Air environment monitoring station humidity monitoring method, equipment and air environment monitoring station

Abstract

The present application relates to a humidity monitoring method, equipment and air environment monitoring station for an air environment monitoring station, and belongs to the field of environmental monitoring technology. The air environment monitoring station of the present application is provided with a plurality of humidity sensing devices for monitoring the environmental humidity. By acquiring the environmental humidity data collected by the humidity sensing device in real time, when there is an abnormal humidity situation, the source direction of the humidity interference in the environment is determined, and real-time image data of the corresponding direction is obtained; the real-time image data is matched with the interference model in the environmental interference model library, and after matching the interference model, it is determined that there is actual interference behavior; and corresponding response data is generated according to the type of actual interference behavior. The present application combines humidity monitoring with image analysis technology, and adopts advanced data processing and artificial intelligence algorithms to automatically process and analyze environmental humidity problems, which can realize fast and accurate diagnosis and location of problems.

Description

Technical Field

[0001] The present application belongs to the technical field of environmental monitoring, and specifically relates to a humidity monitoring method and equipment for an air environment monitoring station, and an air environment monitoring station. Background Art

[0002] Air pollution is a global public health issue. Pollutants such as particulate matter (PM2.5 and PM10), sulfur dioxide, and carbon monoxide have extremely serious impacts on human health. Air monitoring stations can provide key data, providing real-time data for urban planners and managers to help assess air quality and make scientific decisions.

[0003] At present, it is difficult to find problems in air environment monitoring stations at the first time. In terms of problem diagnosis, it usually relies on manual analysis of monitoring data, which is time-consuming and labor-intensive, and cannot detect and solve problems in a timely manner. The lack of intelligent diagnosis and analysis makes it impossible to accurately diagnose environmental problems, which is prone to misjudgment.

[0004] Therefore, how to overcome the defects of the existing air environment monitoring stations, such as difficult diagnosis, low efficiency and low accuracy, is a topic worthy of study for those skilled in the art. Summary of the invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a humidity monitoring method, device and air environment monitoring station for an air environment monitoring station, so as to solve the problems of difficult diagnosis, low efficiency and low accuracy of existing air environment monitoring stations.

[0006] In a first aspect of the present application, a humidity monitoring method for an air environment monitoring station is provided, wherein the air environment monitoring station is provided with a plurality of humidity sensing devices distributed therein for monitoring the ambient humidity, and the method comprises:

[0007] Acquire the environmental humidity data collected by the humidity sensing device in real time, determine whether there is an abnormal humidity situation based on the environmental humidity data, and when there is an abnormal humidity situation, determine the source direction of the humidity interference in the environment;

[0008] According to the source direction of the humidity interference, the image acquisition device of the air environment monitoring station is called to obtain real-time image data in the corresponding direction;

[0009] Matching the real-time image data with an interference model in an environmental interference model library, and determining the presence of actual interference behavior after matching the interference model; wherein the environmental interference model library has pre-established a plurality of different interference models for identifying actual interference behavior;

[0010] After determining that there is actual interference behavior, corresponding response data is generated according to the type of the actual interference behavior, and the response data includes the interference behavior type, the interference behavior time, the impact of the interference behavior on the humidity data, and recommended response measures.

[0011] Optionally, the process of establishing the interference model includes:

[0012] Collect image data of various potential disturbance behaviors including artificial spraying and fog cannons;

[0013] Key visual features are extracted from the image data, and the extracted key visual features are used as feature vectors as reference data for the recognition model; a convolutional neural network is used to train a classifier to identify different types of interference behaviors, and label data of known types of interference behaviors are used for training during the training process.

[0014] Optionally, matching the real-time image data with an interference model in an environmental interference model library, and determining the existence of actual interference behavior after matching the interference model includes:

[0015] Extract key features from the real-time image data to obtain a real-time image feature vector ;

[0016] Using the cosine similarity formula To calculate the real-time image feature vector and the interference model feature vector The similarity between

[0017] By comparing the similarity between the real-time image data and each interference model, it is determined which interference model the scene of the real-time image data matches, and after matching the interference model, it is determined that there is actual interference behavior.

