Deep learning-based environmental disaster monitoring method, system, and device

Abstract

The present application relates to the technical field of environmental disaster monitoring and evaluation, and in particular to an environmental disaster monitoring method, system and device based on deep learning, which comprises the following steps: acquiring environmental monitoring image data; pre-processing the environmental monitoring image data to obtain pre-processed environmental monitoring images; identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result; and analyzing and evaluating risks based on the identification result and historical meteorological data. Thus, the present application can greatly improve the efficiency and accuracy of environmental disaster monitoring and evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental disaster monitoring and evaluation, and in particular to an environmental disaster monitoring method, system and device based on deep learning. BACKGROUND

[0002] In industries such as the power industry and other high-risk industries, disasters such as icing phenomena pose a threat to the safety of equipment and infrastructure. Traditional disaster monitoring techniques usually rely on sensors or manual patrols and cannot grasp disaster situations in real time and comprehensively.

[0003] For example, Chinese patent CN118629159A discloses an icing distribution pole tower broken pole early warning method and system, which collects temperature, humidity, wind speed and inclination angle data using sensors, and then analyzes and judges the data to confirm the danger and issue a warning. Although this method can achieve early warning of danger, it can only judge the icing distribution pole tower broken pole and cannot achieve real-time early warning. For another example, Chinese patent CN115278578A discloses a digital twin system for disaster prevention and mitigation in a substation, which monitors natural monitoring data related to damage to the substation by natural disasters through a wireless monitoring sensor network, determines the disaster type through correlation analysis, and analyzes the damage to the substation by natural disasters based on natural monitoring data and constructed fulcrum string models, rigid body models, etc. through a substation disaster physical analysis module. Although this method can improve the efficiency of early warning to some extent, it still cannot achieve real-time early warning and cannot comprehensively grasp the actual disaster situation. Moreover, the above-mentioned solutions all need to arrange various sensors in advance, which is complex to operate and implement. Therefore, there is an urgent need for a method that is easy to implement, can achieve real-time early warning, and can help staff to comprehensively grasp the disaster situation. SUMMARY

[0004] Therefore, the purpose of the present application is to provide an environmental disaster monitoring method, system and device based on deep learning to overcome the low efficiency and poor accuracy problems in the current disaster monitoring field.

[0005] To achieve the above purpose, the present application adopts the following technical solutions:

[0006] In a first aspect, the present application provides an environmental disaster monitoring method based on deep learning, comprising:

[0007] obtaining environmental monitoring image data;

[0008] preprocessing the environmental monitoring image data to obtain preprocessed environmental monitoring images;

[0009] identifying an environmental disaster site in the environmental monitoring image through a deep learning model to obtain an identification result;

[0010] Based on the identification result and historical meteorological data, a risk is analyzed and evaluated.

[0011] Further, in some embodiments of the present application, the environment monitoring image data is obtained, including:

[0012] The environment monitoring image data is obtained from a preset environment monitoring system, and the environment monitoring image data is obtained by one or more of a satellite remote sensing device, a drone or a monitoring camera.

[0013] Further, in some embodiments of the present application, the environment monitoring image data is preprocessed to obtain preprocessed environment monitoring image, including:

[0014] The environment monitoring image data is sequentially subjected to denoising, image enhancement and size standardization processing;

[0015] Wherein, the denoising processing includes denoising processing by Gaussian filtering and mean filtering; the image enhancement processing includes using histogram equalization technology to improve the contrast of image data; the size standardization processing includes scaling the image data to the input size required by the preset deep learning model.

[0016] Further, in some embodiments of the present application, the environment disaster site in the environment monitoring image is identified by the deep learning model to obtain an identification result, including:

[0017] The environment monitoring image is subjected to semantic segmentation by a deep learning model based on PSPNet to identify the disaster area in the environment monitoring image.

[0018] Wherein, the deep learning model based on PSPNet is generated by a transfer learning strategy and pre-trained model parameters.

[0019] Further, in some embodiments of the present application, the risk is analyzed and evaluated based on the identification result and historical meteorological data, including:

[0020] Based on the historical meteorological data and the identification result of the deep learning model, a rule-based target learning model is used to analyze and evaluate potential risks.

[0021] Further, in some embodiments of the present application, the risk is analyzed and evaluated based on the identification result and historical meteorological data, further including:

[0022] Based on the identification result, historical disaster records and equipment importance, a comprehensive risk score is generated.

[0023] Further, in some embodiments of the present application, the analyzing and evaluating the risk based on the identification result and the historical meteorological data further comprises: issuing a warning based on a result of the evaluating the risk and the comprehensive risk score.

