RTU intelligent data analysis method, remote terminal equipment, medium and product

Abstract

An RTU intelligent data analysis method, remote terminal equipment, medium and product, relating to the field of electrical digital data processing technology. The method includes: collecting real-time hydrological data through sensors at preset time intervals; obtaining real-time video data through cameras; determining the target hydrological parameter value according to real-time video data when all sensors collecting target hydrological parameters fail and the target hydrological parameters belong to the hydrological parameters in the first hydrological parameter set; obtaining the target hydrological parameter value through a prediction model when all sensors collecting target hydrological parameters fail and the target hydrological parameters belong to the hydrological parameters in the second hydrological parameter set; updating real-time hydrological data according to the target hydrological parameters and the target hydrological parameter values; reporting real-time hydrological data and real-time video data according to a preset data reporting mechanism. The implementation of the above technical solution improves the accuracy of the hydrological data transmitted by RTU to the control center.

Description

Technical Field

[0001] The present application relates to the technical field of electrical digital data processing, and in particular to an RTU intelligent data analysis method, remote terminal equipment, medium and product. Background Art

[0002] In the field of hydrological monitoring, remote terminal units (RTUs) play a vital role. As a device that can collect data at remote sites and transmit it to the central control center, RTU is a key component for realizing automated, intelligent monitoring and management of hydrological monitoring systems. It can collect various field data, including but not limited to temperature, flow, water level and other information, and transmit this information to the remote control center to provide operators with real-time feedback on the field situation.

[0003] At present, RTU usually collects various hydrological data through sensors, and then sends the hydrological data directly to the control center for data processing and abnormality judgment.

[0004] However, due to the long-term exposure of sensors to complex hydrological environments, such as high humidity, water flow impact, sediment erosion, etc., they are prone to failure. When a sensor fails, the hydrological data obtained by the RTU is prone to deviation, missing or even errors, which in turn makes the hydrological data sent to the control center abnormal, resulting in cumbersome data processing, long judgment and response time for abnormal situations, and affecting the timeliness of data processing. Summary of the invention

[0005] The present application provides an RTU intelligent data analysis method, remote terminal equipment, medium and product for improving the accuracy of hydrological data transmitted by RTU to a control center.

[0006] In a first aspect, the present application provides an RTU intelligent data analysis method, the method comprising: collecting real-time hydrological data through sensors at preset time intervals; the real-time hydrological data includes water level, flow rate, water temperature, turbidity, pH value, and dissolved oxygen content; acquiring real-time video data through a camera; in the case that all sensors collecting target hydrological parameters fail and the target hydrological parameter belongs to a hydrological parameter in a first hydrological parameter set, determining the target hydrological parameter value according to the real-time video data; the target hydrological parameter is one of all hydrological parameters; the first hydrological parameter set is a set of hydrological parameters whose hydrological parameter values ​​are obtained according to images or videos, including water level, flow rate, and turbidity; when collecting target hydrological parameters, When all sensors of the target hydrological parameters fail and the target hydrological parameter belongs to the hydrological parameters in the second hydrological parameter set, the values ​​of the environmental parameters and the hydrological parameters other than the target hydrological parameters in the real-time hydrological data are input into the prediction model to obtain the target hydrological parameter value; the second hydrological parameter set is a set of other hydrological parameters in the real-time hydrological data except the hydrological parameters in the first hydrological parameter set, including water temperature, pH value and dissolved oxygen; the real-time hydrological data is updated according to the target hydrological parameters and the target hydrological parameter value; the real-time hydrological data and the real-time video data are reported according to a preset data reporting mechanism; the data reporting mechanism includes a hydrological data reporting mechanism and a video data reporting mechanism.

[0007] By adopting the above technical solution, when a sensor failure is detected, the correct hydrological parameter values ​​that the sensor may collect during the period from the sensor failure to the staff replacing or repairing the sensor are obtained by performing image recognition on the video frames in the real-time video or by using a prediction model. Then, the abnormal hydrological parameter values ​​in the hydrological data transmitted to the control center are replaced with the correct hydrological parameter values, thereby reducing the abnormal data transmitted to the control center during this period, avoiding a large amount of abnormal data transmission caused by sensor failure, and improving the accuracy of the hydrological data transmitted by the RTU to the control center.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the real-time hydrological data and the real-time video data are reported according to a preset data reporting mechanism, specifically including: determining whether a target abnormality occurs based on the real-time hydrological data and the real-time video data; the target abnormality includes abnormal water flow, increased water pollution, and intrusion by unfamiliar users; if so, determining the target reporting object and target reporting data based on the target abnormality; adjusting the reporting objects in the preset first hydrological data reporting mechanism and the preset video data reporting mechanism based on the target reporting object; reporting the target reporting data according to the adjusted first hydrological data reporting mechanism; reporting the real-time video data in real time according to the adjusted video data reporting mechanism; if not, reporting the real-time hydrological data according to the preset second hydrological data reporting mechanism.

[0009] By adopting the above technical solution, abnormal situations can be judged on site as soon as the hydrological data and video data are collected, avoiding the problem of slow response to abnormal situations caused by network delays and improving the speed of abnormal response. When abnormal situations are found, they are actively reported through different reporting mechanisms, realizing the active identification and reporting of abnormal situations by remote terminal devices, without having to wait until the data is transmitted to the control center, and then the control center makes abnormal judgments, which improves the timeliness of data processing and enables data processing as soon as it is obtained.

[0010] In combination with some embodiments of the first aspect, in some embodiments, if the target abnormality is a water flow abnormality, the target reporting object and the target reporting data are determined according to the target abnormality, specifically including: when the target abnormality is a water flow abnormality, predicting the water flow change within a preset time in the future through a machine learning model to obtain a prediction result; using the prediction result and the real-time hydrological data as the target reporting data; when the target abnormality is aggravated water pollution, determining the emission area where the pollution source is located and the cause of the aggravated pollution; using the emission area, the cause of the aggravated pollution and the real-time hydrological data as the target reporting data; when the target abnormality is an invasion by a stranger user, determining the location of the stranger user according to the real-time video data; using the location of the stranger user and the real-time hydrological data as the target reporting data; when the target abnormality is a water flow abnormality or an invasion by a stranger user, adding the monitoring site closest to the current monitoring area to the preset reporting object to obtain the target reporting object; when the target abnormality is aggravated water pollution, adding the monitoring site closest to the emission area to the preset reporting object to obtain the target reporting object.

[0011] By adopting the above technical solution, the target reporting object and target reporting data can be accurately determined for different target abnormal situations. When the water flow is abnormal, the machine learning model is used to predict the future water flow changes, and the prediction results and real-time hydrological data are used as reporting data to provide a basis for making water conservancy decisions; when water pollution worsens, the emission area and cause of pollution sources are clarified, and reported together with real-time hydrological data to provide a basis for treating water pollution problems; when a stranger invades, its location is determined based on the real-time video and reported together with the real-time hydrological data, so that the staff can locate the stranger's location at the first time. At the same time, in determining the reporting object, the monitoring station closest to the abnormal area is added to the reporting object, so that the staff closest to the abnormal area can quickly rush to the abnormal area to deal with the abnormal situation, thereby improving the efficiency of abnormal handling.

[0012] In combination with some embodiments of the first aspect, in some embodiments, when the target abnormal condition is aggravated water pollution, determining the emission area where the pollution source is located and the cause of the aggravated pollution, specifically including: when the target abnormal condition is aggravated water pollution and the water flow rate is greater than a preset threshold, obtaining historical video data of other remote terminal devices in the direction opposite to the water flow direction to obtain a historical video data set; determining the emission area where the pollution source is located based on the historical video data set; when the emission area does not exist in the historical emission area, determining the cause of the aggravated pollution is a newly added emission area; when the emission area exists in the historical emission area, determining the cause of the aggravated pollution is an increase in the pollutant emission in the emission area.