[0018] Optionally, matching the real-time image data with an interference model in an environmental interference model library, and determining the existence of actual interference behavior after matching the interference model includes:

[0019] If the similarity between the water flow characteristics in the real-time image data and the water flow characteristics in the artificial spray interference model is greater than a first threshold, and the humidity change rate of the environmental humidity data at the corresponding position exceeds a preset range, it is determined that there is actual interference behavior of artificial spraying; wherein the water flow characteristics include water flow shape, water flow size, water flow spray direction, water flow continuity, and the interaction mode between the water flow and the surrounding environment;

[0020] If the feature similarity between the vehicle outline and spray pattern in the real-time image data and the fog cannon vehicle interference model is greater than a second threshold, and the humidity change rate of the ambient humidity data at the corresponding position exceeds a preset range, it is determined that there is actual interference behavior from the fog cannon vehicle.

[0021] Optionally, judging whether there is an abnormal humidity situation according to the environmental humidity data includes:

[0022] The environmental humidity data series With time series Together they form data points for trend analysis;

[0023] Use linear regression model to fit the data and calculate the humidity change rate ;in, and They are the means of time and ambient humidity data respectively;

[0024] Collect historical environmental humidity data and count different time windows The number and range of fluctuations of ambient humidity data within ;

[0025] Determining whether the environmental humidity data exceeds a third threshold value, and if the environmental humidity data exceeds the third threshold value, determining that an abnormal humidity condition exists;

[0026] If the ambient humidity data does not exceed the third threshold, but the humidity change rate is greater than the fourth threshold, and / or the current ambient humidity data deviates from the fluctuation times and fluctuation range of the same historical period, it is determined that a temperature anomaly exists.

[0027] Optionally, the third threshold is dynamically determined using an adaptive threshold adjustment algorithm;

[0028] Collect historical environmental data ;

[0029] The normal distribution model is used to analyze the distribution law of the data and determine the value of the third threshold under different seasons, different time periods and different environmental conditions; the normal distribution probability density function is assumed to be ,in, is the mean, is the standard deviation, and the value of the third threshold is determined by the mean and the standard deviation;

[0030] Obtain the current environmental multivariate data, and determine the change value of the third threshold based on the current environmental multivariate data and historical environmental data ;

[0031] use The value of the third threshold is adjusted. is the adjusted third threshold.

[0032] Optionally, after receiving a new data sample, the environmental interference model library further includes:

[0033] Based on the small batch gradient descent optimization algorithm, small batch data in the new data sample is selected to calculate the gradient and update the model parameters; let the model parameters be , the loss function is , the small batch data is , then the gradient descent update formula is ;in, is the learning rate;

[0034] When the model is updated, the difference between the new data sample and the existing model parameters is calculated, and only the model parameters related to the new data sample are adjusted.

[0035] Optionally, when there is an abnormal humidity situation, determining the source direction of the humidity interference in the environment includes using any of the following methods to determine:

[0036] The humidity sensing devices are distributed in different directions and positions in the environment. When it is detected that one or more collected environmental humidity data exceeds a third threshold, the source direction of the humidity interference in the environment is determined by comparing the humidity data differences of the humidity sensing devices.

[0037] Obtain the change trend and time series of the environmental humidity data collected by each humidity sensing device, analyze and determine which humidity sensing devices first detect that the environmental humidity data exceeds the third threshold, and determine the source direction of the humidity interference in the environment;

[0038] Historical humidity data and known interference behaviors are used as training data to train the long short-term memory network model. The environmental humidity data collected by each humidity sensor device is input into the trained model to output the source direction of the humidity interference in the environment.

[0039] In a second aspect of the present application, there is provided an air environment monitoring station humidity monitoring device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, an air environment monitoring station humidity monitoring method according to any one of the above-described methods is implemented.

[0040] In a third aspect of the present application, an air environment monitoring station is provided, comprising:

[0041] Humidity sensing devices are distributed at multiple locations in the air environment monitoring station;

[0042] Image acquisition devices are distributed at multiple locations in the air environment monitoring station;

[0043] According to the humidity monitoring equipment of the air environment monitoring station described above;

[0044] and a communication device for communicating with the humidity monitoring device of the air environment monitoring station and outputting the response data when it is determined that there is actual interference behavior.

