Intelligent control method and system for cover lifting of waterproof light box

Abstract

The application relates to the technical field of waterproof lamp boxes, and discloses a cover lifting intelligent control method and system for a waterproof lamp box, the method comprising the following steps: acquiring humidity and environmental dynamic parameters; pre-processing the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental characteristic data set. When the humidity does not exceed a preset humidity interval, it is determined that the waterproof lamp box is normal. When the humidity exceeds the preset humidity interval, it is determined that the waterproof lamp box is abnormal, and the environmental characteristic data set is input into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range. The air heat exchange power is calculated according to the environmental dynamic parameters, the initial cover lifting angle range and the air heat exchange power are optimized based on a genetic algorithm, the optimal population is obtained through iterative optimization, and the individual with the highest fitness is selected as the cover lifting angle. The cover lifting time is calculated to obtain an initial cover lifting time, and the initial cover lifting time and the cover lifting angle are optimized based on a PSO algorithm to obtain a cover lifting control instruction. The method can realize dynamic and accurate cover lifting control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of waterproof light boxes, in particular to a cover lifting intelligent control method and system for a waterproof light box. BACKGROUND

[0002] Waterproof light boxes are widely used in outdoor advertising, traffic signs, public lighting and other fields. Their main function is to provide stable visual information display under various environmental conditions. Since waterproof light boxes are exposed to outdoor environments for a long time, they need to cope with the influence of rain, humidity, wind speed, light and other external factors. In particular, in high humidity environments, water vapor condensation is easily produced inside the light box, affecting the stability of electronic components and even shortening the service life of the equipment. Therefore, intelligent environmental regulation technology for waterproof light boxes has become a research focus. Through a reasonable cover lifting mechanism and control system, internal air circulation is improved, humidity accumulation is reduced, and the reliability and durability of the equipment in complex environments are ensured.

[0003] The existing waterproof light box environmental control technology is mainly based on a fixed threshold triggering mechanism. When the internal humidity of the light box is detected to exceed the preset threshold, the system performs a cover lifting action to increase air circulation and reduce humidity. Some solutions also trigger the cover lifting when the temperature is too high to help the equipment dissipate heat. This type of control method presets fixed humidity, temperature or wind speed thresholds. Once the external environmental parameters meet the conditions, the cover lifting or closing operation is performed according to the predetermined mode. For example, if the humidity is below a certain threshold, the cover lifting mechanism will automatically close to reduce the further entry of external air. However, since this type of method relies on fixed parameter settings, it fails to comprehensively consider the linkage of environmental change trends, light levels, wind speeds and other factors, making the cover lifting control relatively single and difficult to dynamically adapt to changing outdoor environments.

[0004] In summary, the core feature of the existing technology is static adjustment. Its cover lifting control method relies on preset operation logic and fails to intelligently optimize for real-time environmental changes. Due to the lack of comprehensive analysis of multi-source environmental data, the system cannot dynamically adapt to different climate conditions and humidity changes, resulting in poor adaptability of the cover lifting control and limited precision, causing problems such as excessive or lagging cover lifting in complex environments, affecting the overall operation effect. SUMMARY

[0005] The present application provides a cover lifting intelligent control method and system for a waterproof light box to achieve dynamic and precise cover lifting control.

[0006] In a first aspect, to solve the above technical problems, the present application provides a cover lifting intelligent control method for a waterproof light box, comprising:

[0007] obtaining humidity and environmental dynamic parameters;

[0008] Preprocess the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature dataset;

[0009] When the humidity does not exceed the set normal humidity range, it is determined that the waterproof light box is normal.

[0010] When the humidity exceeds the set normal humidity range, it is determined that the waterproof light box is abnormal, and the environmental feature dataset is input into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range.

[0011] According to the environmental dynamic parameters, the air heat exchange power is calculated to obtain the air heat exchange power.

[0012] Based on a genetic algorithm, the initial cover lifting angle range and the air heat exchange power are used to generate an initial optimization population, and the initial optimization population is iteratively optimized to obtain an optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover lifting angle and output.

[0013] According to the environmental dynamic parameters, the cover lifting time is calculated to obtain an initial cover lifting time.

[0014] Based on a PSO algorithm, the initial cover lifting time and the cover lifting angle are iteratively optimized to obtain a cover lifting control instruction.

[0015] Preferably, the environmental dynamic parameters include: illumination data, temperature data, and wind speed.

[0016] The illumination data includes: illumination intensity, illumination change rate, illumination duration, and sunshine angle.

[0017] The temperature data includes: the temperature inside the light box and the temperature outside the light box.

[0018] Preferably, the preprocessing of the humidity and the environmental dynamic parameters based on the data fusion algorithm to obtain the environmental feature dataset comprises:

[0019] The humidity and the environmental dynamic parameters are normalized to obtain a normalized multi-source environmental dataset.

[0020] Based on a Kalman filtering method, the multi-source environmental dataset is subjected to noise filtering to obtain a denoised multi-source environmental dataset.

[0021] Based on a principal component analysis method, the denoised multi-source environmental dataset is subjected to feature extraction to obtain an environmental feature dataset.

[0022] The environmental feature data set includes humidity features, light intensity, light change rate features, light duration features, sunshine angle features, light box internal temperature features, light box external temperature features, and wind speed features.

[0023] Preferably, when the humidity exceeds the normal humidity interval, it is determined that the waterproof light box is abnormal, and the environmental feature data set is input into the pre-constructed dynamic adjustment model to obtain an initial cover angle range, including:

[0024] The dynamic adjustment model is trained by an LSTM neural network model;

[0025] The feature data set is adjusted in time dimension through the input layer of the dynamic adjustment model to obtain preprocessed time series data;

[0026] The preprocessed time series data is identified in time series through the LSTM layer of the dynamic adjustment model to obtain a time series feature vector;

[0027] The time series feature vector is mapped and converted through the fully connected layer of the dynamic adjustment model to obtain an environmental feature embedding vector;

[0028] The environmental feature embedding vector is calculated through the output layer of the dynamic adjustment model to obtain an initial cover angle range.

[0029] Preferably, the air heat exchange power is calculated according to the environmental dynamic parameters to obtain an air heat exchange power, including:

[0030] The air heat exchange power is calculated by the following formula:

[0031]

[0032] wherein, is the air heat exchange power, is the air specific heat capacity, is the wind speed, is the light box internal temperature, is the light box external temperature, is the air density.

[0033] Preferably, an initial optimization population is generated based on a genetic algorithm on the initial cover angle range and the air heat exchange power, and the initial optimization population is iteratively optimized to obtain an optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover angle and output, including:

[0034] Based on the initial cover angle range, a plurality of cover angle individuals are generated, and the air heat exchange power corresponding to the cover angle individuals is calculated to obtain an initial optimization population;

[0035] According to the air heat exchange power of the initial optimization population, the initial optimization population is iteratively optimized to obtain an optimized population;

[0036] Based on a preset optimization condition, it is judged whether the optimized population meets an optimization termination requirement. If yes, the optimized population is taken as an optimal population, and an individual with the highest fitness value in the optimal population is selected as a cover lifting angle and outputted;

[0037] If the optimized population does not meet the preset optimization condition, the genetic operation is continuously performed until the optimization termination condition is met;

[0038] The preset optimization condition is fitness value convergence or reaching a maximum iteration number.

[0039] Preferably, the cover lifting time is calculated according to the environmental dynamic parameters to obtain an initial cover lifting time,

[0040] The initial cover lifting time is calculated by the following formula:

[0041]

[0042] wherein, is the initial cover lifting time, is the temperature inside the light box, is the temperature outside the light box, is the light intensity, is the light duration, is the light change rate, is the sunshine angle, is a normalization coefficient, is a time adjustment factor, is an angle influence coefficient.

[0043] Preferably, the initial cover lifting time and the cover lifting angle are iteratively optimized based on the PSO algorithm to obtain a cover lifting control instruction, which comprises:

[0044] The initial cover lifting time and the cover lifting angle are initialized to obtain an initial state particle swarm;

[0045] The fitness of each particle in the initial state particle swarm is calculated to obtain a fitness value of each particle;

[0046] Based on the fitness value of each particle, the optimal fitness is calculated to obtain an optimal fitness value;

[0047] If the optimal fitness value reaches the preset optimization termination condition, a particle with the highest fitness value is selected as a final optimization result, and a cover lifting time and a cover lifting angle of the particle are selected as the cover lifting control instructions and outputted;

[0048] If the optimal fitness value does not satisfy the preset optimization termination condition, the cover lifting time and the cover lifting angle of the initial state particle group are updated and iteratively optimized until the preset optimization termination condition is reached.

[0049] The cover lifting control instructions include a cover lifting time and a cover lifting angle.

[0050] In a second aspect, the present application provides a waterproof light box cover lifting intelligent control system, comprising:

[0051] A data acquisition module is configured to acquire humidity and environmental dynamic parameters.

