significance is calculated by the following formula:

[0168] ​​​​​​​​​​​​​​​​

[0169] where a refers to the "left width" of the integration window. It defines the number of time steps to go back (to the left) from the candidate event point to integrate. b refers to the "right width" of the integration window. It defines the number of time steps to go forward (to the right) from the candidate event point to integrate. The values of parameters a and b together determine the range of the time window for calculating the significance . The role of this window is to integrate the rate of change of light intensity in a continuous time window around the candidate point .

[0170] According to the calculated significance indicators of the sunrise / sunset events corresponding to each time candidate point , the time candidate point with the highest significance indicator of the sunrise event is selected as the sunrise candidate point; the time candidate point with the highest significance indicator of the sunset event is selected as the sunset candidate point.

[0171] S3-4 to S3-7 above perform event candidate point detection and significance evaluation. The purpose is to objectively find all possible time candidate points of sunrise / sunset events through mathematical methods (including rate of change calculation, smoothing filtering, extreme value searching) according to the prediction search counting window, and determine one candidate point corresponding to each real sunrise / sunset event by evaluating its reliability (significance indicator).

[0172] S3-8: Perform light intensity and color temperature multimodal cross verification on the sunrise candidate point and the sunset candidate point according to the light intensity rate of change smoothing curve and the color temperature rate of change smoothing curve in the same period to determine the sunrise time and the sunset time.

[0173] As shown in Figure 3 and Figure 4 , a real astronomical (sunrise / sunset) event is accompanied by a specific spectral change. In this embodiment, the verification basis for the sunrise event is that the candidate point of the sunrise event must simultaneously satisfy that the light intensity rate of change and the color temperature rate of change nearby present significant local maximum characteristics. That is, the light intensity and the color temperature need to rise rapidly synchronously. The verification basis for the sunset event is the same: the candidate point of the sunrise event must simultaneously satisfy that the light intensity rate of change and the color temperature rate of change nearby present significant local maximum characteristics.​The light intensity and the color temperature simultaneously present a significant local minimum. That is, the light intensity and the color temperature simultaneously rapidly decrease.

[0174] As a specific example, first, a smooth curve of the light intensity change rate is obtained corresponding to the time, and the smooth curve of the color temperature change rate is obtained in the same way as the smooth curve of the light intensity change rate, which is not described herein. Then, it is determined whether the light intensity change rate and the color temperature change rate simultaneously present a significant local maximum within a certain time range (for example, within a time window of 15 to 30 minutes before and after the candidate sunrise time and the color temperature change rate simultaneously present a significant local minimum; if yes, the candidate sunrise time is determined as the true sunrise time. Similarly, it is determined whether the light intensity change rate and the color temperature change rate simultaneously present a significant local minimum within a certain time range (for example, within a time window of 15 to 30 minutes before and after the candidate sunset time and the color temperature change rate simultaneously present a significant local maximum; if yes, the candidate sunset time

[0175] The above S3-8 step uses the physical law that the light intensity and the color temperature change synchronously to perform secondary verification of the candidate time by light intensity and color temperature double-modal fusion, which can effectively filter out pseudo events caused by light intensity mutation (such as cloud blocking), thereby significantly improving the accuracy of event true time determination. Meanwhile, this step is the key to realizing high robustness in this embodiment, which effectively excludes interference of single modal such as local cloud blocking and flying birds, and ensures that only true astronomical events can be finally confirmed.

[0176] S3-9: The local noon time is calculated according to the sunrise time and the sunset time, and is marked as a reference phase point of the daily time phase axis.

[0177] Specifically, after the sunrise time and the sunset time are determined, the local noon time corresponding to the oscillator count can be calculated by the following formula. The principle is that the local noon time is the midpoint of the sunrise time and the sunset time.

[0178] The reference phase point of the daily time phase axis is also called the absolute reference point (phase =0) of the daily time phase axis. That is, the calculated local noon time is used as the reference phase point for subsequent construction of the local standard optical feature model and time calculation.

[0179] Through the above step S3-9, the determination of the local noon time and the subsequent anchoring of the reference phase point are completed.