[0013] By adopting the above technical solution, by obtaining the historical video data of other remote terminal devices in the opposite direction of the water flow, a historical video data set is formed, and the emission area is determined based on this. The impact of water flow on the diffusion of pollutants is taken into account, and the accuracy of locating the pollution source is effectively improved by using the reverse tracing method. At the same time, the currently determined emission area is compared with the historical emission area to quickly determine whether the increase in pollution is due to the addition of a new emission area or the increase in pollutant emissions in the existing emission area. It provides a key basis for staff to formulate targeted pollution control measures, helps to curb pollution at the source, and makes the control work more accurate and efficient.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after the step of determining the target reporting object and target reporting data based on the target abnormal situation, the method also includes: evaluating the target processing difficulty and target processing time of the target abnormal situation based on the target abnormal situation and historical abnormal processing data; and reporting the target processing difficulty and the target processing time to the target reporting object.

[0015] By adopting the above technical solution, the target processing difficulty and target processing time of the target abnormal situation are evaluated by combining the target abnormal situation with the historical abnormal processing data. When the target reporting object receives the abnormal information, it can not only understand the situation of the abnormality itself, but also have a more accurate estimate of the difficulty and time required to handle the abnormality, which provides a basis for relevant departments to allocate resources, arrange personnel and plan response plans.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of evaluating the processing difficulty and target processing time of the target abnormal situation based on the target abnormal situation and historical abnormal processing data, the method also includes: determining the target abnormality level of the target abnormal situation based on the type of the target abnormal situation, the target processing difficulty and the target processing time; the target abnormality level includes level one, level two and level three; when the target abnormality level is level two or level three, executing a corresponding alarm mechanism.

[0017] By adopting the above technical solution, when the abnormality level of the abnormal situation is level 2 or level 3 and the abnormal situation is relatively serious, the alarm mechanism is immediately executed to achieve a rapid response to the abnormal situation, so that the relevant departments can receive the alarm in time and take countermeasures to resolve the abnormal situation as soon as possible.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of acquiring real-time video data through a camera, the method also includes: acquiring the light intensity around the camera; when the light intensity exceeds a preset threshold, acquiring the actual color data of a preset reference object in each video frame in the real-time video data; the preset reference object is a standard color plate fixed on the shore; comparing the actual color data corresponding to each video frame with the preset color data of the preset reference object in sequence to obtain the color similarity of each video frame; when the color similarity of the first video frame does not exceed the preset threshold, calibrating the color of the corresponding first video frame to obtain a second video frame; the first video frame is one of all the video frames in the real-time video data; the color similarity of the second video frame exceeds the preset threshold; replacing the first video frame in the real-time video data with the second video frame.

[0019] By adopting the above technical solution, the color data of each video frame in the video data is calibrated to ensure the color accuracy of the real-time video data, improve the reliability of the recognition results when recognizing the image, and thus improve the accuracy of the hydrological data. At the same time, it provides clearer and more accurate visual information for identifying abnormal situations such as aggravated water pollution and intrusion by unfamiliar users, avoiding misjudgment or omission of abnormal situations due to color distortion, and providing strong support for ensuring the accurate collection and analysis of hydrological data.

[0020] In a second aspect, an embodiment of the present application provides a remote terminal device, including a sensor, a camera, and a data processing server, wherein the data processing server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the exercise prescription generation system to perform the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer-readable storage medium comprising instructions, which, when executed on a remote terminal device, causes the remote terminal device to execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a remote terminal device, enables the remote terminal device to execute the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] It is understandable that the remote terminal device provided in the second aspect, the storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the method provided in the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0025] 1. This application obtains the correct hydrological parameter values ​​collected by the sensor during the period from sensor failure to sensor replacement or repair by staff through image recognition of video frames in real-time video or prediction through a predictive model, and then replaces the abnormal hydrological parameter values ​​in the hydrological data transmitted to the control center with the correct hydrological parameter values, thereby reducing the abnormal data transmitted to the control center during this period and improving the accuracy of the hydrological data transmitted to the control center by the RTU.

[0026] 2. This application realizes the active identification and reporting of abnormal situations by remote terminal devices by identifying different abnormal situations and executing corresponding reporting mechanisms according to the abnormal situations. At the same time, when processing data and judging abnormal situations, there is no need to wait until the data is transmitted to the control center, and then the control center processes the data and judges the abnormal situation. The data can be processed and the abnormal situation can be judged as soon as the data is obtained, which improves the timeliness of data processing.

[0027] 3. This application ensures the color accuracy of real-time video data by calibrating the color data of each video frame in the video data, improves the reliability of the recognition results when recognizing images, thereby improving the accuracy of hydrological data obtained through image recognition and the accuracy of hydrological data transmitted by RTU to the control center. At the same time, it provides clearer and more accurate visual information for identifying abnormal situations such as aggravated water pollution and intrusion by unfamiliar users, avoids misjudgment or omission of abnormal situations due to color distortion, and provides strong support for ensuring the accurate collection and analysis of hydrological data. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a structural diagram of a system architecture to which the RTU intelligent data analysis method in the embodiment of the present application can be applied;

[0029] Figure 2This is a flow chart of the RTU intelligent data analysis method in the embodiment of the present application;

[0030] Figure 3 This is another flow chart of the RTU intelligent data analysis method in the embodiment of the present application;

[0031] Figure 4 It is a schematic diagram of an exemplary hardware structure of a remote terminal device in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0033] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0034] Figure 1 It is a structural diagram of a system architecture to which the RTU intelligent data analysis method in the embodiment of the present application can be applied.

[0035] See also Figure 1 ,The hydrological monitoring system includes remote terminal equipment and control center.

[0036] As the core component of the system, the control center is used to process the data collected by the remote terminal devices and send control instructions to the remote terminal devices.

[0037] The remote terminal device includes a data processing server, a camera, and a sensor. The camera is used to collect image information of the monitoring area and transmit the collected image information to the data processing server. The sensor is used to collect hydrological information of the monitoring area and transmit the collected hydrological information to the data processing server. The data processing server is used to process the image information transmitted by the camera and the hydrological data transmitted by the sensor, and transmit the processed data to the control center.

[0038] Through the above system architecture, the hydrological monitoring system can observe the hydrological conditions and actual on-site images of the monitoring area in real time based on the data collected by remote terminal equipment, providing comprehensive and accurate information support for the operation and management of water conservancy facilities, water resources scheduling, flood prevention and disaster reduction, and other tasks.

[0039] In the related art, high-precision sensors can be used to accurately obtain hydrological data of the monitoring area. However, since the sensors are exposed to complex hydrological environments for a long time, such as high humidity, water flow impact, sediment erosion, etc., they are prone to failure. When the sensor fails, the hydrological data obtained by the RTU is prone to deviation, missing or even errors, which makes the hydrological data sent to the control center abnormal, resulting in cumbersome data processing in the control center, and a long time for judging and responding to abnormal situations, affecting the timeliness of data processing. At the same time, in the process of sending the data collected by the RTU to the control center, due to the large amount of data transmission, network congestion, data loss and other problems are prone to occur, resulting in delayed or incomplete data obtained by the control center, which in turn affects the real-time and accurate judgment of the hydrological situation and the timeliness of data processing.