[0052] A preprocessing module is configured to preprocess the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature data set.

[0053] An abnormality determination module is configured to determine that the waterproof light box is normal when the humidity does not exceed a set normal humidity range.

[0054] A dynamic adjustment module is configured to determine that the waterproof light box is abnormal when the humidity exceeds the set normal humidity range, and input the environmental feature data set into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range.

[0055] An air heat exchange power module is configured to calculate an air heat exchange power based on the environmental dynamic parameters to obtain the air heat exchange power.

[0056] A genetic algorithm module is configured to generate an initial optimization population based on a genetic algorithm for the initial cover lifting angle range and the air heat exchange power, iteratively optimize the initial optimization population, obtain an optimal population, and select an individual with the highest fitness value in the optimal population as a cover lifting angle and output the cover lifting angle.

[0057] A cover lifting time module is configured to calculate an initial cover lifting time based on the environmental dynamic parameters.

[0058] A cover lifting control instruction module is configured to iteratively optimize the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain cover lifting control instructions.

[0059] In a third aspect, the present application also provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the intelligent cover lifting control method of the waterproof light box according to any one of the preceding aspects when executing the computer program. In a fourth aspect, the present application also provides a computer readable storage medium comprising a stored computer program, wherein the computer readable storage medium performs the intelligent cover lifting control method of the waterproof light box according to any one of the preceding aspects when the computer program runs.

[0060] Compared with the prior art, the present application has the following beneficial effects:

[0061] (1) The present application obtains environmental dynamic parameters such as illumination data, temperature data, and wind speed, and pre-processes them in combination with a data fusion algorithm to obtain an environmental feature data set, thereby realizing comprehensive analysis of multiple environmental variables. Compared with the fixed rule control mode used in the prior art, the present application can adapt to changes in different external environments, thereby improving the flexibility and adaptability of cover lifting control.

[0062] (2) The present application optimizes the initial cover lifting angle range using a genetic algorithm, ensures that the cover lifting angle matches the current environmental conditions, and further optimizes the cover lifting time and cover lifting angle based on a particle swarm optimization (PSO) algorithm to form a cover lifting control instruction. Compared with traditional static control schemes, the present application can dynamically adjust the cover lifting parameters in real time based on the calculation results, thereby improving the accuracy and energy efficiency ratio of cover lifting control.

[0063] (3) The present application constructs a dynamic adjustment mechanism through an LSTM neural network model, can predict the trend of environmental changes and adjust the cover lifting angle in advance, realizes real-time and accurate cover lifting control in combination with an optimization algorithm, and effectively improves the environmental adaptability and system reliability of the waterproof light box.

[0064] In summary, the present application breaks through the static adjustment limitations of the prior art in cover lifting control through the technical architecture of environmental perception + intelligent optimization + dynamic adjustment, and realizes dynamic and accurate cover lifting control. BRIEF DESCRIPTION OF DRAWINGS

[0065] Figure 1 is a schematic diagram of the intelligent cover lifting control method of the waterproof light box provided by the first embodiment of the present application;

[0066] Figure 2 is a schematic diagram of the intelligent cover lifting control system of the waterproof light box provided by the second embodiment of the present application. DETAILED DESCRIPTION

[0067] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by those skilled in the art without creative efforts belong to the scope of the present application.

[0068] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by those skilled in the art without creative efforts belong to the scope of the present application. Figure 1 The first embodiment of the present application provides a cover lifting intelligent control method of a waterproof light box, comprising the following steps:

[0069] S11, obtaining humidity and environmental dynamic parameters;

[0070] S12, preprocessing the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature data set;

[0071] S13, when the humidity does not exceed a set normal humidity range, determining that the waterproof light box is normal;

[0072] S14, when the humidity exceeds the set normal humidity range, determining that the waterproof light box is abnormal, and inputting the environmental feature data set into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range;

[0073] S15, calculating air heat exchange power based on the environmental dynamic parameters to obtain the air heat exchange power;

[0074] S16, generating an initial optimization population based on a genetic algorithm on the initial cover lifting angle range and the air heat exchange power, and iteratively optimizing the initial optimization population to obtain an optimal population, and selecting an individual with the highest fitness value in the optimal population as a cover lifting angle for output;

[0075] S17, calculating cover lifting time based on the environmental dynamic parameters to obtain an initial cover lifting time;

[0076] S18, iteratively optimizing the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain a cover lifting control instruction.

[0077] In step S11, the humidity and the environmental dynamic parameters are obtained

[0078] The environmental dynamic parameters include illumination data, temperature data, and wind speed;

[0079] The illumination data includes illumination intensity, illumination change rate, illumination duration, and sunshine angle

[0080] The temperature data includes light box internal temperature and light box external temperature.

[0081] It is worth mentioning that in step S11, the system first performs the acquisition of humidity and environmental dynamic parameters for subsequent intelligent control and dynamic adjustment. The acquisition of humidity and environmental dynamic parameters is the basis for the intelligent control of the waterproof light box cover, directly affecting the optimization calculation of the cover time and cover angle. The execution of this step includes the selection of sensors, the specific way of data collection, the data preprocessing process, and the parameter storage method.

[0082] The acquisition of humidity parameters uses SHT30 digital temperature and humidity sensor, which has high precision and low power consumption, supports I2C interface communication, and can provide real-time humidity data of the internal environment of the light box. The sensor is installed in the environmental monitoring module inside the light box and communicates with the MCU (such as STM32F103C8T6). Every 5 seconds, the MCU sends a trigger instruction through the I2C bus, and the SHT30 sensor returns the current environmental humidity value, which is expressed in percentage (%RH). The data is stored in the system buffer area and used to determine the humidity state of the light box. When the humidity exceeds 80% RH (set threshold), the system determines that the internal humidity of the waterproof light box is abnormal, triggering the dynamic cover strategy.

[0083] The acquisition of environmental dynamic parameters involves light data, temperature data, and wind speed, which are collected by multiple sensors and preprocessed using data fusion algorithms to ensure data accuracy and stability. The acquisition of light data uses BH1750 light sensor, which can directly measure ambient light intensity with an output unit of lux. The sensor is installed on the top of the light box and communicates with the MCU through the I2C interface. Every 2 seconds, the MCU reads the measurement value of the BH1750 sensor and stores it in the environmental parameter database. The light intensity is used to determine whether the external light environment is sufficient, for example, when the external light intensity is lower than 50 lux, the system can trigger the light compensation of the light box. The light change rate is calculated by the light intensity of the adjacent two measurements, reflecting the trend of external environmental light. For example, if the light intensity decreases from 500 lux to 200 lux within 10 seconds, the system determines that the external light environment changes dramatically due to cloud cover or weather changes, triggering the cover adjustment. The light duration is determined by time series analysis of continuous light data, for example, when the light intensity is continuously lower than 50 lux for more than 30 minutes, the system can infer that the external light is insufficient and adjust the cover strategy. The sunlight angle is calculated by the light sensor combined with the IMU sensor (such as MPU6050), and the system calculates the solar azimuth angle based on the current time and geographic location to optimize the cover angle of the light box and improve the light utilization rate.

[0084] The temperature data is acquired by DS18B20 digital temperature sensor, which uses single bus protocol and supports multiple sensor cascade measurement. The system installs one DS18B20 sensor inside and outside the light box respectively to measure the internal and external temperature. The internal temperature of the light box is read by the MCU through the single bus DS18B20 data and stored in the environmental database in Celsius (°C). When the internal temperature is higher than 40°C, the system judges that the internal environment of the light box is overheated and needs to increase the air circulation efficiency. The external temperature of the light box is mainly used to calculate the temperature difference. If the external temperature is low and the internal temperature is high, the system can reduce the cover angle to reduce heat loss.

[0085] The wind speed is measured by FS3000-1005 ultrasonic wind speed sensor, which is installed on the side wall of the waterproof light box and can detect the flow of external air in real time. The wind speed measured by the sensor ranges from 0 to 30 m / s and is converted to digital data by the ADC module, which converts the collected analog signal to digital data and transmits it to the control system for calculation. The update frequency of wind speed data is 1 second / second to ensure that the system can obtain real-time wind speed information for subsequent cover angle adjustment.

[0086] All environmental data is stored in EEPROM or Flash memory and preprocessed with Kalman filter algorithm to remove noise and improve data accuracy. The system updates the stored data every minute and provides it to the subsequent dynamic adjustment model to optimize the cover control strategy.