[0024] In a second aspect, the embodiments of the present application provide a deep learning-based environmental disaster monitoring system, comprising:

[0025] an image acquisition module, configured to acquire environmental monitoring image data;

[0026] an image preprocessing module, configured to preprocess the environmental monitoring image data to obtain preprocessed environmental monitoring images;

[0027] a deep learning model analysis module, configured to identify an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result;

[0028] a risk evaluation module, configured to analyze and evaluate a risk based on the identification result and historical meteorological data.

[0029] Further, in some embodiments of the present application, a warning module is further included, and the warning module is configured to issue a warning based on a result of the risk evaluation module.

[0030] In a third aspect, the embodiments of the present application provide a deep learning-based environmental disaster monitoring device, comprising a processor and a memory, wherein the processor is connected to the memory:

[0031] wherein the processor is configured to call and execute a program stored in the memory;

[0032] the memory is configured to store the program, and the program is at least used to execute the deep learning-based environmental disaster monitoring method described above.

[0033] The present application relates to the technical field of environmental disaster monitoring and assessment, and particularly relates to an environmental disaster monitoring method, system and device based on deep learning, the method comprising: obtaining environmental monitoring image data; pre-processing the environmental monitoring image data to obtain pre-processed environmental monitoring images; identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result; and analyzing and evaluating risks based on the identification result and historical meteorological data. Thus, the present application can greatly improve the efficiency and accuracy of environmental disaster monitoring and assessment. The beneficial effects are as follows: 1. Real-time monitoring can be achieved through the deep learning model, improving the monitoring efficiency. 2. The environmental monitoring images can be obtained from the existing environmental monitoring system without additional hardware costs, and the method is easy to operate and has strong implementation. 3. Disaster monitoring and early warning of various infrastructure vulnerable to disasters such as power lines and transportation facilities can be achieved by adjusting the environmental monitoring image data. 4. The accuracy of monitoring and evaluation can be greatly improved through the pre-processing of the environmental monitoring image data and the use of the deep learning model. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0035] Figure 1 is a flowchart of the environmental disaster monitoring method based on deep learning provided by the present application.

[0036] Figure 2 is a structure diagram of the PPM of the deep learning model in the environmental disaster monitoring method based on deep learning provided by the present application.

[0037] Figure 3 is a structure diagram of the environmental disaster monitoring system based on deep learning provided by the present application.

[0038] Figure 4 is a structure diagram of the environmental disaster monitoring device based on deep learning provided by the present application. DETAILED DESCRIPTION

[0039] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described in detail below. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the scope of protection of the present application.

[0040] Figure 1 is a flowchart of the environment disaster monitoring method based on deep learning provided by the embodiments of the present application, as shown in Figure 1 The environment disaster monitoring method based on deep learning provided by the embodiments of the present application can at least include the following steps:

[0041] S101, obtaining environment monitoring image data.

[0042] Specifically, in the present application, the environment monitoring image data can be directly obtained from the existing environment monitoring system.

[0043] S102, pre-processing the environment monitoring image data to obtain pre-processed environment monitoring image.

[0044] Specifically, it can include sequentially performing denoising, image enhancement and size standardization processing on the obtained environment monitoring image data.

[0045] Among them, the denoising processing includes denoising processing by Gaussian filtering and mean filtering; the image enhancement processing includes using histogram equalization technology to improve the contrast of image data; the size standardization processing includes scaling the image data to the input size required by the pre-set deep learning model, so as to obtain the pre-processed environment monitoring image.

[0046] It should be noted that by using denoising techniques such as Gaussian filtering and mean filtering to denoise the obtained environment monitoring image data, the noise interference in the image data can be reduced, and the image quality can be improved. By using image enhancement techniques such as histogram equalization, the contrast of the image data can be improved, and the features of the icing area in the image data can be made more obvious. By scaling the above-processed image data to the input size required by the deep learning model (for example: 256x256 pixels), the input consistency of the deep learning model can be ensured, and the model recognition accuracy can be improved.

[0047] S103, identifying the environment disaster scene in the environment monitoring image by the deep learning model to obtain the identification result.

[0048] Specifically, the environment monitoring image obtained after the above pre-processing is input into the deep learning model, and the deep learning model is used to identify the environment disaster scene including icing to obtain the identification result (such as whether there is icing and the severity, etc.).

[0049] S104, based on the recognition result and historical meteorological data, analyzing and evaluating the risk.

[0050] Specifically, after obtaining the recognition result of the deep learning model, the potential risk brought by the disaster is analyzed and evaluated in combination with the historical meteorological data.