[0040] By using the RTU intelligent data analysis method in the embodiment of the present application, when a sensor fails, the correct hydrological parameter values ​​from the time the sensor fails to the time the staff replaces or repairs the sensor are obtained by recognizing the image or predicting through a prediction model, and then the abnormal hydrological parameter values ​​in the hydrological data transmitted to the control center are replaced with the correct hydrological parameter values, thereby reducing the abnormal data transmitted to the control center during this period and improving the accuracy of the hydrological data transmitted to the control center by the RTU. At the same time,

[0041] Abnormal situations are judged on-site as soon as the hydrological data and video data are collected, and when abnormal situations are found, they are actively reported through different reporting mechanisms. This enables remote terminal devices to actively identify and report abnormal situations, without having to wait until the data is transmitted to the control center for the control center to make abnormal judgments. This improves the timeliness of data processing, and allows the data to be processed as soon as it is obtained.

[0042] Combine the following Figure 2 To illustrate the method of the embodiment of the present application.

[0043] See also Figure 2 , which is a flow chart of the RTU intelligent data analysis method in an embodiment of the present application.

[0044] S201. Collect real-time hydrological data through sensors at preset time intervals.

[0045] Among them, real-time hydrological data include water level, flow rate, water temperature, turbidity, pH value, dissolved oxygen content, etc. Sensors include but are not limited to water level sensors, flow rate sensors, water temperature sensors, turbidity sensors, pH sensors, dissolved oxygen sensors, etc., which are not limited here.

[0046] Specifically, water level data is obtained through a water level sensor. The water level sensor uses the principle of ultrasonic ranging to transmit ultrasonic waves to the water surface, receive reflected waves, and calculate the water level based on the round-trip time of the sound waves.

[0047] The flow rate is obtained through the flow rate sensor. The flow rate sensor uses the principle of electromagnetic induction. When water flows through the sensor, it cuts the magnetic lines of force to generate induced electromotive force. The water flow rate is calculated by measuring the magnitude of the induced electromotive force.

[0048] The water temperature data is obtained through the water temperature sensor. The water temperature sensor is based on the characteristics of thermistor. Its resistance value changes with temperature. The water temperature is determined by measuring the resistance value.

[0049] Obtain turbidity data through a turbidity sensor. The turbidity sensor uses the principle of light scattering. When light is irradiated onto suspended particles in the water, it will scatter. The turbidity of the water is determined by detecting the intensity of the scattered light.

[0050] The pH value is obtained through a pH sensor. The pH sensor is based on the electrochemical principle and determines the pH value of the solution by measuring the potential difference between the electrode and the reference electrode.

[0051] The dissolved oxygen data is obtained through the dissolved oxygen sensor. The dissolved oxygen sensor uses the membrane electrode method. Oxygen enters the electrolyte through the breathable membrane, and a reduction reaction occurs on the cathode to generate current. The dissolved oxygen content in the water is determined by measuring the current.

[0052] S202: Acquire real-time video data through a camera.

[0053] Real-time video data is collected through the camera in real time.

[0054] S203: When all sensors collecting the target hydrological parameter fail and the target hydrological parameter belongs to the hydrological parameter in the first hydrological parameter set, determine the target hydrological parameter value according to the real-time video data.

[0055] The target hydrological parameter is one of all the hydrological parameters. The first hydrological parameter set is a set of hydrological parameters whose hydrological parameter values ​​can be obtained according to an image or video, including water level, flow velocity, and turbidity.

[0056] Specifically, first determine whether all sensors for collecting target hydrological parameters are faulty. First, combine historical data and current environmental parameters to determine the reasonable parameter value range of each hydrological parameter. Among them, historical data includes historical environmental parameters and their corresponding historical hydrological data. Compare and analyze the current environmental parameters with historical environmental parameters, obtain historical hydrological data corresponding to historical environmental parameters that match the current environmental parameters, and determine, based on the historical hydrological data, that the maximum reasonable parameter value of each hydrological parameter is the maximum value of the corresponding hydrological parameter in the historical hydrological data, and the minimum reasonable parameter value is the minimum value of the corresponding hydrological parameter in the historical hydrological data. Then, compare whether the target hydrological parameter in the real-time hydrological data continues to exceed the corresponding reasonable parameter value range within a preset time range. If so, it is determined that the sensor corresponding to the target hydrological parameter is faulty. In the case of a spare sensor, the hydrological parameter is continuously collected through the spare sensor, and it is determined whether the spare sensor is faulty. If so, it is determined that all sensors for collecting the target hydrological parameters are faulty.

[0057] In some embodiments, it is also possible to further determine whether the sensor for collecting the target hydrological parameter is faulty based on the correlation characteristics between each hydrological parameter and the environmental parameter. For example, the water level is proportional to the rainfall. If the rainfall is monitored to increase but the water level drops, it is determined that the sensor for collecting the water level is faulty.

[0058] When all sensors collecting the target hydrological parameter fail and the target hydrological parameter belongs to the hydrological parameter in the first hydrological parameter set, the target hydrological parameter value is determined according to the real-time video data and the target hydrological parameter.

[0059] If the target hydrological parameter is water level, firstly, obtain real-time video data and determine the position of fixed reference points (such as bridge piers in the monitoring area, corners of specific buildings on the shore, etc.) in the picture through image recognition technology, as well as the actual height of the water surface and the fixed reference points. Then, calculate the current water level value based on the pre-established relationship between the water surface and the fixed reference points and the proportional relationship between the image pixels and the actual physical size.

[0060] In some embodiments, a model can also be established through a machine learning algorithm to obtain water level data. The model uses a large amount of real-time video data containing different water level states and their corresponding actual water level measurements as training samples. During the training process, the model will automatically learn the complex relationship between the features in the video image and the water level, such as the mapping relationship between the texture of the water surface in the video, color changes, and the relative position relationship with a fixed reference point and the actual water level value. After sufficient training, when new real-time video data is input, the model can accurately predict the water level value corresponding to the current video screen based on the learned pattern.

[0061] If the target hydrological parameter is flow velocity. First, a sequence of continuous video frames is selected from the real-time video data. In these video frames, the characteristic objects in the water flow (such as floating leaves, small wood blocks and other natural objects, or artificial floating objects with obvious markings) are identified by image recognition technology. Then, for each frame of the image, the image recognition algorithm is used to determine the position coordinates of the characteristic object in the picture. By analyzing the changes in the position coordinates of the characteristic objects in adjacent video frames and combining the time interval of the video frames, the displacement of the characteristic object per unit time is calculated. Then, according to the pre-established proportional relationship between the image pixels and the actual physical size, the displacement of the characteristic object in the image is converted into the actual physical displacement, and the actual moving speed of the characteristic object is obtained, which is used as the flow velocity of the water flow.

[0062] In some embodiments, in order to improve the accuracy of flow rate calculation, a machine learning algorithm can also be used to establish a model. The model uses a large amount of real-time video data containing different flow rate states and their corresponding actual flow rate measurements as training samples. During the training process, the model will automatically learn the complex relationship between the features in the video image and the flow rate, such as the mapping relationship between the texture, color changes, and movement trajectory of the feature objects in the video and the actual flow rate value. After sufficient training, when new real-time video data is input, the model can accurately predict the water flow velocity value corresponding to the current video screen based on the pattern it has learned.

[0063] If the target hydrological parameter is turbidity. First, obtain real-time video data and use image recognition technology to analyze the visual features of the water body in the video, such as color and texture. Different turbidity of water has different characteristics of reflection and scattering light, which presents different colors and textures in the video image. Then, according to the pre-established relationship model between turbidity and these visual features, the extracted color and texture features are substituted into the model to calculate the approximate turbidity value of the current water body.