[0087] In step S12, the humidity and the environmental dynamic parameters are preprocessed based on a data fusion algorithm to obtain an environmental feature data set, including:

[0088] The humidity and the environmental dynamic parameters are normalized to obtain a normalized multi-source environmental data set;

[0089] The multi-source environmental data set is filtered based on the Kalman filter method to obtain a denoised multi-source environmental data set;

[0090] The denoised multi-source environmental data set is feature extracted based on the principal component analysis method to obtain an environmental feature data set;

[0091] The environmental feature data set includes humidity features, light intensity, light change rate features, light duration features, sunlight angle features, light box internal temperature features, light box external temperature features and wind speed features.

[0092] It is worth mentioning that in step S12, the humidity and environmental dynamic parameters are pre-processed based on a data fusion algorithm to obtain the environmental feature dataset. The main purpose of data preprocessing is to eliminate outliers in environmental parameters, reduce noise interference, and improve the effectiveness and representativeness of data through reasonable feature extraction methods. This step covers three main links of normalization processing, Kalman filter noise reduction and principal component analysis feature extraction.

[0093] First, after obtaining the humidity and environmental dynamic parameters, normalization processing is needed. The environmental parameters are derived from different sensors, and the measured values of different sensors have different units and ranges. For example, the humidity sensor outputs relative humidity values, with a value range of 0 to 100, while the wind speed sensor outputs values of 0 to 30 meters per second. If the data is not normalized, different dimensional data will cause errors in subsequent calculations, affecting the analysis results. Therefore, normalization methods are used to adjust all data to the same numerical range, such as 0 to 1, to ensure that different features have the same weight in the calculation process and improve the stability of the calculation. Specifically, humidity data will be converted into a standardized value, so that the minimum humidity corresponds to 0 and the maximum humidity corresponds to 1, and the intermediate values are mapped to the interval in a linear proportion. Similarly, wind speed, light intensity and temperature data will also undergo the same normalization process to make them on the same numerical scale for comparison and calculation.

[0094] After completing the normalization process, the data needs to be denoised. When sensors collect data, they may be affected by environmental interference, such as short-term dramatic fluctuations in wind speed sensors due to turbulence around the airflow, and sudden changes in data due to shadow obstruction affecting light sensors. These noises can reduce the reliability of the data, thereby affecting subsequent feature extraction and calculation results. Therefore, the Kalman filter method is used to filter the noise of the normalized multi-source environmental data set. Kalman filtering is a recursive filtering method suitable for time series data, which can smooth the data by estimating the optimal state of the current data point to reduce random errors. For example, in wind speed data processing, if the sensor outputs a data point much higher or lower than the previous and next values at a certain time, the Kalman filter will correct the current data based on historical data and prediction models, thereby reducing the impact of sudden values and making the data more stable and reliable. Similarly, for temperature data, if there is a dramatic fluctuation in a short period of time (for example, the temperature inside the light box suddenly rises by 10 degrees Celsius at a certain moment, while the data before and after changes smoothly), the filter will judge that the data point is an outlier and replace it with a more smooth trend value to ensure the continuity and reasonableness of the data.

[0095] After the noise reduction process is completed, feature extraction needs to be performed on the de-noised multi-source environmental data set to extract the most representative information and improve the efficiency and accuracy of subsequent calculations. Principal component analysis is used for feature extraction, which can reduce the dimensionality of the data while preserving the main information. In the multi-source environmental data set, although multiple parameters such as humidity, light intensity, light change rate, light duration, sunshine angle, lamp box internal temperature, lamp box external temperature, and wind speed are collected by sensors, not all parameters are equally important for flap control. For example, in some cases, the light change rate is more reflective of the dynamic changes in the external environment than the light intensity, and the wind speed is more influential on the stability of the internal temperature of the lamp box than the humidity. Principal component analysis combines variables with high correlation by calculating the correlation between variables, thereby reducing the dimensionality of the data. For example, in the light data processing process, light intensity, light change rate, and light duration have high correlation, so they can be combined into a comprehensive feature to reduce the amount of calculation and avoid the interference of redundant information.

[0096] In the application process, the results of feature extraction are stored as environmental feature data sets and used for subsequent flap control calculations. For example, the environmental data acquisition system obtains the following data at a certain time point: humidity 60%, light intensity 500 lux, light change rate 5%, light duration 3 hours, sunshine angle 30 degrees, lamp box internal temperature 35 degrees Celsius, lamp box external temperature 25 degrees Celsius, and wind speed 10 meters per second. After normalization, these data are converted to a standardized range, such as humidity to 0.6, light intensity to 0.5, light change rate to 0.05, and wind speed to 0.33. Then, Kalman filtering is used to smooth the data, such as sudden changes in light change rate, which are adjusted to values more consistent with the overall trend. Finally, principal component analysis is used to extract the most representative features, such as combining light data into a principal component variable and extracting the combined effect of temperature and wind speed into another variable, thereby reducing computational complexity and improving system response speed.

[0097] In step S13, when the humidity does not exceed the set normal humidity range, it is determined that the waterproof lamp box is normal;

[0098] It is worth noting that in step S13, when it is detected that the current environmental humidity does not exceed the pre-set normal humidity range, it is determined that the waterproof lamp box is in a normal state and no flap adjustment operation is needed. The core of this step is to set a reasonable normal humidity range and determine whether the humidity is within this range based on sensor data to ensure the stability of the internal environment of the lamp box while reducing unnecessary flap movements, reducing energy consumption and mechanical wear.

[0099] In practical applications, the setting of the normal humidity range is based on the tolerance range of the internal components of the light box to humidity, environmental climate conditions, and the actual use requirements of the waterproof light box. Generally speaking, electronic components can work stably in a relative humidity range of 40% to 70%, so the normal humidity range is set to 40% to 70%. When the environmental humidity is in this range, the sealing and protection performance of the waterproof light box is sufficient to ensure the normal operation of the internal electronic equipment, avoiding static accumulation due to too low humidity or electronic component short circuit due to too high humidity.

[0100] The humidity detection uses a high-precision humidity sensor installed inside or outside the light box to monitor the humidity level by regularly collecting data. For example, at a certain time point, the system detects that the internal humidity of the light box is 55% and the external humidity is 60%. Since 55% falls within the normal range of 40% to 70%, the system determines that the internal humidity of the light box is stable and does not need to perform the cover lifting operation. In another scenario, if the external humidity of the light box is 80%, but the internal humidity remains at 65%, the cover lifting operation will not be triggered either, because the waterproof performance of the light box can effectively isolate the influence of external humidity.

[0101] In step S14, when the humidity exceeds the normal humidity range, it is determined that the waterproof light box is abnormal, and the environmental feature data set is input into the pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range, including:

[0102] The dynamic adjustment model is trained by an LSTM neural network model;

[0103] The time dimension of the feature data set is adjusted through the input layer of the dynamic adjustment model to obtain pre-processed time series data;

[0104] The pre-processed time series data is subjected to time series recognition through the LSTM layer of the dynamic adjustment model to obtain a time series feature vector;

[0105] The time series feature vector is subjected to mapping conversion through the fully connected layer of the dynamic adjustment model to obtain an environmental feature embedding vector;

[0106] The environmental feature embedding vector is calculated through the output layer of the dynamic adjustment model to obtain an initial cover lifting angle range.

[0107] It is worth mentioning that the training process of LSTM neural network mainly includes data preparation, forward propagation, loss calculation, back propagation and parameter update, and the training will continue until the set termination condition is met. First, in the data preparation stage, the system extracts time series features from a large amount of historical environmental data and divides them into training set, validation set and test set. The training data contains humidity, light intensity, light change rate, light duration, sunshine angle, temperature and wind speed and other parameters. Each input sample is constructed in a sliding window manner, for example, set the time window length to 10 minutes, that is, the model inputs 10 minutes of continuous environmental data each time to predict the cover angle at the next time step.

[0108] In the forward propagation process, LSTM receives input data and filters and stores information through the forget gate, input gate and output gate. The model will calculate the cell state based on the input of the current time step and the hidden state of the previous time step, and update the hidden state, so as to capture the long-term dependence of time series data. The output of the LSTM layer is mapped to the cover angle range through the fully connected layer, and the final prediction value is generated by the output layer.

[0109] In order to measure the prediction effect of the model, the system uses mean square error (MSE) as the loss function to calculate the error between the predicted value and the true value. After loss calculation, the error signal will be transmitted back to each network layer through the back propagation algorithm, and the gradient will be calculated using the truncated back propagation through time (BPTT) method to update the weights and biases of LSTM. The optimization algorithm uses Adam optimizer, and the learning rate is initially set to 0.001 and dynamically adjusted during training to ensure training stability and convergence speed.

[0110] During the training process, the system will continuously iterate and update the parameters of LSTM to gradually reduce the value of the loss function. The training termination condition is set as one of the following three conditions: first, the training round reaches the maximum upper limit, generally set to 200 rounds, if the model still does not reach the convergence standard after 200 rounds of training, stop training; second, the validation set loss decreases by less than 0.0001 in the last 10 rounds of training, indicating that the model has reached the optimal state and no longer needs to be significantly optimized; third, the training set loss is lower than the set threshold, generally set to MSE less than 0.01, to ensure that the model accuracy is high enough. If any of the conditions are met, the training process is terminated, and the model with the smallest error on the validation set is selected as the final training result.