[0180] In this embodiment, through the above step S3, the optical periodic characteristics of the local environment are monitored and analyzed in real time, the occurrence time of the sunrise / sunset event can be accurately identified, and the accurate local noon time can be calculated accordingly; at the same time, it is set as the absolute reference origin of the daily time phase axis (T=0), which establishes the time and space reference for all subsequent calculations.

[0181] In step S4, the light intensity data sequence and the color temperature data sequence of continuous M days are mapped to the standard phase axis with the local noon time as the reference phase point, and the local standard optical characteristic model is constructed; wherein M is an integer greater than or equal to 2.

[0182] In some specific embodiments, as shown in Figure 5 and Figure 6 , the light intensity data sequence and the color temperature data sequence collected in continuous M days (M≥2) are respectively mapped to the unified standard phase axis hour with the local noon time as the reference phase point; and the arithmetic mean of all M data points mapped to the same discrete phase point is calculated, thereby constructing the localized standard optical characteristic model:

[0183] .

[0184] In some specific embodiments, the step S4 further comprises:

[0185] The first-order differential values of the light intensity data and the first-order differential values of the color temperature data of each phase point in the local standard optical characteristic model are calculated and stored respectively. That is, the light intensity change rate and the color temperature change rate of each phase point.

[0186] In this embodiment, through the above step S4, the multi-day optical data is aligned and averaged with the local noon time as the reference, and the optical characteristic model representing the typical daily variation law of the local environment is constructed. Therefore, the following three points can be achieved:

[0187] 1. Autonomous learning ability: the model is completely constructed through self-learning, without pre-setting any geographical data.

[0188] 2. Anti-interference filtering capability: through multi-day averaging, single-day random weather interference is effectively smoothed, and stable rules are extracted.​

[0189] 3. Model feature rich: not only contains the absolute value of light intensity / color temperature, but also contains the rate of change model, which provides multi-dimensional features for subsequent high-precision matching.

[0190] In step S5, the instantaneous rate of change of the original light intensity data sequence Iraw(t) and the original color temperature data sequence Craw(t) obtained by real-time acquisition in step S1 is calculated, and the feature vector of the current time is constructed; the Euclidean distance between the feature vector and the model feature vector of the local standard optical feature model is calculated, and the optimal estimation phase is obtained, and the current local time is calculated.

[0191] In some specific embodiments, in the running phase, the original light intensity data sequence Iraw(t) and the original color temperature data sequence Craw(t) collected in real time are respectively marked as the current optical data sequence , and the corresponding instantaneous rate of change is calculated.

[0192] Then, the feature vector Vnow of the current time is constructed .

[0193] Then, the Euclidean distance between the feature vector Vnow and the model feature vector Vmodel of the local standard optical feature model on the standard phase axis is calculated .

[0194] Then, by solving the optimization problem, the optimal estimation phase that makes the Euclidean distance minimum is calculated according to the following formula (6):

[0195] Formula (6): .

[0196] Finally, the current local time is calculated according to the following formula (7):

[0197] Formula (7): .

[0198] In this embodiment, by the above step S5 in the system running phase, the real-time optical data collected is matched with the learned model, and the high-precision local time is accurately calculated. Therefore, the following two points can be achieved:

[0199] 1. Significantly improve the calculation accuracy by multi-modal feature matching: use multi-dimensional feature vectors containing absolute values and rates of change for matching, which has much higher accuracy and robustness than single light intensity threshold judgment.

[0200] ​2. Global Optimal Search: The system determines the time by searching across all candidate phase points on the entire daily timescale, employing a rigorous and scientific method. In other words, this embodiment's system iterates through all candidate phase points on the entire daily timescale, sequentially calculating the real-time feature vector V_now and the standard model V_model for each phase point. The Euclidean distance D(_i) _i); then, through a global search, find the optimal estimated phase that minimizes the Euclidean distance. _est, thus ensuring the global optimality of the time projection results.

[0201] In step S6, latitude and longitude information is calculated based on the system's initial reference date, the calculated daytime duration, and the time difference between the local noon time and Coordinated Universal Time.