[0064] This relationship model is built based on machine learning algorithms (such as convolutional neural networks, etc.), and is trained using a large amount of real-time video data containing different turbidity states and their corresponding actual turbidity measurements as training samples. Before training the model, first of all, these real-time video data must be preprocessed to improve image quality through operations such as image enhancement and noise reduction. Then, advanced image recognition and computer vision techniques are used to extract key visual features related to turbidity from video frames, including color features and texture features. After feature extraction, the model is trained, and the model parameters are optimized through cross-validation. Finally, the model is repeatedly adjusted and optimized based on evaluation indicators such as mean square error, mean absolute error, and determination coefficient, so that the trained model can accurately reflect the mapping relationship between turbidity and visual features.

[0065] S204. When all sensors collecting target hydrological parameters fail and the target hydrological parameters belong to the hydrological parameters in the second hydrological parameter set, the values ​​of the environmental parameters and the hydrological parameters other than the target hydrological parameters in the real-time hydrological data are input into the prediction model to obtain the target hydrological parameter values.

[0066] The second hydrological parameter set is a set of other hydrological parameters in the real-time hydrological data except the hydrological parameters in the first hydrological parameter set, including water temperature, pH value and dissolved oxygen content.

[0067] Specifically, environmental parameters are obtained, including rainfall, temperature, air pressure, humidity, sunshine duration, etc. These environmental factors will have different degrees of impact on hydrological parameters. For example, the amount of rainfall is directly related to the replenishment of river water, which in turn affects hydrological parameters such as flow and water level; the temperature will change the evaporation rate of the water body, affecting the total amount of water and related hydrological characteristics; changes in air pressure can affect the stress on the surface of the water body and affect the movement of the water flow to a certain extent; humidity will affect the exchange of water vapor content in the air with the water body, indirectly affecting the hydrological parameters; sunshine duration affects the temperature distribution and evaporation process of the water body.

[0068] At the same time, there will be corresponding correlations between hydrological parameters. For example, the height of the water level will affect the flow rate. Generally speaking, when the water level rises, the cross-sectional area of ​​the water will increase, and the flow rate may decrease accordingly when other conditions remain unchanged. The change in flow rate will affect the turbidity of the water body. When the flow rate is faster, the ability of the water flow to carry particles such as sediment is enhanced, which may lead to increased turbidity. Water temperature will not only affect the dissolved oxygen content of the water body. Generally, when the water temperature rises, the solubility of oxygen in water decreases, and the dissolved oxygen content will decrease. It will also have a certain impact on the pH value of the water, because changes in water temperature may change the rate of various chemical reactions in the water, thereby affecting the concentration of hydrogen ions and causing changes in pH. In addition, changes in turbidity will affect the absorption and scattering of light by the water body, thereby indirectly affecting the water temperature, and the nature of suspended particles in the water may also have an effect on the pH value and dissolved oxygen content.

[0069] The values ​​of hydrological parameters other than the target hydrological parameters in the real-time hydrological data are obtained, and the obtained values ​​are input into the prediction model to obtain the target hydrological parameter values. The prediction model is trained based on a large amount of historical data. In the historical data, the values ​​of environmental parameters and various hydrological parameters at different times are recorded in detail.

[0070] Before training the model, the historical data needs to be preprocessed. First, fill in the missing values ​​in the data. You can use the mean, median, or interpolation method based on machine learning algorithms to ensure the integrity of the data. Then, detect and correct outliers to avoid their adverse effects on model training. For example, use the box plot method to identify and replace outliers with reasonable boundary values. Take the environmental parameters and various hydrological parameters in the preprocessed training data as input, and use the target hydrological parameters as output labels to train the prediction model.

[0071] Build prediction models. Build prediction models based on ensemble learning models or deep learning models, such as random forest regression, gradient boosting trees, long short-term memory networks (LSTM), etc.

[0072] In the model training phase, the model is trained using training data. During the training process, the early stopping method is used to prevent the model from overfitting. That is, when the performance indicators of the model on the validation set (such as mean square error or determination coefficient) no longer improve in multiple consecutive training rounds, the training is stopped and the current optimal model is saved. After multiple rounds of iterative training, the model gradually learns the complex internal connections and potential laws between environmental parameters and various hydrological parameters.

[0073] The values ​​of environmental parameters and hydrological parameters other than the target hydrological parameters in real-time hydrological data are input into the trained model to obtain the predicted parameter values ​​of the target hydrological parameters. Taking the random forest regression model as an example, the input data will be input into multiple decision trees at the same time. Each decision tree classifies and predicts the input data according to the rules it has learned. Finally, the random forest model averages the prediction results of all decision trees to obtain the predicted value of the target hydrological parameter.

[0074] After obtaining the predicted parameter value, determine whether the predicted parameter value is within the reasonable parameter value range corresponding to the target hydrological parameter. If so, use the predicted parameter value as the target hydrological parameter value; if not, use the prediction model to predict the parameter value of the target hydrological parameter again.

[0075] S205. Update real-time hydrological data according to target hydrological parameters and target hydrological parameter values.

[0076] Replace the parameter values ​​of the target hydrological parameters in the real-time hydrological data with the target hydrological parameter values.

[0077] S206. Report real-time hydrological data and real-time video data according to a preset data reporting mechanism.

[0078] Report real-time hydrological data according to the preset hydrological data reporting mechanism, and report video data in real time according to the preset video data reporting mechanism.

[0079] Among them, the data reporting mechanism includes hydrological data reporting mechanism and video data reporting mechanism.

[0080] Specifically, according to the hydrological data reporting mechanism, the hydrological data is transmitted to the designated hydrological data server in a specific data format through network communication at fixed time intervals. During the data transmission process, if there is an abnormal situation such as a network failure, the cache function is automatically started to temporarily store the data that has not been successfully transmitted in the local storage device. After the network returns to normal, the cached data is resent first to ensure the integrity and continuity of the data.

[0081] According to the video data reporting mechanism, the video data is transmitted in real time to the designated hydrological data server in a specific data format through network communication. At the same time, in order to ensure the security of the video data, multi-layer encryption technology is used in the transmission process to prevent the data from being stolen or tampered with. If the video data transmission is interrupted, the interruption time and location information are recorded. After the network is restored, the subsequent video clips are uploaded from the breakpoint to achieve stable reporting of real-time video data.

[0082] In an embodiment of the present application, when a sensor failure is detected, the correct hydrological parameter values ​​from the time the sensor failure occurs to the time the staff replaces or repairs the sensor are obtained by performing image recognition on the video frames in the real-time video or by using a prediction model. Then, the abnormal hydrological parameter values ​​in the hydrological data are replaced with the correct hydrological parameter values, thereby reducing the abnormal data transmitted to the control center during this period and improving the accuracy of the hydrological data transmitted to the control center by the RTU.

[0083] Combine the following Figure 3 To further illustrate the method of the embodiment of the present application.

[0084] See also Figure 3 , is another flow chart of the RTU intelligent data analysis method in an embodiment of the present application.

[0085] S301. Collect real-time hydrological data through sensors at preset time intervals.

[0086] S302: Acquire real-time video data through a camera.

[0087] Steps S301, S302 and Figure 2 In the illustrated embodiment, steps S201 and S202 are similar, and the descriptions of steps S201 and S202 may be referred to, and will not be repeated here.

[0088] S303: Obtain the light intensity around the camera.

[0089] The ambient light sensor collects the light intensity around the camera. The ambient light sensor senses the surrounding light through the photosensitive element and converts the light signal into an electrical signal. After being processed by the internal circuit, it outputs the light intensity data in the form of a digital signal.

[0090] S304: When the light intensity exceeds a preset threshold, actual color data of a preset reference object in each video frame in the real-time video data is obtained.