[0111] For example, during actual training, if the model's training set MSE has dropped to 0.0095 at the 120th round, and the validation set MSE has changed by less than 0.0001 in the past 10 rounds of training, the system will trigger the early termination mechanism and stop training. The basis for this setting is to avoid overfitting and improve the model's generalization ability in real-world environments. If the training round is set too high, such as 500 rounds, it will cause the model to overfit the training data, making its prediction effect in real-world scenarios decline. Therefore, setting a reasonable termination condition is crucial for the training of the LSTM model, ensuring that it is effectively trained while avoiding waste of computing resources and improving the real-time performance and accuracy of the waterproof light box intelligent control system.

[0112] Abnormal changes in humidity can affect the internal environment of the light box, which in turn affects the performance and service life of the electronic components inside the light box. Therefore, when the humidity exceeds the normal range set, the system needs to be adjusted in time to ensure the stability of the internal environment of the light box. To this end, this method is based on the LSTM neural network model, which dynamically adjusts the cover angle through analysis and calculation of the environmental feature dataset, thereby optimizing the air circulation effect of the light box and improving environmental adaptability.

[0113] First, the system will continuously monitor the humidity parameters inside the light box and compare them with the preset normal humidity range. In this scheme, the normal range of humidity is set between 40% and 70%, which is determined based on the best working environment for the electronic components inside the light box. If the humidity is below 40%, the air is too dry, which can cause static electricity to accumulate, affecting the normal operation of the circuit board. If the humidity exceeds 70%, condensation occurs inside the light box, causing the electronic components to be damp and increasing the risk of short circuits. Therefore, once the humidity exceeds 70%, the system will determine that the internal environment of the light box is abnormal and enter the dynamic adjustment mode.

[0114] In the case of excessive humidity, the system will input the current environmental feature dataset into the LSTM neural network model for analysis. This dataset includes humidity, temperature, light intensity, light change rate, light duration, sunlight angle, and wind speed, which are collected by various sensors inside the light box. For example, the humidity sensor measures the current humidity value, the temperature sensor detects the temperature changes inside and outside the light box, the light sensor monitors the light intensity around the light box, and the wind speed sensor measures the air flow outside. After these data are collected, they will be normalized, noise filtered, and feature extracted to ensure the accuracy and effectiveness of the data.

[0115] At the input layer of the LSTM model, the system first adjusts the time dimension of the feature data, converting it into a format suitable for time series analysis. Since LSTM neural networks are specifically designed for processing time series data, the input data needs to be arranged in chronological order and formed into data windows to allow the neural network to learn the patterns of historical data changes. For example, the system uses the humidity changes in the past 10 minutes as input, combined with the trends of temperature and wind speed, to predict future humidity fluctuations.

[0116] Subsequently, the data enters the LSTM layer, which is primarily responsible for identifying and analyzing time series data. Through the coordinated action of the forget gate, input gate, and output gate, the LSTM network can effectively extract key time series features while avoiding information loss. For example, in the case of abnormal humidity rise, the LSTM network can identify the speed, duration, and periodicity of the humidity rise, and combine factors such as temperature and wind speed to determine the cause of the humidity rise. If the humidity rises quickly and is accompanied by a drop in temperature, it is due to excessive external air humidity; if the humidity rises slowly and the temperature remains stable, it is due to water vapor accumulation inside the light box.

[0117] The data processed by the LSTM layer enters the fully connected layer, which maps the time series feature vector to the environmental feature embedding vector to facilitate subsequent calculation of the cover angle. The fully connected layer further integrates the feature information extracted by the LSTM layer through weighted operations of multiple neurons. For example, if the system detects that the humidity is rising and the wind speed is low, the model will tend to calculate a larger cover angle to enhance air circulation efficiency; if the humidity is rising but the wind speed is high, the cover angle will be relatively small to avoid excessive wind pressure affecting the stability of the light box structure.

[0118] Finally, the data enters the output layer of the LSTM model, which is responsible for calculating the initial cover angle range. This angle range is calculated based on the environmental feature embedding vector and is used to guide subsequent cover adjustment operations. The system compares the calculated cover angle range with the safety threshold to ensure that the cover operation does not affect the overall stability of the light box. For example, in the case of rising humidity and low wind speed, the system calculates a cover angle range of 5 to 15 degrees, while in the case of high wind speed, the range is adjusted to 3 to 8 degrees to reduce the impact of wind resistance.

[0119] Through the LSTM neural network model, the system can intelligently predict the humidity change trend based on historical data and real-time environmental parameters, and calculate a reasonable cover lifting angle range, thereby improving the dynamic adaptability of cover lifting control. Compared with the traditional fixed cover lifting strategy, this method can flexibly adjust the cover lifting angle under different environmental conditions, improve the air circulation efficiency, ensure the stability of the internal environment of the light box, and at the same time avoid unnecessary energy loss and safety hazards caused by excessive cover lifting.

[0120] In step S15, the air heat exchange power is calculated according to the environmental dynamic parameters, and the air heat exchange power is obtained, including:

[0121] The air heat exchange power is calculated by the following formula:

[0122]

[0123] wherein, is the air heat exchange power, is the specific heat capacity of air, is the wind speed, is the internal temperature of the light box, is the external temperature of the light box, is the air density.

[0124] It is worth noting that in step S15, the system calculates the air heat exchange power according to the environmental dynamic parameters to evaluate the heat exchange effect of the air circulation in the waterproof light box, thereby assisting the optimization decision of the cover lifting angle and time. The air heat exchange power is a key indicator to measure the heat removal effect of air circulation, and its calculation is based on the specific heat capacity of air, air density, wind speed and temperature difference between the inside and outside of the light box. The core purpose of this calculation is to determine the contribution of air flow to temperature regulation, to ensure that the cover lifting angle can adapt to the change of environmental temperature, and to improve the stability of the internal environment of the light box.

[0125] In the specific calculation process, the system first needs to obtain the specific heat capacity and air density of the air. The specific heat capacity of air is the heat absorbed or released by unit mass of air when the temperature changes by unit. The specific heat capacity of air depends on temperature and pressure. At standard atmospheric pressure (101.325 kPa) and 20 degrees Celsius, the specific heat capacity of air is about 1005 J / (kg·K). However, due to the differences between the actual application environment and the standard atmospheric pressure and temperature, the specific heat capacity needs to be corrected according to the environmental temperature and pressure. The system uses real-time weather data or environmental information measured by the built-in air pressure sensor and temperature sensor, and adjusts the air specific heat capacity dynamically by consulting the specific heat capacity temperature correction table. For example, when the environmental temperature rises to 40 degrees Celsius, the air specific heat capacity increases to 1010 J / (kg·K), and when the temperature drops to 0 degrees Celsius, the specific heat capacity decreases to 1000 J / (kg·K). Therefore, the system needs to first correct the value of the air specific heat capacity according to the real-time temperature when calculating the air heat exchange power, to ensure the accuracy of the calculation.

[0126] Air density is another important parameter that affects air heat exchange power. Air density refers to the mass of air per unit volume, which is also affected by temperature and pressure. At standard atmospheric pressure and 20 degrees Celsius, the air density is about 1.204 kg / m³. The system measures the real-time temperature through external environmental sensors, and combines the standard atmospheric pressure or the measured actual air pressure to correct the air density in the current environment using the ideal gas state equation. For example, if the measured environmental temperature is 30 degrees Celsius, the air density will decrease to 1.164 kg / m³, and at 0 degrees Celsius, the air density will increase to 1.293 kg / m³. To improve the calculation accuracy, the system can filter the air density through multiple sampling methods to reduce measurement errors and obtain more accurate air density values.

[0127] After obtaining the specific heat capacity and air density of the air, the system needs to measure the wind speed inside the light box through the wind speed sensor. The wind speed sensor uses an ultrasonic wind speed sensor or a hot-wire wind speed sensor, which is installed at the ventilation port position inside the light box to monitor the air flow in real time. The wind speed data collected by the wind speed sensor is filtered and combined with the temperature data to input into the calculation formula. The temperature inside the light box is obtained by the temperature sensor installed inside the light box, and the external temperature is measured by the temperature sensor arranged outside the light box. The temperature difference is used to evaluate the temperature driving force of air heat exchange.

[0128] During the calculation, the system first substitutes the obtained specific heat capacity of air, air density, wind speed, and temperature difference into the calculation formula to calculate the air heat exchange power value. The calculated air heat exchange power reflects the heat that can be taken away by air circulation per unit time, thereby indirectly determining the ability of temperature regulation inside the light box. If the air heat exchange power is high, it indicates that the current ventilation condition is good, and the temperature regulation effect is strong, and the cover lifting angle can be appropriately reduced to reduce the influence of external air. If the air heat exchange power is low, it indicates that the current air circulation is not smooth, and the cover lifting angle needs to be increased to enhance air flow and improve temperature regulation capability.