[0202] In some specific implementations, the system initial reference date is used as an initial, coarse initial date reference. The system can obtain this information in the following ways:

[0203] a) User input: When the electronic device is first activated, the user inputs the approximate month or season through a simple interface (such as a DIP switch or screen selection);

[0204] b) Pre-configuration: During the manufacturing process of electronic devices, approximate geographical area or time zone information is pre-configured into the device;

[0205] c) Long-term self-learning: If no prior information can be obtained, the system will enter a long-term learning mode, continuously observing the duration of daylight. It analyzes long-term trends and automatically detects key astronomical events such as the summer solstice and winter solstice to deduce absolute dates. However, this method has a relatively long convergence time.

[0206] In some specific implementations, latitude information The calculation process includes:

[0207] The system obtains an initial date reference. Then, based on the sunrise time obtained in the previous steps... and sunset time The length of daylight was calculated. Then, using the solar declination formula or a pre-defined lookup table, the geographical latitude can be deduced. The estimated value. Initial geographical latitude. The estimated value includes a range of uncertainties (e.g.) As data accumulates and the model continues to learn, this range will gradually converge.

[0208] Longitude information The calculation process includes:

[0209] The local noon time obtained according to the preceding step , and the time difference between the local noon time and 12:00 of the world standard time (UTC) corresponds directly to the longitude offset :

[0210] ;

[0211] wherein, ΔT is the time difference between the local noon time and 12:00 of the UTC time;

[0212] If the local noon time occurs after 12:00 of the UTC time, ΔT is positive, corresponding to east longitude; otherwise, ΔT is negative, corresponding to west longitude.

[0213] In the embodiment, the approximate longitude and latitude where the electronic device is located are inversely deduced by using the calculated day length and initial date information through the above step S6. Thus, the time information can be converted into spatial information (longitude and latitude); and the perception of time and space integration is realized, so that the device function is far beyond simple timekeeping.

[0214] Please refer to Figure 7 . Another embodiment of the present application based on the above embodiment also provides a time and space perception system based on light intensity and color temperature dual modal fusion, as shown in Figure 7 , mainly including an optical sensing module, a data processing module, an oscillator and an output module. Among them,

[0215] The optical sensing module is configured to directly perceive the physical environment, and the core function is to synchronously and continuously collect the original signals of the light intensity and color temperature, which are two key optical physical quantities of the environment, to obtain the original light intensity data sequence and the original color temperature data sequence, and to send them to the data processing module.

[0216] The data processing module is configured as a computer program execution core component, responsible for running all computer programs, and also has the function of storing key data and modules; specifically, it can execute all steps contained in the time and space perception method based on light intensity and color temperature dual modal fusion described in the previous embodiment; here, the specific step content will not be described in detail, please refer to the description of the previous embodiment.

[0217] The oscillator, as an oscillator inside the electronic device, is configured to provide a reference counting function for the system.

[0218] and the output module is configured to output the current local time and the longitude and latitude information calculated by the data processing and storage module.

[0219] In some embodiments, the optical sensing module can be an integrated digital ambient light sensor that integrates a luminometer and a colorimeter. The optical sensor can directly output digital quantities that have been internally pre-processed, i.e., a raw light intensity data sequence and a raw color temperature data sequence, thereby simplifying the amount of subsequent signal processing.

[0220] As a specific example, the optical sensing module can be implemented using a spectral sensor of model TSL2591 or AS7341 of AMS, which is connected to the data processing module through an I2C interface and collects light intensity and color temperature data once per minute.

[0221] In some embodiments, the data processing module includes at least a microprocessor and a memory.

[0222] Preferably, the microprocessor uses a processor core with a floating-point unit (FPU), such as an ARM Cortex-M4, which is responsible for executing all algorithmic logic including signal pre-processing, event detection, model learning, and time position calculation. For example, the microprocessor can be implemented using a chip of model RK3576.

[0223] Preferably, the memory preferably uses a non-volatile Flash memory as a computer-readable storage medium for storing computer programs, locally learned standard optical feature models, historical event records, and system configuration parameters, etc. For example, the memory can be implemented using a 1 GB Flash.

[0224] As a specific example, the data processing module can be implemented using a chip of model RK3576 with an additional 1 GB Flash.