[0051] The environmental disaster monitoring method based on deep learning provided by the application can greatly improve the efficiency and accuracy of environmental disaster monitoring and evaluation.

[0052] Further, in the present application, the environmental monitoring image data can be obtained from the existing environmental monitoring system, such as the image data from satellite remote sensing, unmanned aerial vehicle aerial photography, monitoring camera and other sources, which can cover power lines, traffic facilities and other infrastructure vulnerable to disasters.

[0053] In addition, the obtained historical environmental monitoring image data can also be applied to the training process of the deep learning model, and for this purpose, the environmental monitoring image data is obtained from the existing image database or historical image collection system, wherein the types of data include image data taken under different weather conditions, so as to ensure that the subsequent icing phenomenon can be effectively identified under various environments.

[0054] In the present application, for the deep learning model, the present application first selects a PSPNet(Pyramid Scene Parsing Network) basic deep learning model, which performs well in image segmentation and recognition tasks, especially suitable for environmental monitoring in complex background. On this basis, a large number of labeled icing image data sets are used for model training to ensure that the final deep learning model has good generalization ability. In the training process of the deep learning model, the scheme of the present application includes using the pre-trained model parameters for application, and the recognition accuracy of the model is further improved through targeted training. The PPM structure diagram of the deep learning model is as shown in Figure 2 The principle is consistent with the basic principle of the deep learning model in the prior art, and can be understood by referring to the deep learning model in the prior art, which will not be described in detail here.

[0055] After obtaining the deep learning model through early training, the environmental monitoring image data obtained during actual monitoring is preprocessed and input into the deep learning model. The deep learning model can perform semantic segmentation on the image and identify the icing area to obtain the recognition result.

[0056] After obtaining the recognition result, the potential risks are analyzed and evaluated in combination with historical meteorological data by using a rule-based algorithm or a machine learning model (such as a decision tree or a random forest) to analyze and evaluate the risks (i.e., analyze the potential risks that may be caused by icing) and provide a warning based on the evaluation result, such as using the evaluation result in a warning system or module to provide a reference to relevant decision makers.

[0057] In addition, after obtaining the above-mentioned recognition result, a comprehensive risk score can be generated based on various factors such as the recognition result of the icing area obtained based on the above-mentioned deep learning model, historical disaster records, and equipment importance (in some embodiments, the result obtained by analyzing and evaluating the potential risks by using the above-mentioned rule-based algorithm or machine learning model can also be included) to assist decision makers in making response measures.

[0058] Based on the same inventive concept, the present application also provides a deep learning-based environmental disaster monitoring system, Figure 3 is a structural schematic diagram of the deep learning-based environmental disaster monitoring system provided by the embodiments of the present application, as Figure 3 shown, the deep learning-based environmental disaster monitoring system provided by the present application comprises:

[0059] an image acquisition module 21 for acquiring environmental monitoring image data; an image preprocessing module 22 for preprocessing the environmental monitoring image data to obtain preprocessed environmental monitoring images; a deep learning model analysis module 23 for identifying an environmental disaster site in the environmental monitoring images by using a deep learning model to obtain a recognition result; wherein identifying the environmental disaster site in the environmental monitoring images comprises identifying icing in the environmental monitoring images; and a risk evaluation module 24 for analyzing and evaluating risks based on the recognition result and historical meteorological data.

[0060] Further, in some embodiments of the present application, the environmental disaster monitoring system further comprises a warning module for providing a warning based on the evaluation result of the risk evaluation module.

[0061] It should be noted that the deep learning-based environmental disaster monitoring system provided by the present application adopts a modular design, and the above-mentioned various modules can transmit data through an API interface to ensure the flexibility and scalability of the system. Meanwhile, the deep learning-based environmental disaster monitoring system also provides a user interface for providing a friendly user interface, and users can view real-time monitoring data, recognition results, and risk evaluation reports through the interface. The system can be applied to multiple fields such as the power industry, transportation, and meteorological monitoring, and provides decision support for relevant units through real-time monitoring and evaluation.

[0062] In addition, as described in the above method embodiments, the deep learning-based environmental disaster monitoring system provided in this application can utilize existing monitoring equipment and image data without additional hardware costs. At the same time, the environmental disaster monitoring system can also be integrated into existing environmental monitoring systems for disaster early warning and real-time monitoring in the power industry, transportation facilities and other high-risk scenarios, making it highly implementable.

[0063] Regarding the deep learning-based environmental disaster monitoring system in the above embodiments, the specific methods by which each module performs its operations have been described in detail in the embodiments of the deep learning-based environmental disaster monitoring method, and will not be elaborated here.