[0129] For example, in actual application, the internal temperature of the light box is 35 degrees Celsius, the external temperature is 25 degrees Celsius, the wind speed sensor measures a wind speed of 2 m / s, the air density is measured as 1.184 kg / m³, and the specific heat capacity is taken as 1005 J / (kg·K). The calculated air heat exchange power is 23.68 W. If the calculated air heat exchange power is lower than a certain set threshold, for example, 15 W, the system determines that the current heat exchange effect is insufficient, and the cover lifting angle needs to be adjusted to enhance air circulation. If the heat exchange power is higher than a certain upper limit, for example, 50 W, the cover lifting angle needs to be reduced to prevent excessive ventilation from causing excessive fluctuations in the internal temperature of the light box.

[0130] The accuracy of this calculation step is crucial to the cover lifting control system. By reasonably calculating the air heat exchange power, the system can dynamically adjust the cover lifting angle, so that the waterproof light box can maintain stable temperature under different environmental conditions, avoid imbalance of temperature and humidity inside the light box due to drastic changes in external environment, and thereby improve the reliability and service life of the waterproof light box.

[0131] In step S16, based on the initial cover lifting angle range and the air heat exchange power, an initial optimization population is generated, and the initial optimization population is iteratively optimized to obtain an optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover lifting angle for output.

[0132] Based on the initial cover lifting angle range, a plurality of cover lifting angle individuals are generated, and the air heat exchange power corresponding to the cover lifting angle individuals is calculated to obtain an initial optimization population;

[0133] According to the air heat exchange power of the initial optimization population, the initial optimization population is iteratively optimized to obtain an optimized population;

[0134] Based on a preset optimization condition, it is determined whether the optimized population meets the optimization termination requirement. If yes, the optimized population is taken as the optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover lifting angle for output;

[0135] If the optimized population does not meet the preset optimization condition, continue to perform genetic operations until the optimization termination condition is met.

[0136] The preset optimization condition is the convergence of fitness value or reaching the maximum number of iterations.

[0137] It is worth noting that in the intelligent control system of the waterproof light box, the optimization of the cover angle is an important link to ensure the stability of the internal environment of the light box and maximize the efficiency of air circulation. Genetic algorithm is used to optimize the cover angle, so that it always maintains the best state under different environmental conditions. Genetic algorithm simulates the process of biological evolution, continuously optimizes the cover angle through selection, crossover and mutation operations, and finally obtains the optimal solution. The core of this step is to establish an optimized population and find the optimal cover angle through iterative optimization, while setting reasonable termination conditions to ensure the efficiency and accuracy of the optimization process.

[0138] First, in the initialization stage of optimization, an initial optimization population needs to be generated. This population consists of multiple cover angle individuals, each representing a possible cover angle scheme. The cover angle values are based on the initial cover angle range provided by the dynamic adjustment model, which is calculated by the LSTM neural network model and ranges from 0 degrees to 90 degrees. To ensure the diversity of the initial population, a random generation method can be used to randomly select 50 to 100 cover angle individuals within this range. For example, in a certain initialization, the randomly generated population may include individuals with cover angles of 20 degrees, 35 degrees, 50 degrees, etc. These individuals will evolve in the subsequent optimization process to find the optimal cover angle.

[0139] Secondly, the fitness value of each cover angle individual is calculated. The goal of the fitness function is to measure whether the air heat exchange power of the cover angle individual under the current environmental conditions meets the optimization target. The calculation of air heat exchange power needs to combine multiple environmental dynamic parameters, including wind speed, temperature difference, air density and specific heat capacity, etc. Wind speed data is collected by a wind speed sensor, temperature data is provided by internal and external temperature sensors of the light box, and air density and specific heat capacity are corrected according to meteorological data or standard values determined by experiments. For example, in a certain environment, the wind speed is measured to be 5 m / s, the internal temperature of the light box is 30℃, and the external temperature is 22℃. The calculated air heat exchange power of a certain cover angle individual is 3.8 W, and the fitness value of this individual is 3.8. In the entire population, the fitness value of each cover angle individual will be calculated to evaluate its performance under the current environment.

[0140] In the optimization process, the core operations of the genetic algorithm include selection, crossover, and mutation. First, the selection operation is used to select individuals with higher fitness from the current population into the next generation, and excellent individuals are retained, while individuals with lower fitness are eliminated. For example, if the fitness value of a certain flip angle individual is 4.0, and the fitness value of another individual is only 2.5, the former has a higher probability of being selected into the next generation. The selection process uses roulette selection or tournament selection to ensure that individuals with higher fitness have a greater chance of passing on their genes. Then, in the crossover operation stage, the selected parent individuals undergo gene recombination to generate new offspring individuals. The crossover method can use single-point crossover or double-point crossover. For example, the flip angles of two parent individuals are 30 degrees and 50 degrees, respectively, and after crossover, the offspring individuals may inherit some of the parent characteristics, generating new flip angle individuals such as 40 degrees or 45 degrees. Finally, in the mutation operation, the flip angle of the offspring individual is adjusted with a certain probability to increase the diversity of the population. For example, the flip angle of a certain offspring individual is 38 degrees, and after mutation it may become 36 degrees or 40 degrees, thus exploring new search space and avoiding the optimization process from falling into a local optimal solution.

[0141] The setting of optimization termination conditions is the key to ensuring computational efficiency and optimization effect. Optimization termination conditions include fitness value convergence, reaching the maximum number of iterations, or reaching the target air heat exchange power value. When the fitness value of the population changes by less than a set threshold, for example, less than 0.01, in consecutive generations, it is considered that the optimization has converged, and the iteration can be stopped. For example, if the air heat exchange power of the optimal individual in the population changes by no more than 0.01 in consecutive 10 generations, it is considered that the optimization has reached a stable state. In addition, in order to prevent the calculation process from being too long and causing resource waste, the maximum number of iterations can be set, for example, 100 generations, if the convergence condition is still not reached within this range, the optimization is terminated to avoid excessive calculation. The third way to terminate optimization is to set a target air heat exchange power value, for example, 4.5 W, when the fitness value of a certain individual reaches or exceeds this value, it is considered that the optimization goal is achieved, and the optimization process is terminated in advance.

[0142] When the optimization terminates, the individual with the highest fitness value from the final optimization population is selected as the final optimal cover angle and is used for cover control. For example, in a certain optimization process, the final selected optimal cover angle is 42 degrees, and the corresponding air heat exchange power is 4.6 W, then the system takes this cover angle as the final control parameter. Subsequently, the system sends this angle value to the cover actuator, and adjusts the cover to 42 degrees through the servo motor to optimize the air flow effect. The system will continuously monitor environmental parameters such as wind speed, temperature changes, etc. in subsequent periods, and restart the optimization process if necessary to ensure that the dynamic adjustment of the cover angle always adapts to the current environment. For example, if the wind speed rises from 5 m / s to 8 m / s during subsequent monitoring, the air heat exchange power may change, at which time the system will re-optimize to adjust the cover angle to maintain optimal air flow efficiency.

[0143] The entire optimization process combines real-time environmental data acquisition, air heat exchange power calculation, and efficient genetic algorithm optimization. In practical applications, to improve optimization efficiency, parallel computing techniques can be combined, such as using multi-threaded computing to speed up fitness calculation and population iteration. At the same time, during the optimization process, an adaptive mutation strategy can be introduced to further improve search capability. For example, when the optimization process approaches convergence, the mutation probability can be reduced to reduce the randomness of the search; when the optimization falls into local optimum, the mutation probability can be increased to expand the search range and improve the possibility of finding the global optimal solution.

[0144] In step S17, the cover time is calculated according to the environmental dynamic parameters to obtain the initial cover time, including:

[0145] The initial cover time is calculated by the following formula:

[0146]

[0147] wherein, is the initial cover time, is the internal temperature of the light box, is the external temperature of the light box, is the light intensity, is the light duration, is the light change rate, is the sun angle, is the normalization coefficient, is the time adjustment factor, is the angle influence coefficient.

[0148] It's worth noting that calculating the opening time is a crucial step in the intelligent control of the waterproof lightbox's lid. It determines the length of time the lid needs to be open under different environmental conditions to ensure stable internal temperature and humidity and optimize airflow efficiency. To accurately calculate the opening time, multiple environmental dynamic parameters need to be considered, including the lightbox's internal and external temperatures, light intensity, light duration, light variability rate, and solar angle. Based on these parameters, the calculation formula is corrected using normalization coefficients, time adjustment factors, and angle influence coefficients to ensure the calculation's rationality and dimensional consistency.