[0045] The humidity monitoring method of the air environment monitoring station provided by the present application obtains the environmental humidity data collected by the humidity sensing device in real time, determines the source direction of the humidity interference in the environment when there is an abnormal humidity situation, and obtains real-time image data in the corresponding direction; matches the real-time image data with the interference model in the environmental interference model library, and determines the existence of actual interference behavior after matching the interference model; generates corresponding response data according to the type of actual interference behavior. The present application combines humidity monitoring with image analysis technology, and uses advanced data processing and artificial intelligence algorithms to automatically process and analyze environmental humidity problems, which can realize rapid and accurate diagnosis and positioning of problems, and promote the development of environmental monitoring technology to a higher level of automation and intelligence. In addition, the present application also provides humidity monitoring equipment for air environment monitoring stations and air environment monitoring stations with the above-mentioned technical effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0047] Figure 1 A flowchart of a specific implementation of the air environment monitoring station humidity monitoring method provided in this application;

[0048] Figure 2 A flowchart of the specific implementation process of determining whether there is an abnormal humidity situation based on the environmental humidity data;

[0049] Figure 3 This is the structural block diagram of the humidity monitoring equipment of the air environment monitoring station;

[0050] Figure 4 This is a schematic diagram of the deployment of air environment monitoring stations. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. It should be noted that, in the absence of conflict, the embodiments in the present disclosure and the features in the embodiments can be combined, separated, interchanged and / or rearranged with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0052] The terms used here are for the purpose of describing specific embodiments, and are not intended to be restrictive. As used here, unless the context clearly indicates otherwise, the singular forms "one (kind, person)" and "said (the)" are also intended to include plural forms. In addition, when the terms "comprise" and / or "include" and their variations are used in this specification, it is explained that there are stated features, integral bodies, steps, operations, parts, assemblies and / or their groups, but it is not excluded that there are or add one or more other features, integral bodies, steps, operations, parts, assemblies and / or their groups. It should also be noted that, as used here, the terms "substantially", "approximately" and other similar terms are used as approximate terms and not as degree terms, so that they are used to explain the inherent deviations of the measured values, calculated values ​​and / or the values ​​provided that will be recognized by those of ordinary skill in the art.

[0053] It is difficult to detect problems in current air environment monitoring stations at the first time. In terms of problem diagnosis, it is usually dependent on manual analysis of monitoring data, which is inefficient and prone to misjudgment. Based on this, the present application provides an air environment monitoring station with a highly integrated intelligent monitoring system. Using sensors and monitoring equipment, real-time monitoring of the environment inside and outside the station can be achieved.

[0054] The sensing equipment may specifically include:

[0055] Humidity sensing devices are distributed in multiple locations in the air environment monitoring station to monitor the ambient humidity of the air environment monitoring station. For example, multiple humidity sensing devices arranged outside the station building, together with humidity monitoring devices within 5-500 meters around the station building, form a humidity monitoring network, which can obtain real-time humidity changes in different areas around the station building 24 hours a day. These sensing devices are able to withstand severe weather conditions and provide accurate humidity readings.

[0056] Image acquisition devices are distributed at multiple locations outside the air environment monitoring station. The image acquisition devices can be high-resolution cameras and / or infrared cameras that operate around the clock to capture environmental changes or human activities that may be related to humidity anomalies.

[0057] Gas concentration monitoring equipment is distributed at multiple locations in the air environment monitoring station room to detect gas concentration data.

[0058] Equipment operation status monitoring equipment is distributed in multiple locations in the air environment monitoring station room to monitor the equipment in the monitoring station.

[0059] Through the Internet of Things technology, humidity monitoring, station building surrounding environment monitoring and various monitoring equipment inside the station building (such as gas concentration monitoring equipment, equipment operation status monitoring equipment, etc.) are organically connected.

[0060] In addition, servers and storage devices are also set up. The storage devices have sufficient data storage capacity to save the large amount of data collected from sensors and cameras. The servers have high performance to support real-time processing and analysis of data, supporting functions such as time series analysis, image recognition, and machine learning algorithms.

[0061] You can also set up communication equipment and establish a stable wireless network system to ensure seamless data transmission from sensors and cameras to the server. You can also set up network security equipment: deploy necessary network security measures to protect information security and privacy during data transmission.

[0062] A flowchart of a specific implementation of the air environment monitoring station humidity monitoring method provided in this application is as follows Figure 1 As shown, the method includes:

[0063] S101: Acquire environmental humidity data collected by a humidity sensor device in real time, determine whether there is an abnormal humidity situation based on the environmental humidity data, and when there is an abnormal humidity situation, determine the source direction of the humidity interference in the environment.