[0091] Among them, the preset reference object is a standard color plate fixed on the shore.

[0092] Specifically, after receiving the real-time video data transmitted by the camera, the remote terminal device decodes and processes it, and restores the encoded video data into a processable video frame image. When the light intensity exceeds the preset threshold, the position of the preset reference object in each video frame is located by the image recognition algorithm. After locating the position of the preset reference object, the color data of all pixels in the area where it is located in each video frame is extracted. The color data is usually represented in the RGB (red, green, blue) color space. The average value or weighted average value of all pixels in the preset reference object area in each video frame in the corresponding color space is calculated to obtain the actual color data of the preset reference object in each video frame.

[0093] In some cases, in order to more conveniently analyze color information, it may be necessary to convert the RGB color space to other color spaces, such as HSV (hue, saturation, brightness) or CIELAB (a device-independent color space). In this case, you can select a suitable color space conversion algorithm for conversion based on specific analysis requirements.

[0094] S305 : Compare the actual color data corresponding to each video frame with the preset color data of the preset reference object in sequence to obtain the color similarity of each video frame.

[0095] Specifically, the actual color data and the preset color data of the preset reference are first converted to the same color space. Then, according to the specific color space, the color similarity of each video frame is calculated by the corresponding similarity calculation method. For example, if the RGB color space is used, the Euclidean distance formula can be used to calculate the distance between the actual color data and the preset color data in the three-dimensional space. The smaller the distance, the higher the color similarity. This distance value is converted into a similarity value through a certain mapping relationship; if the HSV color space is used, the differences in hue, saturation and brightness can be calculated separately, and then these three differences can be combined according to certain weights to obtain an overall similarity value.

[0096] S306: When the color similarity of the first video frame does not exceed a preset threshold, calibrate the color of the corresponding first video frame to obtain a second video frame.

[0097] The first video frame is one of all video frames in the real-time video data. The color similarity of the second video frame exceeds a preset threshold.

[0098] Specifically, according to the specific color space, the color data of each pixel in the first video frame is calculated according to the corresponding calculation formula to obtain the color data of each pixel after calibration. The color data of each pixel in the first video frame is updated to the color data of each pixel after calibration to obtain a second video frame. The color similarity of the preset reference object in the second video frame is calculated to determine whether it exceeds the preset threshold. If it exceeds, repeat the adjustment process; if it does not exceed, complete the calibration operation.

[0099] If the RGB color space is used, the difference between the actual color data and the red, green and blue color channels in the preset color data can be calculated respectively. According to the calculated difference, the color data of each pixel in the first video frame is calculated according to the corresponding calculation formula. If the HSV color space is used, the difference between the hue, saturation and brightness in the actual color data and the preset color data can be calculated respectively. According to the calculated difference, the color data of each pixel in the first video frame is calculated according to the corresponding calculation formula.

[0100] S307: Replace the first video frame in the real-time video data with the second video frame.

[0101] A first video frame in the real-time video data is replaced by a second video frame obtained by calibrating the first video frame.

[0102] S308: When all sensors collecting the target hydrological parameter fail and the target hydrological parameter belongs to the hydrological parameter in the first hydrological parameter set, determine the target hydrological parameter value according to the real-time video data.

[0103] S309. When all sensors collecting target hydrological parameters fail and the target hydrological parameters belong to the hydrological parameters in the second hydrological parameter set, the values ​​of the hydrological parameters other than the target hydrological parameters in the environmental parameters and real-time hydrological data are input into the prediction model to obtain the target hydrological parameter values.

[0104] S310. Update real-time hydrological data according to target hydrological parameters and target hydrological parameter values.

[0105] Steps S308-S310 and Figure 2 Steps S203 - S205 in the illustrated embodiment are similar, and the descriptions of steps S203 - S205 may be referred to, and will not be repeated here.

[0106] S311. Determine whether a target abnormality occurs based on real-time hydrological data and real-time video data.

[0107] If yes, and the target abnormal condition is abnormal water flow, execute step S313; if yes, and the target abnormal condition is aggravated water pollution, execute step S316; if yes, and the target abnormal condition is intrusion by an unfamiliar user, execute step S322; if no, execute step S312.

[0108] Among them, target abnormal situations include abnormal water flow, increased water pollution, and intrusion by unfamiliar users.

[0109] Specifically, for abnormal water flow, the current cross-sectional area of ​​the river is first calculated through the correspondence between the water level and the cross-sectional area of ​​the river, and then the current water flow is obtained by multiplying the current cross-sectional area of ​​the river and the current flow rate. Then, the rainfall is obtained, and the preset threshold of the water flow is adjusted according to the correspondence between the rainfall and the water flow to obtain the current water flow threshold. Finally, it is determined whether the current water flow exceeds the current water flow threshold. If it exceeds, it is determined that the target abnormality occurs, and the target abnormality is the water flow abnormality.

[0110] In case of worsening water pollution, the pollutant detection model is used to identify the number of pollutants or the scope of pollution in the monitoring area based on real-time video data. The pollutant detection model is built based on deep learning algorithms (such as convolutional neural networks, etc.). During the model training process, the model is trained by collecting a large number of labeled pollutant images, so that the model can learn the characteristic performance of different pollutants in the image, including color, shape, texture, etc., and then can quickly and accurately identify various types of pollutants in the monitoring area when real-time video data is input, accurately calculate the number of pollutants, and determine their pollution scope.

[0111] The real-time video data is input into the pollutant detection model, and the model will analyze each frame of the video according to the trained mode. First, the deep learning algorithm such as convolutional neural network is used to extract the feature information in the image and identify different types of pollutants in the monitoring area. For example, different pollutants such as solid waste, oil pollution, and chemical leakage are distinguished based on the previously learned color, shape, texture and other features. If the pollutant is a solid pollutant, the identified pollutants are counted and the number of pollutants is accurately calculated. If the pollutant is not a solid pollutant, the distribution of the pollutant in the image is analyzed to determine its pollution range.

[0112] After obtaining the number of pollutants or the pollution range, determine whether the water pollution has worsened based on the historical flow rate, the historical number of pollutants or the pollution range, the current flow rate, and the current number of pollutants or the pollution range. When the difference between the historical flow rate and the current flow rate does not exceed the preset range, and the current number of pollutants is greater than the historical number of pollutants, or the current pollution range is greater than the historical pollution range, it is determined that the water pollution has worsened; otherwise, a first relationship model is established based on the relationship between the flow rate and the number of pollutants, and a second relationship model is established based on the relationship between the flow rate and the pollutant range. The number of pollutants or the pollution range at the current flow rate is predicted by the relationship model and compared with the current actual number of pollutants or the pollution range. If the deviation between the predicted value and the actual value exceeds a certain threshold, it is determined that the water pollution has worsened.

[0113] The first relationship model obtains a function expression of flow velocity and pollutant quantity by performing linear regression analysis on historical data; the second relationship model obtains a function expression by performing nonlinear fitting on historical data. Substituting the flow velocity into the above function expression, the predicted pollutant quantity or pollution range is obtained.

[0114] In some embodiments, the difference between the turbidity and the preset standard turbidity, the difference between the pH value and the preset standard pH value, and the difference between the dissolved oxygen content and the preset standard dissolved oxygen content can also be determined based on real-time hydrological data, and the current pollution index value can be calculated according to preset weights. The current pollution index value is compared with the historical pollution index value. If the current pollution index value is greater than the historical pollution index value and the difference is greater than the preset threshold, it is determined that the water pollution has worsened.

[0115] For intrusion by unfamiliar users, we first use the target detection algorithm to identify whether there are people in the monitoring area based on real-time video data. We use advanced deep learning-based target detection algorithms to pre-train a model for personnel detection. The model is trained on a large-scale image dataset containing annotated people to learn various characteristic patterns of people in the image.