[0064] Based on the same inventive concept, this application also provides an environmental disaster monitoring device based on deep learning. Figure 4 This is a schematic diagram of the structure of the deep learning-based environmental disaster monitoring device provided in the embodiments of this application, as shown below. Figure 4 As shown, the device may include a processor 31 and a memory 32, with the processor 31 connected to the memory 32. The processor 31 is used to call and execute a program stored in the memory 32. The memory 32 is used to store a program, which is at least used to execute the aforementioned deep learning-based environmental disaster monitoring method.

[0065] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0066] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.

[0067] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0068] It should be understood that each part of the present application can be realized by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, it can be realized by any one or a combination of the following technologies known in the art: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0069] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by a program instructing the relevant hardware, and the program can be stored in a computer readable storage medium, and when executed, includes one or a combination of steps of the embodiment method.

[0070] In addition, each functional unit in each embodiment of the present application can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of software functional module. The integrated module, if realized in the form of software functional module and sold or used as an independent product, can also be stored in a computer readable storage medium.

[0071] The above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

[0072] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above-mentioned terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0073] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1.A deep learning-based environmental disaster monitoring method, characterized by, The method comprises the following steps: acquiring environmental monitoring image data; preprocessing the environmental monitoring image data to obtain preprocessed environmental monitoring images; identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result; analyzing and evaluating risks based on the identification result and historical meteorological data; wherein the step of identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result comprises the following steps: performing semantic segmentation on the environmental monitoring images through a deep learning model based on PSPNet to identify disaster areas in the environmental monitoring images; and the deep learning model based on PSPNet is generated through a transfer learning strategy and pre-trained model parameters; the step of analyzing and evaluating risks based on the identification result and historical meteorological data comprises the following steps: based on historical meteorological data and the identification result of the deep learning model, a rule-based target learning model is used to analyze and evaluate potential risks; and based on the identification result, historical disaster records and equipment importance, a comprehensive risk score is generated. 2.The deep learning-based environment disaster monitoring method of claim 1, wherein, The step of acquiring environmental monitoring image data comprises the following steps: acquiring the environmental monitoring image data from a preset environmental monitoring system, wherein the environmental monitoring image data is acquired through one or more of satellite remote sensing equipment, a drone or a monitoring camera. 3.The deep learning-based environment disaster monitoring method of claim 1, wherein, The step of preprocessing the environmental monitoring image data to obtain preprocessed environmental monitoring images comprises the following steps: sequentially performing denoising, image enhancement and size standardization processing on the environmental monitoring image data; wherein the denoising processing comprises denoising processing through Gaussian filtering and mean filtering; the image enhancement processing comprises using histogram equalization technology to improve the contrast of image data; and the size standardization processing comprises scaling the image data to the input size required by a preset deep learning model. 4.The deep learning-based environment disaster monitoring method of claim 1, wherein, The step of analyzing and evaluating risks based on the identification result and historical meteorological data further comprises the following steps: based on the result of the risk evaluation and / or the comprehensive risk score, issuing a warning. 5.A deep learning-based environmental disaster monitoring system, characterized by, The method comprises the following steps: an image acquisition module for acquiring environmental monitoring image data; an image preprocessing module for preprocessing the environmental monitoring image data to obtain preprocessed environmental monitoring images; a deep learning model analysis module for identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result; a risk evaluation module for analyzing and evaluating risks based on the identification result and historical meteorological data; wherein the step of identifying an environmental disaster site in the environmental monitoring images through a deep learning model to obtain an identification result comprises the following steps: performing semantic segmentation on the environmental monitoring images through a deep learning model based on PSPNet to identify disaster areas in the environmental monitoring images; and the deep learning model based on PSPNet is generated through a transfer learning strategy and pre-trained model parameters; The risk analysis and evaluation is based on the identification result and historical meteorological data, including: based on the historical meteorological data and the identification result of the deep learning model, using a rule-based target learning model to analyze and evaluate potential risks; based on the identification result, historical disaster records and equipment importance, generating a comprehensive risk score. 6.The deep learning based environment disaster monitoring system according to claim 5, wherein, Further comprising a warning module, the warning module is used for warning based on the evaluation result of the risk evaluation module. 7.A deep learning-based environment disaster monitoring device, characterized by, A processor and a memory are included, the processor is connected with the memory: Wherein, the processor is used for calling and executing the program stored in the memory; The memory is used for storing the program, and the program is at least used for executing the deep learning-based environmental disaster monitoring method in any one of claims 1-4.

Images