[0064] Specifically, 1 to 5 humidity sensing devices are arranged on the roof of the station building, and multiple humidity sensing devices or monitoring devices with humidity monitoring functions are arranged 5 to 500 meters around the station building to form a humidity monitoring network, so as to conduct 24-hour real-time tracking and diagnosis of environmental humidity data. After obtaining the environmental humidity data collected by the humidity monitoring network in real time, it is determined whether there is an abnormal humidity situation based on the environmental humidity data.

[0065] As a specific implementation method, judging whether there is a humidity anomaly based on the ambient humidity data can be obtained by comparing the collected ambient humidity data with a set threshold. For example, when the ambient humidity data exceeds the set threshold, it is determined that there is a humidity anomaly, and the subsequent tracing process of the humidity anomaly is initiated.

[0066] When there is an abnormal humidity situation, determining the source direction of the humidity disturbance in the environment includes using any of the following methods:

[0067] Method 1: The humidity sensing devices are distributed in different directions and positions in the environment. When it is detected that one or more collected environmental humidity data exceeds a third threshold, the source direction of the humidity interference in the environment is determined by comparing the humidity data differences of each humidity sensing device.

[0068] Method 2: Obtain the change trend and time series of the ambient humidity data collected by each humidity sensing device, analyze and determine which humidity sensing devices first detect that the ambient humidity data exceeds the third threshold, and determine the source direction of the humidity interference in the environment.

[0069] Method 3: Use historical humidity data and known interference behaviors as training data to train the long short-term memory network model. Input the environmental humidity data collected by each humidity sensor device into the trained model to automatically output the source direction of the humidity interference in the environment.

[0070] S102: According to the source direction of the humidity interference, call the image acquisition device in the air environment monitoring station to obtain real-time image data in the corresponding direction.

[0071] By calling the real-time image data collected by the image acquisition device, in the embodiment of the present application, the real-time image data in the corresponding direction is obtained according to the source direction of the humidity interference, without obtaining all the data, and the target data can be found quickly and accurately for analysis and diagnosis.

[0072] S103: Matching the real-time image data with an interference model in an environmental interference model library, and determining whether actual interference behavior exists after matching the interference model.

[0073] Wherein, a plurality of different interference models for identifying actual interference behaviors are pre-established in the environmental interference model library.

[0074] As a specific implementation method, the process of establishing the interference model includes the following steps:

[0075] Collect image data of various potential interference behaviors. For example, this includes manual spraying and fog cannon operations. Specifically, data can be obtained from historical records, surveillance videos, or specific experiments.

[0076] The key visual features are extracted from the image data, and the extracted key visual features are used as feature vectors as reference data for the recognition model. For example, for artificial sprinklers, the key visual features may include water flow shape, water flow size, water flow spray direction, water flow continuity, and the interaction between water flow and the surrounding environment.

[0077] A convolutional neural network is used to train the classifier to identify different types of interference behaviors, and labeled data of known interference behavior types are used for training.

[0078] Various verified interference models are integrated into an environmental interference model library for easy access. The environmental interference model library can be easily updated and expanded to include new interference types.

[0079] When matching the interference model, a specific image recognition algorithm and pattern matching technology can be used to compare various visual elements in the real-time image data with the feature patterns in the data model one by one. When a feature matching a certain model is found, the system will identify the corresponding interference type. Specifically: extract key features from the real-time image data to obtain a real-time image feature vector ; Using the cosine similarity formula To calculate the real-time image feature vector and the interference model feature vector by comparing the similarity between the real-time image data and each interference model, determine which interference model matches the scene of the real-time image data, and determine the existence of actual interference behavior after matching the interference model.

[0080] In an embodiment of the present application, by matching real-time image data with interference models in an environmental interference model library, it is monitored whether the real-time image data matches a certain type of interference model in the environmental interference model library, thereby determining whether there is interference from manual spraying or fog cannon operations.

[0081] For artificial spraying, the specific implementation of this process can be: if the similarity between the water flow characteristics in the real-time image data and the water flow characteristics in the artificial spraying interference model is greater than a first threshold, and the humidity change rate of the ambient humidity data at the corresponding position exceeds a preset range, then it is determined that there is actual interference behavior of artificial spraying. Water flow characteristics can include water flow shape, water flow size, water flow spray direction, water flow continuity, and the interaction mode between water flow and the surrounding environment.