[0116] The video frames in the pre-real-time video data are input into the model in sequence, and the model extracts features from the image to capture the key feature information in the image. Then, based on these features, target prediction is performed, and the position information of the person target detected in the image in the video frame and the target category (person) are output. Among them, the detected person target is marked with a fixed-shape bounding box, and the position information includes the coordinates of the upper left corner and the lower right corner of the bounding box.

[0117] For the detected person target, extract the facial image from its bounding box area. Use a special face recognition algorithm (such as the FaceNet model based on deep learning, etc.) to feature encode the extracted facial image and generate a facial feature vector. Compare the feature vector with the facial feature vectors in the user information database one by one to calculate the similarity score. If the similarity score is lower than the set threshold, the person is determined to be a stranger. At the same time, it is determined that the target abnormality has occurred and the target abnormality is an intrusion by a stranger.

[0118] In some embodiments, if facial recognition fails, the judgment can be assisted by identifying the behavior patterns of the analyst.

[0119] Based on the observed and recorded behavioral data of authorized personnel's normal activities in the monitoring area collected in advance, the collected behavioral data are analyzed and modeled using machine learning algorithms (such as hidden Markov model (HMM), recurrent neural network (RNN) and its variant long short-term memory network (LSTM), etc.). For example, using HMM, the behavior sequence of personnel can be regarded as a series of hidden states (such as normal walking, short stay, operating equipment, etc.) and observable states (such as position coordinates, speed, etc.), and the conversion probability between hidden states and observable states under the normal behavior mode of authorized personnel can be learned by training the model.

[0120] When a person target is detected, its behavior data in the monitoring area is tracked in real time. Through video analysis technology, the person's real-time position coordinate sequence, movement speed, direction change and other information are obtained, and this information is divided according to a certain time window to form a behavior sequence.

[0121] The real-time behavior sequence is input into the pre-trained behavior pattern model, and the matching score between the behavior sequence and the normal behavior pattern of the authorized person is calculated. For example, in the HMM model, the matching score can be obtained by calculating the likelihood probability of the behavior sequence in the model.

[0122] Based on a preset matching threshold, if the matching score of the real-time behavior sequence is lower than the threshold, it indicates that the behavior pattern of the person is significantly different from the normal behavior pattern of the authorized person, and the person may be an unfamiliar user.

[0123] S312: Report real-time hydrological data according to a preset second hydrological data reporting mechanism.

[0124] Among them, the second hydrological data reporting mechanism includes reporting objects, reporting frequency, etc.

[0125] Specifically, first determine the data receiver according to the reporting object specified in the second hydrological data reporting mechanism. The receiver can be a server or other monitoring terminal, which is not limited here. Then, establish a communication link between itself and the receiver through a network connection. After the connection is established, according to the reporting frequency, at a certain time interval, the data is sent to the receiver through a network protocol.

[0126] S313. When the target abnormality is abnormal water flow, the water flow change within a preset time in the future is predicted by a machine learning model to obtain a prediction result.

[0127] Specifically, the current water flow and current environmental parameters are input into the machine learning model, and the model predicts the change of water flow within a preset time in the future to obtain the prediction result. The machine learning model is built based on machine learning algorithms (such as linear regression algorithm, support vector machine regression algorithm, neural network algorithm, etc.). Before building the model, a large amount of historical water flow data and corresponding historical environmental parameter data are required. These environmental parameters cover factors that may affect water flow, such as temperature, rainfall, wind speed, and air pressure. The model is trained using the above data, and by continuously adjusting the parameters of the model, such as the coefficients in the linear regression algorithm, the kernel function parameters in the support vector machine regression algorithm, and the weights and biases in the neural network algorithm, the model can learn the mapping relationship between historical water flow and historical environmental parameters.

[0128] The current water flow and current environmental parameters are input into the machine learning model, and the model operates on the input current data according to the mapping relationship between historical water flow and historical environmental parameters learned during the training process. Taking the model built by the linear regression algorithm as an example, the model will perform a weighted summation of the current water flow and current environmental parameters based on the coefficients obtained in training, and add a constant term to obtain the predicted water flow value within a preset future time.

[0129] S314. Use the prediction results and real-time hydrological data as target reporting data.

[0130] According to the preset data format, the prediction results and real-time hydrological data are integrated into a whole data, and the data is used as the target reporting data.

[0131] S315: Add the monitoring site closest to the current monitoring area to the preset reporting objects to obtain the target reporting object.

[0132] According to the preset location information of each monitoring site and the location information of the current monitoring area, the distance between each monitoring site and the current monitoring area is calculated, and the corresponding monitoring site with the closest distance is screened out. If the monitoring site is not in the preset reporting object, the monitoring site is added to the preset reporting object to obtain the target reporting object.

[0133] The preset reporting objects include the main control center, the monitoring site that manages the monitoring area, etc. The main control center is not in the monitoring site, and the monitoring site manages one or more monitoring areas. The location information includes specific location coordinate information.

[0134] S316. When the target abnormal condition is aggravated water pollution and the water flow velocity is greater than a preset threshold, historical video data of other remote terminal devices in a direction opposite to the water flow direction are obtained to obtain a historical video data set.

[0135] Specifically, when the target abnormal condition is that water pollution is aggravated and the water flow velocity is greater than a preset threshold, the water flow direction is determined. The water flow direction can be obtained through a water flow direction sensor, and the water flow direction can also be determined according to the movement direction of pollutants in real-time video data.

[0136] The historical video data of other remote terminal devices in the direction opposite to the water flow direction are obtained to obtain a historical video data set. According to the location information of the monitoring area and the location information of other monitoring areas, other monitoring areas located in the direction opposite to the water flow direction are screened out. Communication is established with remote terminal devices in other monitoring areas to obtain the historical video data of other remote terminal devices in the time range from the time point before the current time point to the time point of the preset time interval of the current time interval to the current time point, and obtain a historical video data set, wherein each historical video data in the historical video data set corresponds to the monitoring area one by one, and the sorting order of the historical video data is sorted from near to far according to the distance from the monitoring area.

[0137] S317. Determine the emission area where the pollution source is located based on the historical video data set.

[0138] Specifically, the historical video data corresponding to the first monitoring area that is closest to the monitoring area among other monitoring areas is first input into the pollutant detection model, and the pollutants in the video are identified and analyzed using the model to obtain pollutant information in the first monitoring area, including the type, quantity or pollution range of the pollutants. At the same time, the movement direction of the pollutants in the first monitoring area is determined according to the water flow direction of the first monitoring area, or the movement direction of the pollutants is identified according to the real-time video data of the first monitoring area.

[0139] Then, the pollutant type in the first monitoring area is compared with the pollutant type in the monitoring area to determine whether the same type of pollutant exists. If the same type of pollutant exists, the second monitoring area closest to the first monitoring area in the direction opposite to the direction of movement of the pollutant is determined according to the movement direction of the pollutant, and the historical video data corresponding to the second monitoring area in the historical video data set is input into the pollutant detection model to obtain the pollutant information in the second monitoring area. At the same time, the movement direction of the pollutant in the second monitoring area is determined.