[0149] First, environmental data is acquired. Temperature data for the inside and outside of the waterproof lightbox is measured by temperature sensors, and the temperature data is expressed in degrees Celsius. Illumination intensity is collected by a light sensor, and its unit is lux. Illumination duration is calculated by the light sensor combined with the system timestamp, and its unit is seconds. The rate of change of illumination is the rate at which illumination intensity changes over time, and its unit is lux per second. The solar angle can be acquired by a light angle sensor installed near the lightbox, or it can be calculated based on sunrise and sunset times combined with geographical location. All this data, after acquisition, will be normalized and adjusted by the data processing module.

[0150] During the data processing phase, it is necessary to match the dimensions of each parameter to ensure the correctness of the calculation formula. The unit for temperature difference is Kelvin, the unit for light intensity is lux, and the unit for light duration is seconds. Direct calculations will result in a mismatch between these units, therefore a normalization coefficient needs to be introduced. This allows the unit of measurement for the temperature difference term to be converted to seconds. Normalization coefficient. The value of needs to be selected based on experimental data. The value ranges from 0.01 to 0.05, with units of seconds per Kelvin multiplied by lux per second. This value ensures that the contribution of temperature difference to the opening time is consistent with reality, avoiding the overestimation or neglect of the temperature effect. For example, in an experimental environment, the internal temperature of the light box is 35 degrees Celsius, the external temperature is 25 degrees Celsius, the light intensity is 10000 lux, and the light duration is 3600 seconds. If no temperature difference is introduced... If the calculated time is not accurate, it may deviate from the actual requirement. Therefore, it is necessary to select an appropriate time. After setting the value, the opening time can be reasonably matched with the actual environment.

[0151] Next, we calculate the effect of the rate of change of illumination on the opening time. The rate of change of illumination represents the change in illumination intensity per unit time, and its unit is lux per second. When this term is directly added to the calculation, the units are mismatched, so it needs to be multiplied by a time adjustment factor. Time adjustment factor Its function is to convert the rate of change of illumination into a time unit of seconds, giving it practical physical meaning in time calculations. The value ranges from 100 to 500 seconds, depending on the drastic change in light intensity under experimental conditions. For example, in a clear, cloudless environment, the rate of change in light intensity is relatively small, making a higher value suitable. For example, a value of 400 seconds is suitable, but in cloudy weather, the rate of change in sunlight is greater, so a smaller value is appropriate. A value like 150 seconds. Reasonable. This value ensures that the effect of changes in light intensity on the opening time is consistent with reality, avoiding over-response or lag.

[0152] In addition, the influence of the solar angle also needs to be taken into account in the calculation. The solar angle represents the angle at which sunlight shines on the lightbox, affecting the degree of heating inside the lightbox. When the solar angle is small, the direct impact of the sunlight is small, while when the solar angle is large, the sunlight will be stronger, thus affecting the setting of the opening time. Since the cosine value of the solar angle is dimensionless, it needs to be multiplied by an angle influence coefficient during the calculation. This ensures that it can be correctly applied to time calculations. Angle influence coefficient. The duration ranges from 5 to 30 seconds, depending on the lightbox's installation location, seasonal variations, and local climate conditions. For example, in equatorial regions, due to the larger angle of sunlight... A higher value, such as 25 seconds, is suitable. However, in high-latitude regions, due to the smaller angle of sunlight, A lower value, such as 10 seconds, is suitable. (Appropriate selection) The value ensures that the influence of the solar angle is reasonably reflected in the calculation of the cover-opening time, avoiding excessive deviations in cover-opening time in different seasons and latitude regions.

[0153] After normalization adjustments, the calculated results are summed to obtain the final lid-opening time. This time controls the lid-opening action of the lightbox, ensuring that under specific environmental conditions, the opening time meets air circulation requirements without causing an imbalance in the internal environment. For example, in a certain experimental test environment, the internal temperature of the lightbox is 40 degrees Celsius, the external temperature is 28 degrees Celsius, the illuminance is 15000 lux, the illumination duration is 5400 seconds, the illuminance variation rate is 50 lux per second, and the sunlight angle is 60 degrees. The calculated opening time is 120 seconds. This calculation ensures that the lightbox can be opened for ventilation in a timely manner under these conditions without excessively affecting the external environment due to an excessively long opening time.

[0154] In summary, this method, by introducing normalization coefficients, time adjustment factors, and angle influence coefficients, ensures dimensional matching of various physical quantities, making the calculated lid-opening time conform to practical application requirements. The function of the temperature difference is to normalize the temperature difference to seconds, The function of the light change rate is to ensure that the time influence of the light change rate is reasonable, The function of the sunshine angle is to quantify the influence of the sunshine angle into the cover lifting time calculation.

[0155] In step S18, the initial cover lifting time and cover lifting angle are iteratively optimized based on the PSO algorithm to obtain cover control instructions, including:

[0156] The initial cover lifting time and the cover lifting angle are initialized to obtain an initial state particle group;

[0157] The fitness of each particle is calculated based on the initial state particle group to obtain the fitness value of each particle;

[0158] The optimal fitness value is calculated based on the fitness value of each particle to obtain the optimal fitness value;

[0159] If the optimal fitness value meets the preset optimization termination condition, the particle with the highest fitness value is selected as the final optimization result, and the cover lifting time and cover lifting angle of the particle are selected as the optimal cover control instructions for output;

[0160] If the optimal fitness value does not meet the preset optimization termination condition, the cover lifting time and cover lifting angle of the initial state particle group are updated and iteratively optimized until the preset optimization termination condition is met;

[0161] The cover control instructions include cover lifting time and cover lifting angle.

[0162] It is worth noting that in the intelligent control process of the waterproof light box, the accurate calculation of the cover lifting time and the cover lifting angle directly affects the balance between the air circulation efficiency and the waterproof performance. In order to optimize the control parameters of the cover lifting, the particle swarm optimization algorithm is used to iteratively optimize the initial cover lifting time and the cover lifting angle, so as to obtain the optimal cover control instructions. The goal of this step is to use intelligent optimization method to make the combination of cover lifting time and angle optimal for air circulation efficiency, while ensuring the stability of the internal environment of the light box.

[0163] Firstly, the first step of the optimization process is to initialize the particle swarm. In the particle swarm optimization algorithm, each particle represents a possible solution, and in this scheme, each particle is composed of a flap time value and a flap angle value. The initial values of the particles are derived from the initial flap time and initial flap angle obtained from the pre-computation, and reasonable search ranges are set according to empirical values. For example, the flap angle may vary between 0 and 90 degrees, while the flap time may fluctuate between 0 and 60 seconds. Assuming that the number of initialized particle swarms is 50, the system will randomly generate 50 different combinations of flap time and flap angle within the range to ensure the diversity of the optimization process.

[0164] Secondly, after the initialization of the particle swarm is completed, the fitness value of each particle needs to be calculated. The fitness value is a standard for measuring the pros and cons of the current particle solution, and in this optimization process, the fitness function is mainly calculated based on the air flow efficiency calculation formula. The design of the fitness function needs to consider the temperature difference change caused by air flow after flapping, wind speed influence, and light intensity and other parameters. In order to ensure the accuracy of the calculation, the system will call real-time sensor data, such as wind speed sensor, temperature and humidity sensor and light sensor, to obtain the current environmental parameters. In a specific example, if the flap angle of a particle is 45 degrees and the flap time is 30 seconds, the calculated fitness value is 4.2, while the flap angle of another particle is 50 degrees and the flap time is 25 seconds, the calculated fitness value is 4.5, then the fitness value of the latter is higher, which means its optimization effect is better.

[0165] Next, in each round of iteration, the system will update the optimization according to the fitness value of the particle. The optimization process includes particle velocity update and position update, where velocity represents the trend of particle solution change, and position represents the current solution value of the particle. The update method is mainly affected by the particle's own historical optimal solution and the group's optimal solution. For example, if a particle performs well in the past few rounds of iteration, the algorithm will tend to let it continue to optimize in that direction, while if a solution in the group performs extremely well, other particles will also converge to that solution. In a specific application, if the global optimal solution of a certain round of iteration corresponds to a flap angle of 40 degrees and a flap time of 28 seconds, most individuals in the next round of particle swarm may converge to that solution, thereby accelerating the optimization convergence speed.

[0166] In the iterative optimization process, optimization termination conditions need to be set to ensure computational efficiency and prevent the algorithm from falling into meaningless long-time calculation. The optimization termination conditions include two cases. One is the convergence of fitness value, that is, when the optimal fitness value of the particle swarm as a whole changes by less than a set threshold after multiple iterations, the system considers that the optimal solution has been found, for example, when the optimal fitness value changes by less than 0.01 in 10 consecutive iterations, the optimization is stopped. The second is to reach the maximum number of iterations, set to 100 to 200 rounds, to prevent the algorithm from converging slowly for a long time. If the fitness value has not changed significantly after 150 iterations, the system will forcibly stop the optimization and select the current optimal solution as the final cover control command.