[0082] For artificial spraying, the image recognition algorithm is used to detect whether there are continuous water-like objects in the video image, and whether the spray direction and range of the water flow conform to the characteristic pattern of artificial spraying. At the same time, a comprehensive judgment is made in combination with the sudden increase in humidity data and other relevant surrounding environmental data. For example, a comprehensive judgment can also be made in combination with a sound sensor. The working sound characteristics of the spraying equipment detected by the sound sensor can also assist in judging the actual interference behavior of artificial spraying.

[0083] Another specific implementation method may be: if the similarity between the vehicle outline and spray pattern in the real-time image data and the characteristics of the fog cannon interference model is greater than a second threshold, and the humidity change rate of the ambient humidity data at the corresponding position exceeds a preset range, then it is determined that there is actual interference behavior of the fog cannon, and the operation of the fog cannon has an impact on the humidity. Among them, the key visual features include the vehicle outline, which is used to identify the shape and size of the vehicle in the image and match it with the typical outline of the fog cannon. It also includes: Spray pattern: observe the shape, density and coverage of the spray, and these features should conform to the characteristics of the fog cannon spraying.

[0084] For fog cannon operation, we first identify the vehicle outline in the real-time image data, and accurately determine the unique appearance characteristics of the fog cannon based on the deep learning target detection algorithm. Then analyze its spray form. The water mist sprayed by the fog cannon usually has a large diffusion range and a specific atomization effect. The characteristic parameters of the water mist, such as droplet size and distribution density, are extracted through image processing technology and compared with the model characteristics of the fog cannon operation in the environmental interference model library. At the same time, based on the rapid changes in humidity data and the movement trajectory of the vehicle (the movement trajectory of the vehicle can be obtained through continuous frame image analysis), we further determine whether there is an impact of the fog cannon operation on humidity.

[0085] The embodiment of the present application can use a feature extraction algorithm based on a convolutional neural network to extract representative feature vectors. These feature vectors can accurately distinguish between fog cannon operations and other ordinary vehicles or scenes, and improve recognition accuracy. The specific convolutional neural network structure can adopt classic ResNet (residual network) or MobileNet (mobile network) architectures, and be fine-tuned according to actual scenarios.

[0086] This step identifies whether there is actual interference behavior by matching the real-time image data with the interference model in the environmental interference model library.

[0087] S104: after determining that there is actual interference behavior, generating corresponding response data according to the type of the actual interference behavior, the response data includes the type of interference behavior, the time of the interference behavior, the impact of the interference behavior on the humidity data, and recommended response measures.

[0088] Once the actual interference behavior is identified, the corresponding response data is generated according to the type of the actual interference behavior. The interference behavior type can clearly indicate the specific type of interference, such as manual spraying and fog cannon operation. The interference behavior time records the specific time when the interference occurred. Time data is important for tracing the cause of the event and analyzing the impact period of the interference on the monitoring data. The impact of the interference behavior on the humidity data, such as an abnormal increase or decrease in humidity readings. This helps to assess the severity and duration of the interference. Based on the type and impact of the interference, specific recommended response measures are generated. Recommended response measures can include operational adjustments (such as changing monitoring parameters or procedures), emergency measures (such as shutting down monitoring equipment in the affected area), and early warning prompts (sending early warning information to relevant personnel).

[0089] This application has established a rich library of environmental interference models and feature models to optimize the identification of common interference factors around the station building (such as artificial spraying, fog cannon operations, surrounding construction, etc.). Through deep learning analysis of monitoring video images, the characteristics of interference behaviors (such as the spray pattern of fog cannons, the water flow characteristics of artificial spraying, etc.) are extracted, and combined with the changes in monitoring data, the type and source direction of interference factors can be accurately determined. In the diagnostic analysis, the influence of interference factors on the monitoring data is automatically eliminated, and the true conditions of the environment inside and outside the station building are restored, making the monitoring results more accurate and reliable.

[0090] Based on the above embodiments, this application constructs a comprehensive diagnostic framework that comprehensively considers the multi-dimensional characteristics of monitoring data. Figure 2 As shown in the flowchart of the specific implementation process of judging whether there is a humidity abnormality according to the environmental humidity data, the process of judging whether there is a humidity abnormality according to the environmental humidity data in the present application may include the following steps:

[0091] S201: The environmental humidity data sequence With time series Together they form data points for trend analysis.