[0225] In some embodiments, the oscillator is a high-stability oscillator that can provide a high-precision, high-stability reference clock tick for the electronic device system, in particular to the data processing module.

[0226] Preferably, the oscillator uses a temperature-compensated crystal oscillator (TCXO) or an oven-controlled crystal oscillator (OCXO) with a frequency of 24 MHz and a stability better than ±1 ppm. The clock signal generated by the oscillator will directly drive the clock circuit of the microprocessor in the data processing module.

[0227] In some embodiments, the output module transmits the local time and latitude and longitude information calculated by the data processing module to external devices or upper-layer applications in the form of standard digital signals, such as a display screen, for display.

[0228] Optionally, the output module can include a general serial communication interface, such as UART, I2C, SPI, USB; if the output module is a display terminal, it can also include a screen interface LVDS, V-by-one, eDP and the like.

[0229] Another embodiment of the present application is further extended based on the above two embodiments, and further provides a computer readable storage medium having a computer program stored thereon, wherein the computer program can realize the steps of the time and space perception method based on the fusion of light intensity and color temperature dual mode according to the above embodiment after being executed by the data processing module in the previous embodiment. Here, the specific step content will not be described in detail, and details can be referred to the description of the previous embodiment.

[0230] In some specific embodiments, the computer readable storage medium according to the present embodiment is the memory according to the previous embodiment. The computer program is specifically executed by the microprocessor in the data processing module.

[0231] Those skilled in the art can understand that all or part of the above technical solutions can be realized by a computer program to instruct related hardware, and the program can be stored in a computer readable storage medium. The program can include the above-mentioned method flow when executed. The program can also achieve the beneficial effects of the corresponding method after being executed by the processor.

[0232] The storage medium can be a disk, an optical disc, a read-only memory (ROM) or a random access memory (RAM), etc.

[0233] The present application based on any of the above embodiments further provides a specific application scenario:

[0234] An outdoor charging pile display screen deployed in a service area in northwest China uses the technical solution provided by any of the above embodiments of the present application to realize accurate time synchronization and spatial positioning.

[0235] Deployment and initialization: the installer sets "summer" through local operation, and the system obtains the initial date reference ≈180.

[0236] Self-learning stage: after 3 days of data collection, the system successfully detects that the local noon time occurs at the oscillator count , and measures the average day length ≈ 15 hours; combined with the initial date reference , the latitude of the system is estimated ≈ 44°N; meanwhile, the local light intensity-color temperature-phase model is successfully constructed.

[0237] Running phase: in the afternoon of a certain day, the system collects real-time data: = 55,000 lux, = -1,800 lux / minute, = 5,300 K, = -50 K / minute. After matching with the model, the current phase = +5.2 hours. Therefore, the local time is determined as + 5.2 hours = 17:12; meanwhile, the longitude is estimated as ≈ 85°E.

[0238] Output display: the system displays the time "17:12" and the location "44°N, 85°E" through the display screen. The main controller displays the accurate local time accordingly and automatically adjusts the light according to the built-in "dusk" mode, realizing energy saving.

[0239] As can be seen, the spatiotemporal perception scheme based on light intensity and color temperature dual modal fusion provided by the application can adapt to different time zones autonomously, provide accurate time in line with local people's work and rest habits, and has geographical location perception capability, which embodies its superior autonomy and practicality.

[0240] In summary, the spatiotemporal perception method, system and medium based on light intensity and color temperature dual modal fusion provided by the application have the following characteristics:

[0241] (1) complete independence and high autonomy: completely free from dependence on any external time, space reference facility, realizing full autonomous spatiotemporal perception capability in any place with natural light circulation.

[0242] (2) global universality and adaptive learning: through self-learning mechanism to construct local model, can automatically adapt to different optical environments from equator to polar circle, from desert to coast around the world, without preloading any geographical data.

[0243] (3) high robustness: adopting multi-day data average to construct model, and fusing multi-modal features of absolute values and change rates of light intensity and color temperature, can effectively suppress the interference brought by single-day weather mutation (such as rain, haze, short-term cloud cover), and guarantee the stability of output.