[0167] After optimization termination, the system selects the particle with the highest fitness value from the final optimized particle swarm and outputs its corresponding cover time and cover angle as the final cover control command. For example, in a certain optimization process, the final optimal solution is a cover angle of 42 degrees and a cover time of 27 seconds, and the system will send this parameter to the execution control module to make the waterproof light box perform the cover operation according to this parameter. This process ensures the optimal matching of cover time and angle, thereby improving air circulation efficiency and effectively avoiding the problems of excessive cover or insufficient cover.

[0168] If the convergence speed is too slow or the diversity of the solution is insufficient during the optimization process, the system can adjust the inertia weight, acceleration factor or particle swarm size to improve the optimization efficiency. For example, a larger inertia weight can be set in the early iteration process to increase the search range and prevent the algorithm from falling into local optimum, and the inertia weight can be gradually reduced in the later period to make the particle swarm more concentrated around the optimal solution. In addition, dynamic adjustment strategies can be introduced, such as increasing the mutation probability when the fitness distribution of individuals in the population is too concentrated, making the solution distribution more diverse, thereby enhancing the global search ability of the algorithm.

[0169] Finally, the optimization process uses the particle swarm optimization algorithm to continuously adjust the cover time and cover angle, enabling the waterproof light box to achieve the optimal cover strategy under different environmental conditions. The optimized cover control command not only improves the air circulation efficiency, but also ensures the balance of temperature and humidity inside the light box, reduces energy consumption, and improves the overall intelligent level of the system.

[0170] The following describes the working process of the present application in a more common scenario. Please refer to Figure 2 , which is Figure 1 the working scenario diagram of the method.

[0171] In a common outdoor advertising light box application scenario, such as a waterproof light box installed on an outdoor billboard, it is exposed to the external environment for a long time and may be affected by factors such as high humidity, strong light, or strong wind. The goal of the system is to intelligently adjust the cover angle of the light box to optimize air circulation, reduce internal humidity, improve protection performance, and ensure the stable operation of the electronic equipment inside the light box.

[0172] In step S11, the system first collects humidity data and environmental dynamic parameters such as light data, temperature data, and wind speed through the data acquisition module. These data come from multiple sensors installed inside and outside the light box, including humidity sensors, light sensors, temperature sensors, and wind speed sensors. For example, when the light box is in the rainy season in a coastal city in the south, the air humidity is maintained above 90% for a long time, and the internal temperature rises to 40 degrees Celsius due to long-time lighting. Changes in these environmental parameters will directly affect the air flow state inside the light box.

[0173] In step S12, the system preprocesses the humidity and environmental dynamic parameters based on the data fusion algorithm to obtain an accurate environmental feature dataset. First, normalize the various types of data to eliminate the influence between different data units. Then, use the Kalman filter method to filter the data noise and remove abnormal data caused by sensor measurement errors or environmental mutations. Finally, use principal component analysis to extract features from the denoised data to obtain the environmental feature dataset. For example, in the case of an outdoor advertising light box, the light intensity is above 5000 lux during the day and drops to 0 lux at night. Using the data fusion algorithm can effectively identify abnormal changes and ensure accurate judgment of the system on the environmental state.

[0174] In step S13, the system first determines whether the humidity exceeds the set normal humidity range. The set humidity normal range is generally between 30% and 70% based on experimental data and actual application requirements. When the humidity inside the light box is below 70%, the system determines that the light box is in a normal state and does not need to adjust the cover. For example, in dry autumn or sunny weather, the humidity level will not exceed the standard, and the system will maintain the original state without performing the cover operation to ensure the protection performance of the light box.

[0175] In step S14, when the humidity exceeds the normal range, for example, more than 80%, the system determines that the humidity inside the light box is abnormal, and inputs the environmental feature dataset into the pre-constructed LSTM neural network dynamic adjustment model to calculate the initial cover angle range suitable for the current environment. The LSTM model has been trained with a large amount of historical data and can predict the optimal cover angle range according to different humidity, wind speed, and temperature trends. For example, in the case of humidity exceeding 85%, the LSTM model outputs a preliminary adjustment suggestion of a cover angle range of 15 to 30 degrees.

[0176] In step S15, the system calculates the air heat exchange power based on the environmental dynamic parameters. The air heat exchange power is used to evaluate the efficiency of air circulation in the lamp box after the cover is lifted. Parameters such as air specific heat capacity and air density are used in the calculation. For example, when the wind speed is high (e.g., 10 m / s), the air heat exchange power will be relatively large, while in the windless state, the heat exchange effect is low. Therefore, in winter when it is dry but the wind speed is low, the system needs to lift the cover at a larger angle to increase air flow, while in summer when the humidity is high but the wind speed is large, appropriate lifting of the cover can achieve good heat exchange effect.

[0177] In step S16, the system optimizes the initial cover lifting angle range and air heat exchange power based on the genetic algorithm to obtain the optimal cover lifting angle. The system first generates an initial optimization population, each individual corresponding to a cover lifting angle, and calculates its air heat exchange power to form an initial optimization population. Then, in the iterative process of the genetic algorithm, the system performs selection, crossover and mutation operations on the cover lifting angle individuals, so that individuals with higher fitness are more likely to be inherited to the next generation, until the optimization termination condition is met, such as reaching the maximum number of iterations 100 times or the fitness value changes less than 0.01. After multiple optimizations, the cover lifting angle with the highest fitness is finally output as the optimal solution, for example, 30 degrees.

[0178] In step S17, the system calculates the cover lifting time based on the environmental dynamic parameters to obtain the initial cover lifting time. The calculation of the cover lifting time takes into account factors such as temperature difference, light change rate, light duration and solar angle, to ensure that the cover lifting operation can improve air circulation efficiency without causing additional energy consumption. For example, in summer with strong direct sunlight and long light duration, the cover lifting time is set to 15 minutes, while in the low-temperature environment at night, the cover lifting time only needs 5 minutes to reduce heat loss.

[0179] In step S18, the system iteratively optimizes the initial cover lifting time and cover lifting angle based on the particle swarm optimization algorithm to obtain the final cover lifting control instruction. The particle swarm optimization algorithm continuously adjusts the cover lifting angle and time in the iteration process to maximize the air circulation efficiency. For example, during the optimization process, the system finds that adjusting the cover lifting angle to 25 degrees and setting the cover lifting time to 12 minutes results in the highest air circulation efficiency, so the system finally selects these parameters as the control instruction.

[0180] The present application realizes dynamic adjustment of the waterproof lamp box humidity abnormal condition through the intelligent cover lifting control method, has higher environmental adaptability and control precision compared with the traditional fixed cover lifting method. Through time series analysis of environmental data by the LSTM neural network model, the cover lifting angle adjustment can be optimized based on historical trends, improving the rationality of decision-making. Combined with air heat power calculation and genetic algorithm optimization, the selection of cover lifting angle is more scientific, ensuring the maximum improvement of air flow effect, reducing the humidity inside the lamp box, and preventing electronic components from being damaged by moisture. In addition, based on the PSO algorithm, the cover lifting time and angle are further optimized, making the cover lifting process more stable and efficient, reducing unnecessary structural wear and tear, and improving the service life of the equipment. The present method realizes dynamic and accurate cover lifting control.

[0181] Reference Figure 2 The second embodiment of the present application provides a cover lifting intelligent control system for a waterproof lamp box, comprising:

[0182] A data acquisition module is configured to acquire humidity and environmental dynamic parameters.

[0183] A preprocessing module is configured to preprocess the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature data set.

[0184] An abnormality determination module is configured to determine that the waterproof lamp box is normal when the humidity does not exceed the set normal humidity range.

[0185] A dynamic adjustment module is configured to determine that the waterproof lamp box is abnormal when the humidity exceeds the set normal humidity range, and input the environmental feature data set into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range.

[0186] An air heat power module is configured to calculate air heat power based on the environmental dynamic parameters to obtain air heat power.

[0187] A genetic algorithm module is configured to generate an initial optimization population based on a genetic algorithm for the initial cover lifting angle range and the air heat power, and iteratively optimize the initial optimization population to obtain an optimal population, and select an individual with the highest fitness value in the optimal population as the cover lifting angle for output.

[0188] A cover lifting time module is configured to calculate the cover lifting time based on the environmental dynamic parameters to obtain an initial cover lifting time.

[0189] A cover lifting control instruction module is configured to iteratively optimize the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain a cover lifting control instruction.

[0190] It should be noted that the cover lifting intelligent control system of the waterproof lamp box provided by the embodiment of the present application is used to execute all process steps of the cover lifting intelligent control method of the waterproof lamp box of the above-mentioned embodiment, and the working principles and beneficial effects of the two are one-to-one correspondence, thus it is not repeated here.