[0140] Then compare the pollutant type in the second monitoring area with the pollutant type in the monitoring area to determine whether there are pollutants of the same type. If there are pollutants of the same type, determine the third monitoring area closest to the second monitoring area in the direction opposite to the direction of movement of the pollutants according to the direction of movement of the pollutants, and input the historical video data corresponding to the third monitoring area in the historical video data set into the pollutant detection model to obtain the pollutant information in the third monitoring area. Repeat the above operation until there is a target monitoring area in other detection areas, and there is no pollutant type in the target monitoring area that is the same as the pollutant type in the monitoring area, and there is no pollutant type in the next compared monitoring area of ​​the target monitoring area that is the same as the pollutant type in the monitoring area. At this time, it is determined that the last compared monitoring area of ​​the target area is the emission area where the pollution source is located.

[0141] S318. When there is no emission area in the historical emission area, it is determined that the cause of the increased pollution is the newly added emission area.

[0142] Check whether there is an emission area where the pollution source is located in the preset historical emission area table. When there is no emission area in the historical emission area, determine that the cause of the increased pollution is the newly added emission area, and store the emission area where the pollution source is located in the historical emission area table.

[0143] S319. When there is an emission area in the historical emission area, it is determined that the cause of the increased pollution is the increase in the amount of pollutant emissions in the emission area.

[0144] Find out whether there is an emission area where the pollution source is located in the preset historical emission area table. When there is an emission area in the historical emission area, determine that the cause of the increased pollution is the increase in pollutant emissions in the emission area.

[0145] S320. The emission area, the cause of the increased pollution and the real-time hydrological data are used as target reporting data.

[0146] According to the preset data format, the emission area, the cause of pollution aggravation and the real-time hydrological data are integrated into a whole data, and this data is used as the target reporting data.

[0147] S321. Add the monitoring site closest to the emission area to the preset reporting objects to obtain the target reporting object.

[0148] Step S321 is similar to the above-mentioned step S315. Please refer to the description in step S315, which will not be repeated here.

[0149] S322: When the target abnormal situation is an intrusion by a stranger user, determine the location of the stranger user based on the real-time video data.

[0150] Specifically, based on the position information of the person target in the video frame of the real-time video data obtained in step S311, the coordinate range of the person target in the video frame is determined. Then, based on the coordinate range, the position description of the position information in the preset position comparison table (such as the first area by the river, etc.) and the real position in the actual space are obtained. Then, the position description and the real position are integrated into a unified information structure as the location of the stranger user.

[0151] The video frame of real-time video data is divided into multiple small areas, each small area corresponds to a bounding box, and the position comparison table stores the coordinate ranges of the bounding boxes corresponding to the multiple small areas in the video frame. Each coordinate range corresponds to a position description and a real position in the actual space.

[0152] S323: The location of the unknown user and the real-time hydrological data are used as target reporting data.

[0153] According to the preset data format, the location of the unfamiliar user and the real-time hydrological data are integrated into a whole data, and the data is used as the target reporting data.

[0154] S324: Add the monitoring site closest to the current monitoring area to the preset reporting objects to obtain the target reporting object.

[0155] Step S324 is similar to the above-mentioned step S315. Please refer to the description in step S315, which will not be repeated here.

[0156] S325. Adjust the reporting objects in the preset first hydrological data reporting mechanism and the preset video data reporting mechanism according to the target reporting object.

[0157] The reporting objects in the preset first hydrological data reporting mechanism and the preset video data reporting mechanism are replaced with the target reporting objects.

[0158] The first hydrological data reporting mechanism and the second hydrological data reporting mechanism have different reporting frequencies and reporting objects. The video data reporting mechanism includes reporting objects.

[0159] S326. Report the target reporting data according to the adjusted first hydrological data reporting mechanism.

[0160] Step S326 is similar to the above-mentioned step S312. Please refer to the description in step S312, which will not be repeated here.

[0161] S327. Report the real-time video data in real time according to the adjusted video data reporting mechanism.

[0162] Specifically, the data receiver is first determined according to the reporting object specified in the video data reporting mechanism. Then, a communication link is established between the receiver and itself through a network connection. After the connection is established,

[0163] The video data is transmitted in real time to the data receiver in a specific data format through network communication.

[0164] S328. Evaluate the target processing difficulty and target processing time of the target abnormal situation based on the target abnormal situation and historical abnormal processing data.

[0165] Among them, historical anomaly processing data includes specific environmental parameters when the anomaly occurs, specific description of the anomaly (such as the specific value of the water flow exceeding the threshold, the number of pollutants in the water pollution or the size of the pollution range, etc.), processing difficulty and processing time, etc. The target processing difficulty is divided into high, medium and low.

[0166] Specifically, first, according to the target abnormal situation, target historical abnormality processing data with the same abnormal situation is screened out from the historical abnormality processing data.

[0167] If the target abnormal situation is an intrusion by a stranger user, the target processing difficulty of determining the target abnormal situation is high, and the target processing time is the average of all processing times in the target historical abnormal processing data.

[0168] If the target abnormal situation is not caused by an intrusion by an unfamiliar user, the target abnormal situation is matched with each record in the filtered target historical abnormal processing data based on the current environmental parameters, the specific description of the current abnormality, etc. The similarity between the target abnormal situation and each historical abnormal record is evaluated by calculating the difference or similarity coefficient of each parameter.

[0169] When there is a similarity exceeding a preset threshold, the processing difficulty and processing time in the record with the highest similarity are selected as the target processing difficulty and target processing time.

[0170] In the case where there is no similarity exceeding the preset threshold, the preset processing difficulty and processing time are used as the target processing difficulty and target processing time.

[0171] S329: Report the target processing difficulty and target processing time to the target reporting object.

[0172] Specifically, first, the target processing difficulty and target processing time are integrated into a whole data according to the preset data format. Then, the data receiver is determined according to the target reporting object. Then, a communication link is established between itself and the receiver through a network connection. After the connection is established, the data is sent to the receiver through a network protocol.

[0173] S330: Determine a target abnormality level of the target abnormality according to the type of the target abnormality, the target processing difficulty, and the target processing time.

[0174] Among them, the target anomaly levels are divided into level one, level two and level three.

[0175] Specifically, if the target abnormal situation is an intrusion by an unfamiliar user, the target abnormality level is determined to be level three.

[0176] If the target abnormality is not an intrusion by an unfamiliar user, determine the target abnormality level based on the target processing difficulty and the target processing time. If the target processing difficulty is high, determine the target abnormality level to be level three. If the target processing difficulty is medium, determine whether the target processing time exceeds the threshold of the processing time when the target processing difficulty is medium. If it exceeds, determine the target abnormality level to be level three; if it does not exceed, determine the target abnormality level to be level two. If the target processing difficulty is low, determine whether the target processing time exceeds the threshold of the processing time when the target processing difficulty is low. If it exceeds, determine the target abnormality level to be level two; if it does not exceed, determine the target abnormality level to be level one.

[0177] S331. When the target abnormality level is level 2 or level 3, execute the corresponding alarm mechanism.

[0178] If the target abnormality level is level 2, a warning text message is sent to the first emergency contact person through the communication module; if the target abnormality level is level 3, a buzzer alarm is triggered or the corresponding alarm information is played through the speaker, and a warning text message is sent to the second emergency contact person through the communication module.

[0179] Among them, the first emergency contact personnel include security supervisors, regional managers, on-duty personnel, etc. The second emergency contact personnel include security supervisors, regional managers, on-duty personnel, technical support specialists, etc.

[0180] In the embodiment of the present application, when a sensor failure is detected, the correct hydrological parameter value during the period from the sensor failure to the replacement or repair of the sensor by the staff is obtained by performing image recognition on the video frames in the real-time video or by predicting the prediction model, and then the abnormal hydrological parameter value in the hydrological data transmitted to the control center is replaced with the correct hydrological parameter value, thereby reducing the abnormal data transmitted to the control center during this period and improving the accuracy of the hydrological data transmitted by the RTU to the control center. When an abnormal situation is detected, the abnormal situation is actively reported, and the relevant staff is notified in time to pay attention to monitoring and handling the abnormal situation, thereby improving the response speed to the hydrological abnormal situation. At the same time, by calibrating the color data of each video frame in the video data, the color accuracy of the real-time video data is ensured, and the reliability of the recognition result when the image is recognized is improved, thereby improving the accuracy of the hydrological data obtained by image recognition.