[0244] (3) dual output and high value information: simultaneously output high-precision continuous local time and valuable approximate longitude and latitude information, the function is far beyond simple day and night state discrimination, providing rich context for upper-layer application.

[0245] (4) Low cost and low power consumption: The required hardware devices are all general-purpose low-cost components, with costs significantly lower than existing solutions; and the power consumption is extremely low, making it suitable for IoT applications that require long-term battery power.

[0246] It should be noted that any reference signs placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

[0247] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0248] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0249] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0250] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0251] In the present disclosure, unless specifically stated and limited otherwise, a first feature "on" or "under" a second feature can be directly in contact with the second feature, or indirectly in contact with the second feature through an intermediate medium. Also, a first feature "over", "above" and "on top of" a second feature can be directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature "under", "below" and "underneath" a second feature can be directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0252] In the description of the present disclosure, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In the present disclosure, the illustrative description of the above terms should not be understood as necessarily referring to the same embodiment or example. Also, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the skilled person in the art can combine and combine the features of different embodiments or examples described in the present disclosure and the features of different embodiments or examples, without contradiction.

[0253] The above description is only an embodiment of the present disclosure, and does not limit the patent scope of the present disclosure, and any equivalent transformation or direct or indirect application in the related technical field by using the content of the present disclosure and the drawings is also included in the patent protection scope of the present disclosure.

Claims

1. A spatio-temporal perception method based on light intensity and color temperature dual-modal fusion, characterized in that, The method comprises the following steps: Step S1: preprocessing the collected original light intensity data sequence and original color temperature data sequence to obtain a light intensity data sequence and a color temperature data sequence; Step S2: detecting the oscillator count difference between two consecutive sunrise events in the light intensity data sequence according to the light intensity change rate, and calibrating the effective count value of the oscillator according to the oscillator count difference; Step S3: calculating the local noon time according to the light intensity data sequence, the color temperature data sequence and the effective count value, and marking it as the reference phase point of the daily time phase axis; Step S4: mapping the light intensity data sequence and the color temperature data sequence of consecutive M days to the standard phase axis with the reference phase point as the coordinate origin, and constructing a local standard optical feature model; wherein M is an integer greater than or equal to 2; Step S5: calculating the instantaneous change rate of the real-time collected original light intensity data sequence and original color temperature data sequence respectively, and constructing a feature vector at the current time; calculating the Euclidean distance between the feature vector and the model feature vector of the local standard optical feature model, and obtaining the optimal estimation phase, and then calculating the current local time; Step S6: calculating the latitude and longitude information according to the system initial reference date, the calculated day length and the time difference between the local noon time and the world standard time.

2. The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, characterized in that, The step S1 specifically comprises: S1-1: performing sliding average filtering on the real-time collected original light intensity data sequence and original color temperature data sequence respectively to obtain a filtered light intensity data sequence and a filtered color temperature data sequence; S1-2: Calculate the corresponding current time in the filtered light intensity data sequence and the filtered color temperature data sequence, respectively. The change in adjacent sampling points is considered. If the change in adjacent sampling points exceeds a dynamic threshold, then the current sampling point is discarded. The sampling points are determined, and data repair is performed based on the previous sampling point or the linear interpolation method to obtain the light intensity data sequence and the color temperature data sequence, respectively. Wherein, real-time according to the filtered light intensity data sequence and the filtered color temperature data sequence respectively at the current time The previous L continuous sampling points: , according to the following formula (1) dynamic update their respective dynamic threshold ; Equation (1): ; Wherein, L is an integer greater than or equal to 2; the value range of α is 1-15, which is the sensitivity preset according to the actual application scene; j is the summation index variable, j=k-L to j=k-1. 3.The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 2, wherein, the dynamic updating of the respective dynamic threshold value comprising the following sub-steps: Substep A, set the length of the historical data sliding window as L, L is an integer greater than or equal to 2; Sub-step B, acquiring, according to the historical data sliding window, current time points from the filtered light intensity data sequence and the filtered color temperature data sequence respectively previous L continuous sampling points: ; Substep C, according to the sampling points obtained in the substep B, calculate the recent average change rate of light intensity data and the recent average change rate of color temperature data respectively according to the following formula (1-1); Equation (1-1): ; Sub-step D, the dynamic threshold values of the light intensity data and the color temperature data are respectively calculated according to the following formula (1-2) , and the dynamic threshold values of the previous steps are updated ; Equation (1-2): ; Wherein, the value range of α is 1-15, which is the sensitivity preset according to the actual application scene.