[0191] The embodiment of the present application also provides an electronic device. The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a cover lifting intelligent control program of a waterproof lamp box. The processor implements the steps in the above-mentioned various cover lifting intelligent control method embodiments of the waterproof lamp box when executing the computer program, such as the step S11 shown in the figure. Figure 1 The processor implements the functions of the modules / units in the above-mentioned various device embodiments when executing the computer program, such as the cover lifting control instruction module.

[0192] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0193] The electronic device can be a desktop computer, a notebook computer, a palm computer, a smart tablet and the like. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above-mentioned components are only examples of the electronic device, and do not constitute a limitation on the electronic device, and can include more or less components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0194] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the electronic device, and connects all parts of the electronic device through various interfaces and lines.

[0195] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), and the like. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0196] The modules / units integrated in the electronic device can be stored in a computer readable storage medium if they are realized in the form of software function units and sold or used as independent products. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0197] It should be noted that the apparatus embodiments described above are merely illustrative, and the units described as separate units can or can not be physically separate, and the units displayed as units can or can not be physical units, i.e. can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the connection relationship between the modules in the apparatus embodiment provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0198] The above specific embodiments further illustrate the purpose, technical scheme and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A cover lifting intelligent control method of a waterproof lamp box, characterized in that, The method comprises the following steps: acquiring humidity and environmental dynamic parameters; preprocessing the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature dataset; when the humidity does not exceed a preset normal humidity range, determining that the waterproof light box is normal; when the humidity exceeds the preset normal humidity range, determining that the waterproof light box is abnormal, and inputting the environmental feature dataset into a pre-constructed dynamic adjustment model to obtain an initial cover lifting angle range; calculating air heat exchange power based on the environmental dynamic parameters to obtain air heat exchange power; generating an initial optimization population based on a genetic algorithm for the initial cover lifting angle range and the air heat exchange power, and iteratively optimizing the initial optimization population to obtain an optimal population, and selecting an individual with the highest fitness value in the optimal population as the cover lifting angle for output; calculating cover lifting time based on the environmental dynamic parameters to obtain an initial cover lifting time; iteratively optimizing the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain cover lifting control instructions; wherein the calculation of the initial cover lifting time based on the environmental dynamic parameters comprises: the initial cover lifting time is calculated by the following formula: wherein, is the initial lid-up time, is the inside temperature of the light box, is the outside temperature of the light box, is the light intensity, is the light duration, is the light change rate, is the sun angle, is the normalization coefficient, is the time adjustment factor, is the angle influence coefficient; wherein the iteratively optimizing the initial cover lifting time and the cover lifting angle based on the PSO algorithm to obtain the cover lifting control instructions comprises: initializing the initial cover lifting time and the cover lifting angle to obtain an initial state particle swarm; calculating the fitness value of each particle based on the fitness value of each particle to obtain an optimal fitness value; if the optimal fitness value meets a preset optimization termination condition, selecting a particle with the highest current fitness as the final optimization result, and selecting the cover lifting time and the cover lifting angle of the particle as the cover lifting control instructions for output; if the optimal fitness value does not meet the preset optimization termination condition, updating and iteratively optimizing the cover lifting time and the cover lifting angle of the initial state particle swarm until the preset optimization termination condition is met; the cover lifting control instructions comprise cover lifting time and cover lifting angle. The environmental dynamic parameters comprise illumination data, temperature data and wind speed; 2.The lid opening and closing intelligent control method of a waterproof lamp box according to claim 1, characterized in that, the illumination data comprises illumination intensity, illumination change rate, illumination duration and sunshine angle; the temperature data comprises light box internal temperature and light box external temperature. The preprocessing of the humidity and the environmental dynamic parameters based on the data fusion algorithm to obtain the environmental feature dataset comprises: 3.The lid intelligent control method of a waterproof lamp box according to claim 1, characterized in that, normalizing the humidity and the environmental dynamic parameters to obtain a normalized multi-source environmental dataset; performing noise filtering on the multi-source environmental dataset based on a Kalman filtering method to obtain a denoised multi-source environmental dataset; performing feature extraction on the denoised multi-source environmental dataset based on a principal component analysis method to obtain the environmental feature dataset; ​ The environmental feature dataset includes humidity features, illumination intensity, illumination rate of change features, illumination duration features, sunshine angle features, light box internal temperature features, light box external temperature features, and wind speed features. 4.The lid opening and closing intelligent control method of a waterproof lamp box according to claim 1, wherein, When the humidity exceeds the set normal humidity range, it is determined that the waterproof light box is abnormal, and the environmental feature dataset is input into a pre-constructed dynamic adjustment model to obtain an initial cover angle range, including: The dynamic adjustment model is trained by an LSTM neural network model; The time dimension of the feature dataset is adjusted through the input layer of the dynamic adjustment model to obtain pre-processed time series data; The pre-processed time series data is subjected to time series recognition through the LSTM layer of the dynamic adjustment model to obtain a time series feature vector; The time series feature vector is subjected to mapping conversion through the fully connected layer of the dynamic adjustment model to obtain an environmental feature embedding vector; The environmental feature embedding vector is calculated through the output layer of the dynamic adjustment model to obtain the initial cover angle range. 5.The lid intelligent control method of a waterproof lamp box according to claim 1, wherein, The air heat exchange power is calculated according to the environmental dynamic parameters to obtain the air heat exchange power, including: The air heat exchange power is calculated by the following formula: wherein, P a for air heat exchange power, C p for air specific heat capacity, V for wind speed, T i for the inside temperature of the light box, T o for the outside temperature of the light box, p for air density. 6.The lid lifting intelligent control method of a waterproof lamp box according to claim 1, characterized in that, The initial optimization population is generated based on the genetic algorithm and the initial cover angle range and the air heat exchange power, and the initial optimization population is iteratively optimized to obtain an optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover angle and output, including: Based on the initial cover angle range, a plurality of cover angle individuals are generated, and the air heat exchange power corresponding to the cover angle individuals is calculated to obtain an initial optimization population; The initial optimization population is iteratively optimized according to the air heat exchange power of the initial optimization population to obtain an optimized population; It is judged whether the optimized population meets the optimization termination requirement based on a preset optimization condition, if it meets, the optimized population is taken as the optimal population, and the individual with the highest fitness value in the optimal population is selected as the cover angle and output; If the optimized population does not meet the preset optimization condition, the genetic operation is continuously performed until the optimization termination condition is met; The preset optimization condition is fitness value convergence or maximum iteration number.

7. A lid opening intelligent control system of a waterproof light box, used for implementing the lid opening intelligent control method of the waterproof light box according to any one of claims 1 to 6, characterized in that, It includes: The data acquisition module is configured to acquire humidity and environmental dynamic parameters; The preprocessing module is configured to preprocess the humidity and the environmental dynamic parameters based on a data fusion algorithm to obtain an environmental feature dataset; The abnormality determination module is configured to determine that the waterproof light box is normal when the humidity does not exceed the set normal humidity range; The dynamic adjustment module is configured to determine that the waterproof light box is abnormal when the humidity exceeds the set normal humidity range, and input the environmental feature dataset into a pre-constructed dynamic adjustment model to obtain an initial cover angle range; The air heat exchange power module is configured to calculate the air heat exchange power according to the environmental dynamic parameters to obtain the air heat exchange power; The air heat exchange power module is configured to calculate the air heat exchange power according to the environmental dynamic parameters to obtain the air heat exchange power; The genetic algorithm module is configured to generate an initial optimization population based on a genetic algorithm and the initial cover lifting angle range and the air heat exchange power, iteratively optimize the initial optimization population, obtain an optimal population, and select an individual with the highest fitness value in the optimal population as the cover lifting angle and output the cover lifting angle. The cover lifting time module is configured to calculate a cover lifting time based on the environmental dynamic parameters to obtain an initial cover lifting time. The cover lifting control instruction module is configured to iteratively optimize the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain a cover lifting control instruction.

8. An electronic device, comprising: The waterproof light box comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor executes the computer program to implement the cover lifting intelligent control method of the waterproof light box. The genetic algorithm module is configured to generate an initial optimization population based on a genetic algorithm and the initial cover lifting angle range and the air heat exchange power, iteratively optimize the initial optimization population, obtain an optimal population, and select an individual with the highest fitness value in the optimal population as the cover lifting angle and output the cover lifting angle. The cover lifting time module is configured to calculate a cover lifting time based on the environmental dynamic parameters to obtain an initial cover lifting time. The cover lifting control instruction module is configured to iteratively optimize the initial cover lifting time and the cover lifting angle based on a PSO algorithm to obtain a cover lifting control instruction. The waterproof light box comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor executes the computer program to implement the cover lifting intelligent control method of the waterproof light box.

Images