[0092] S202: Fit the data using a linear regression model to calculate the humidity change rate k.

[0093] The calculation formula of humidity change rate k is: ;in, and are the means of time and ambient humidity data respectively.

[0094] S203: Collect historical environmental humidity data and count different time windows The number and range of fluctuations of ambient humidity data within .

[0095] Frequency analysis focuses on the data in a specific time window. The stability of the data is determined by counting the fluctuation frequency.

[0096] S204: Determine whether the environmental humidity data exceeds a third threshold value. If the environmental humidity data exceeds the third threshold value, determine that an abnormal humidity condition exists.

[0097] S205: If the ambient humidity data does not exceed the third threshold, but the humidity change rate is greater than the fourth threshold, and / or the current ambient humidity data deviates from the fluctuation times and fluctuation range of the same historical period, it is determined that a temperature anomaly exists.

[0098] In the humidity diagnosis of this embodiment, the comprehensive environmental humidity data , Humidity change rate in a short period of time , the fluctuation range of humidity in the same period of history Many factors, even the environmental humidity data Does not exceed the set threshold (third threshold ), but the rate of change is abnormal ( > ) and deviate from historical patterns ( ), it is still determined that temperature abnormality exists, thereby significantly improving the accuracy and reliability of diagnosis.

[0099] Furthermore, in the embodiment of the present application, based on the deep statistical analysis of big data and the dynamic perception of real-time environmental factors, intelligent adaptive adjustment of the threshold of the monitoring index is realized. The third threshold can be dynamically determined by the adaptive threshold adjustment algorithm. The specific determination method is:

[0100] Collect historical environmental data .

[0101] The normal distribution model is used to analyze the distribution law of the data and determine the value of the third threshold under different seasons, different time periods and different environmental conditions; the normal distribution probability density function is assumed to be ,in, is the mean, is the standard deviation, and the value of the third threshold is determined by the mean and the standard deviation.

[0102] Obtain the current environmental multivariate data, and determine the change value of the third threshold based on the current environmental multivariate data and historical environmental data .

[0103] use The value of the third threshold is adjusted. is the adjusted third threshold.

[0104] Among them, the perception of the current environmental multivariate data can use a variety of sensors and data acquisition equipment to obtain seasonal information, weather conditions (temperature ,precipitation , air pressure Based on these real-time information, the threshold range of monitoring indicators is automatically adjusted. Taking the hot and humid summer season as an example, by analyzing the distribution of historical summer humidity data, the upper limit of the threshold of humidity anomaly is appropriately increased. ( Based on historical data distribution and real-time environmental factors), it avoids false alarms caused by normal seasonal humidity increases, effectively improving the accuracy and effectiveness of diagnosis.

[0105] In addition, this application also adopts an innovative mechanism that combines the advantages of online learning and incremental learning to achieve continuous optimization and adaptive adjustment of the image recognition model. During operation, new real-time image data and interference behavior sample streams are acquired in real time, and the parameters and structure of the image recognition model are updated in real time. For example, when new construction equipment or working methods appear around the station building, the characteristics of the new samples can be automatically identified and incorporated into the model training, continuously optimizing the model's ability to recognize new interference behaviors, and ensuring that the system always maintains a high recognition accuracy and adaptability.

[0106] After receiving new data samples, the environmental interference model library immediately applies the optimization algorithm based on mini-batch gradient descent, and updates the model with a part (a mini-batch) of the new data each time. Mini-batch data in the new data samples are selected to calculate the gradient and update the model parameters. Let the model parameters be , the loss function is , the small batch data is , then the gradient descent update formula is ;in, is the learning rate.

[0107] When the model is updated, the difference between the new data sample and the existing model parameters is calculated, and only the model parameters related to the new data sample are adjusted.

[0108] This real-time update method enables the model to quickly adapt to the emerging interference behavior characteristics and maintain sensitivity to environmental changes. At the same time, combined with the incremental learning mechanism, instead of re-training all historical data, only the incremental information contained in the new samples is used to update the model. During the model update process, by calculating the difference between the new sample and the existing model parameters, only some model parameters related to the new information are adjusted, thereby ensuring that the model learns new features while effectively avoiding overfitting of historical data and ensuring the generalization ability of the model.