[0181] The above describes the RTU intelligent data analysis method in the embodiment of the present application. The following describes the remote terminal device in the embodiment of the present application in detail in combination with the above RTU intelligent data analysis method.

[0182] See also Figure 4 , which is a schematic diagram of an exemplary hardware structure of a remote terminal device in an embodiment of the present application.

[0183] In some embodiments, the remote terminal device 400 includes a computer device, which can be a terminal device. The computer device includes a processor 401, a memory 402, a sensor module 403, a communication module 404, an input device 405, and an output device 406 connected through a system bus. Among them, the processor 401 of the computer device is used to provide computing and control capabilities. The memory 402 of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database is used to store data. The sensor module 403 of the computer device is used to collect environmental parameters, hydrological data, etc. The communication module 404 of the computer device is used to report data to the reporting object and send text messages to contact personnel, etc. The input device 405 of the computer device is used to receive control instructions sent by staff, etc. The output device 406 of the computer device is used to play alarm information, etc. When the computer program is executed by the processor 401, the RTU intelligent data analysis method in the embodiment of the present application is implemented.

[0184] Those skilled in the art will understand that Figure 4The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0185] In some embodiments of the present application, a computer-readable storage medium is provided, including instructions, which, when executed on the remote terminal device 400, can enable the remote terminal device 400 to execute the RTU intelligent data analysis method in the embodiment of the present application.

[0186] In some embodiments of the present application, a computer program product is further provided. When the computer program product is run on a remote terminal device 400, the remote terminal device 400 executes the RTU intelligent data analysis method in the embodiments of the present application.

[0187] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0188] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0189] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0190] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A RTU intelligent data analysis method, characterized in that: include: Collect real-time hydrological data through sensors at preset time intervals; The real-time hydrological data include water level, flow rate, water temperature, turbidity, pH value and dissolved oxygen content; Obtain real-time video data through the camera; In the case where all sensors for collecting target hydrological parameters fail and the target hydrological parameter belongs to a hydrological parameter in a first hydrological parameter set, determining a target hydrological parameter value according to the real-time video data; the target hydrological parameter is one of all hydrological parameters; the first hydrological parameter set is a set of hydrological parameters whose hydrological parameter values ​​are obtained according to an image or video; In the case that all sensors collecting the target hydrological parameter fail and the target hydrological parameter belongs to the hydrological parameter in the second hydrological parameter set, the values ​​of the environmental parameters and the hydrological parameters other than the target hydrological parameters in the real-time hydrological data are input into the prediction model to obtain the value of the target hydrological parameter; the second hydrological parameter set is a set of other hydrological parameters in the real-time hydrological data except the hydrological parameters in the first hydrological parameter set; updating the real-time hydrological data according to the target hydrological parameter and the target hydrological parameter value; Determine whether a target abnormality occurs according to the real-time hydrological data and the real-time video data; the target abnormality includes abnormal water flow, aggravated water pollution, and intrusion by unfamiliar users; If yes, determine the target reporting object and target reporting data according to the target abnormality; According to the target reporting object, adjusting the reporting objects in the preset first hydrological data reporting mechanism and the preset video data reporting mechanism; Reporting the target reporting data according to the adjusted first hydrological data reporting mechanism; The real-time video data is reported in real time according to the adjusted video data reporting mechanism.

2. The method according to claim 1, characterized in that After the step of determining whether a target abnormality occurs according to the real-time hydrological data and the real-time video data, the method further comprises: If not, the real-time hydrological data is reported according to a preset second hydrological data reporting mechanism.

3. The method according to claim 1, characterized in that If so, according to the target abnormality, determine the target reporting object and target reporting data, specifically including: When the target abnormality is water flow abnormality, the water flow change within a preset time in the future is predicted by a machine learning model to obtain a prediction result; the water flow abnormality refers to the current water flow exceeding the water flow threshold; the machine learning model is constructed based on a machine learning algorithm, and is trained using a large amount of historical water flow data and corresponding historical environmental parameter data to construct the model; Using the prediction result and the real-time hydrological data as target reporting data; When the target abnormal situation is aggravated water pollution, determine the emission area where the pollution source is located and the cause of the aggravated pollution; the aggravated water pollution means that the amount of pollutants or the pollution range in the current detection area is greater than the amount of pollutants or the pollution range in the past; The emission area, the cause of the aggravated pollution and the real-time hydrological data are used as the target reporting data; When the target abnormal situation is an intrusion by a stranger user, the location of the stranger user is determined based on real-time video data; Using the location of the unknown user and the real-time hydrological data as the target reporting data; When the target abnormality is abnormal water flow or intrusion by an unfamiliar user, the monitoring site closest to the current monitoring area is added to the preset reporting object to obtain the target reporting object; When the target abnormal situation is aggravated water pollution, the monitoring site closest to the discharge area is added to the preset reporting objects to obtain the target reporting object.

4. The method according to claim 3, characterized in that When the target abnormal situation is aggravated water pollution, determining the emission area where the pollution source is located and the cause of the aggravated pollution specifically includes: When the target abnormal condition is aggravated water pollution and the water flow velocity is greater than a preset threshold, historical video data of other remote terminal devices in a direction opposite to the water flow direction are obtained to obtain a historical video data set; Determine the emission area where the pollution source is located based on the historical video data set; When the emission area does not exist in the historical emission area, the cause of the increased pollution is determined to be the newly added emission area; When the emission area exists in the historical emission area, it is determined that the cause of the pollution aggravation is an increase in the pollutant emission amount of the emission area.

5. The method according to claim 2, characterized in that: In the case where, after the step of determining the target reporting object and the target reporting data according to the target abnormality, the method further includes: According to the target abnormal situation and historical abnormality processing data, evaluating the target processing difficulty and target processing time of the target abnormal situation; The target processing difficulty and the target processing time are reported to the target reporting object.

6. The method according to claim 5, characterized in that After the step of evaluating the target processing difficulty and target processing time of the target abnormal situation according to the target abnormal situation and the historical abnormal processing data, the method further includes: Determine a target abnormality level of the target abnormality according to the type of the target abnormality, the target processing difficulty and the target processing time; the target abnormality level includes level one, level two and level three; When the target abnormality level is level 2 or level 3, a corresponding alarm mechanism is executed.

7. The method according to claim 1, characterized in that After the step of acquiring real-time video data through the camera, the method further includes: Obtaining the light intensity around the camera; When the light intensity exceeds a preset threshold, actual color data of a preset reference object in each video frame in the real-time video data is obtained; the preset reference object is a standard color plate fixed on the shore; Comparing the actual color data corresponding to each video frame with the preset color data of the preset reference object in sequence to obtain the color similarity of each video frame; When the color similarity of the first video frame does not exceed a preset threshold, calibrate the color of the corresponding first video frame to obtain a second video frame; the first video frame is one of all video frames in the real-time video data; and the color similarity of the second video frame exceeds a preset threshold; The first video frame in the real-time video data is replaced by the second video frame.

8. A remote terminal device, characterized in that: It includes a sensor, a camera and a data processing server, wherein the data processing server includes: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the remote terminal device to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on the remote terminal device, the remote terminal device is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is executed on a remote terminal device, the remote terminal device is caused to execute the method according to any one of claims 1 to 7.

Images