4. The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, characterized in that, The step S2 specifically comprises: S2-1: Calculate the ideal count value for one natural day cycle according to the following equation (2) ; Equation (2): ; wherein, is the ideal period; is the nominal frequency of the oscillator; S2-2: detecting the difference of the oscillator count between two consecutive sunrise events in the light intensity data sequence according to the light intensity change rate and mark it as a natural day cycle; S2-3: Calculate the calibration factor according to the following equation (3) ; Equation (3): ; S2-4: The effective count value of the oscillator is calculated according to the following equation (4) and the original count value n of the oscillator is calibrated as the effective count value ; Equation (4): .

5. The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, characterized in that, The step S3 specifically comprises: S3-1 : The oscillator count difference detected in step S2 is divided by the number of clock cycles per second of the oscillator , is marked as one natural day cycle; S3-2: identifying the sunrise count value and / or sunset count value corresponding to the first and the next consecutive N natural day cycles according to the light intensity data sequence and the color temperature data sequence; N is an integer greater than or equal to 2; S3-3: analyzing the predicted search count window and the count tolerance value according to the sunrise count value and / or sunset count value corresponding to the first and the next consecutive N natural day cycles; S3-4: Calculate the light intensity variation rate of adjacent light intensity data in the light intensity data set corresponding to the prediction search count window in the light intensity data sequence according to the following formula (5) ; Equation (5): ; wherein is a preset time resolution of data collection in step S1; is the light intensity data at time t in the light intensity data sequence; S3-5: the light intensity change rate carrying out sliding smoothing filtering processing to obtain a light intensity change rate smoothing curve ; S3-6: According to the light intensity change rate smooth curve , all time candidate points satisfying the local maximum or local minimum condition are obtained ; S3-7: calculate each time candidate point respectively corresponding sunrise event saliency index or sunset event saliency index, select the sunrise candidate point with the highest sunrise event saliency index and the sunset candidate point with the highest sunset event saliency index; S3-8: According to the light intensity change rate smooth curve The simultaneous period color temperature change rate smooth curve carries out light intensity and color temperature multimodal cross verification on the sunrise candidate point and the sunset candidate point, and determines the sunrise time and the sunset time. S3-9: calculating the local noon time according to the sunrise time and the sunset time, and marking it as the reference phase point of the daily time phase axis.

6. The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, characterized in that, The step S4 further comprises: Taking the arithmetic mean of all data points mapped to the same discrete phase point as the data point of the corresponding phase point.

7. The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, characterized in that, The step S4 further comprises: The first-order differential value of the light intensity data and the first-order differential value of the color temperature data of each phase point in the local standard optical feature model are calculated and stored respectively. 8.The spatio-temporal perception method based on light intensity and color temperature dual-modal fusion according to claim 1, wherein, The optimal estimated phase in the step S5 is obtained, and the current local time is calculated, comprising: The optimal estimated phase is calculated according to the following formula (6); Equation (6): ; wherein is the Euclidean distance; The current local time is calculated according to the following formula (7); Equation (7): ; wherein is local mean time; is the optimal estimated phase.

9. A spatio-temporal perception system based on dual-modal fusion of light intensity and color temperature, characterized in that, Comprising: An optical sensing module configured to acquire a raw light intensity data sequence and a raw color temperature data sequence; A data processing module configured to execute the steps included in the time-space perception method based on the fusion of the light intensity and color temperature dual modalities according to any one of the above claims 1 to 8; An oscillator configured to provide a reference counting function for the system; And an output module configured to output the current local time and the latitude and longitude information calculated by the data processing and storage module.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program can realize the steps included in the time-space perception method based on the fusion of the light intensity and color temperature dual modalities according to any one of the above claims 1 to 8 after being executed by the data processing module in the above claim 9.

Images