[0109] At the same time, in order to prevent the model from overfitting in the process of continuously learning new samples and to ensure the stability and generalization ability of the model, this application also introduces model regularization technology, among which L2 regularization is one of the key means.

[0110] Specifically, L2 regularization constrains the size of model parameters by adding a regularization term to the model's loss function. Let the regularization term coefficient be , the regularization term is , which is used to penalize the situation where the model parameters are too large. During the model training process, the optimization algorithm will consider the influence of the regularization term while pursuing the minimization of the loss function, thereby prompting the model to choose relatively small parameter values. This mechanism helps to reduce the complexity of the model, prevent the model from over-learning the noise and local features in the training data, and improve the model's generalization ability for unseen data. In this way, model regularization technology is combined with online learning and incremental learning to ensure that the adaptive model training algorithm maintains good performance and stability while constantly adapting to new data, providing solid technical support for intelligent analysis based on image recognition.

[0111] In addition, the present application also provides a humidity monitoring device for an air environment monitoring station, such as Figure 3 As shown, the humidity monitoring device of the air environment monitoring station includes a memory 31 and a processor 32. The memory 31 stores a computer program. When the computer program is executed by the processor 32, any of the above-mentioned humidity monitoring methods of the air environment monitoring station is implemented.

[0112] In addition, the present application also provides an air environment monitoring station, such as Figure 4 The deployment diagram of the air environment monitoring station is shown in the figure. The air environment monitoring station includes:

[0113] Humidity sensing equipment, corresponding Figure 4 The humidity sensors are distributed at multiple locations in the air environment monitoring station.

[0114] Image acquisition device, corresponding to Figure 4 The cameras are distributed at multiple locations in the air environment monitoring station.

[0115] And according to the above-mentioned air environment monitoring station humidity monitoring equipment, corresponding to Figure 4 The core monitoring system in .

[0116] The communication device (not shown in the figure) is used to communicate with the humidity monitoring device of the air environment monitoring station, and output the response data when it is determined that there is actual interference behavior.

[0117] All data from IoT devices such as sensors and monitoring equipment are connected to the core monitoring system, namely the humidity monitoring equipment in the air environment monitoring station for further analysis and processing.

[0118] In addition, the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements any of the above-mentioned air environment monitoring station humidity monitoring methods.

[0119] Computer readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0120] The professionals should also be further aware that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0121] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0122] The specific implementation methods described above further illustrate the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only the specific implementation method of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A humidity monitoring method for an air environment monitoring station, characterized in that: The air environment monitoring station is provided with a plurality of humidity sensing devices for monitoring the environmental humidity. The method comprises: Acquire the environmental humidity data collected by the humidity sensing device in real time, determine whether there is an abnormal humidity situation based on the environmental humidity data, and when there is an abnormal humidity situation, determine the source direction of the humidity interference in the environment; According to the source direction of the humidity interference, the image acquisition device of the air environment monitoring station is called to obtain real-time image data in the corresponding direction; Matching the real-time image data with an interference model in an environmental interference model library, and determining the presence of actual interference behavior after matching the interference model; wherein the environmental interference model library has pre-established a plurality of different interference models for identifying actual interference behavior; After determining that there is an actual interference behavior, generating corresponding response data according to the type of the actual interference behavior, the response data including the type of interference behavior, the time of the interference behavior, the impact of the interference behavior on the humidity data, and recommended response measures; The process of establishing the interference model includes: Collect image data of various potential disturbance behaviors including artificial spraying and fog cannons; Key visual features are extracted from the image data, and the extracted key visual features are used as feature vectors as reference data for the recognition model; a convolutional neural network is used to train a classifier to identify different types of interference behaviors, and label data of known types of interference behaviors are used for training during the training process.

2. The humidity monitoring method of an air environment monitoring station according to claim 1, characterized in that: The step of matching the real-time image data with an interference model in an environmental interference model library and determining the existence of actual interference behavior after matching the interference model comprises: Extract key features from the real-time image data to obtain a real-time image feature vector ; Using the cosine similarity formula To calculate the real-time image feature vector and the interference model feature vector The similarity between By comparing the similarity between the real-time image data and each interference model, it is determined which interference model the scene of the real-time image data matches, and after matching the interference model, it is determined that there is actual interference behavior.

3. The humidity monitoring method of an air environment monitoring station according to claim 1, characterized in that: The step of matching the real-time image data with an interference model in an environmental interference model library and determining the existence of actual interference behavior after matching the interference model comprises: If the similarity between the water flow characteristics in the real-time image data and the water flow characteristics in the artificial spray interference model is greater than a first threshold, and the humidity change rate of the environmental humidity data at the corresponding position exceeds a preset range, it is determined that there is actual interference behavior of artificial spraying; wherein the water flow characteristics include water flow shape, water flow size, water flow spray direction, water flow continuity, and the interaction mode between the water flow and the surrounding environment; If the feature similarity between the vehicle outline and spray pattern in the real-time image data and the fog cannon vehicle interference model is greater than a second threshold, and the humidity change rate of the ambient humidity data at the corresponding position exceeds a preset range, it is determined that there is actual interference behavior from the fog cannon vehicle.

4. The humidity monitoring method of an air environment monitoring station according to any one of claims 1 to 3, characterized in that: The determining whether there is an abnormal humidity situation according to the environmental humidity data includes: The environmental humidity data series With time series Together they form data points for trend analysis; Use linear regression model to fit the data and calculate the humidity change rate ;in, and They are the means of time and ambient humidity data respectively; Collect historical environmental humidity data and count different time windows The number and range of fluctuations of ambient humidity data within ; Determining whether the environmental humidity data exceeds a third threshold value, and if the environmental humidity data exceeds the third threshold value, determining that an abnormal humidity condition exists; If the ambient humidity data does not exceed the third threshold, but the humidity change rate is greater than the fourth threshold, and / or the current ambient humidity data deviates from the fluctuation times and fluctuation range of the same historical period, it is determined that a temperature anomaly exists.

5. The humidity monitoring method of an air environment monitoring station according to claim 4, characterized in that: The third threshold is dynamically determined by using an adaptive threshold adjustment algorithm; Collect historical environmental data ; The normal distribution model is used to analyze the distribution law of the data and determine the value of the third threshold under different seasons, different time periods and different environmental conditions; the normal distribution probability density function is assumed to be ,in, is the mean, is the standard deviation, and the value of the third threshold is determined by the mean and the standard deviation; Obtain the current environmental multivariate data, and determine the change value of the third threshold based on the current environmental multivariate data and historical environmental data ; use The value of the third threshold is adjusted. is the adjusted third threshold.

6. The humidity monitoring method of an air environment monitoring station according to any one of claims 1 to 3, characterized in that: After receiving the new data sample, the environmental interference model library further includes: Based on the small batch gradient descent optimization algorithm, small batch data in the new data sample is selected to calculate the gradient and update the model parameters; let the model parameters be , the loss function is , the small batch data is , then the gradient descent update formula is ;in, is the learning rate; When the model is updated, the difference between the new data sample and the existing model parameters is calculated, and only the model parameters related to the new data sample are adjusted.

7. The humidity monitoring method of an air environment monitoring station according to any one of claims 1 to 3, characterized in that: When there is an abnormal humidity situation, determining the source direction of the humidity interference in the environment includes using any of the following methods to determine: The humidity sensing devices are distributed in different directions and positions in the environment. When it is detected that one or more collected environmental humidity data exceeds a third threshold, the source direction of the humidity interference in the environment is determined by comparing the humidity data differences of the humidity sensing devices. Obtain the change trend and time series of the environmental humidity data collected by each humidity sensing device, analyze and determine which humidity sensing devices first detect that the environmental humidity data exceeds the third threshold, and determine the source direction of the humidity interference in the environment; Historical humidity data and known interference behaviors are used as training data to train the long short-term memory network model. The environmental humidity data collected by each humidity sensor device is input into the trained model to output the source direction of the humidity interference in the environment.

8. A humidity monitoring device for an air environment monitoring station, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the humidity monitoring method of an air environment monitoring station according to any one of claims 1 to 7 is implemented.

9. An air environment monitoring station, characterized in that: include: Humidity sensing devices are distributed at multiple locations in the air environment monitoring station; Image acquisition devices are distributed at multiple locations in the air environment monitoring station; The humidity monitoring device of the air environment monitoring station according to claim 8; and a communication device for communicating with the humidity monitoring device of the air environment monitoring station and outputting the response data when it is determined that there is actual interference behavior.

Images