Ecological environment protection supervision method, system, equipment and medium

Abstract

The invention provides an ecological environment protection supervision method, system and device and a medium. The method comprises the following steps: acquiring multi-source data of an ecological environment to be protected and supervised; based on the multi-source data, performing environmental problem identification by using a pollution identification model to obtain an environmental problem identification result and a target area with an environmental problem; based on the obtained historical ecological data of the target area, performing pollution dynamic detection by using a dynamic prediction model to obtain a pollution diffusion path of the target area; and according to the pollution diffusion path, if it is identified that the target area has a pollution diffusion trend, the area is used as a key area to be supervised. According to the invention, environmental problems can be rapidly identified, a target area with the environmental problems can be rapidly positioned, pollution dynamic detection is carried out on the target area, and supervision can be carried out in time when the target area has a pollution diffusion trend, so that the supervision efficiency is improved, and a comprehensive and efficient environmental management system is constructed. The environmental pollution can be quickly and effectively handled, prevented and controlled.

Description

Technical Field

[0001] The present invention relates to the technical field of ecological environment protection supervision, and specifically relates to an ecological environment protection supervision method, system, device and medium. Background Art

[0002] Currently, ecological environment data is mainly collected by various sensors, and then the supervision of ecological environment protection is completed manually by supervisors. Among them, the supervision of ecological environment protection completed manually needs to go through processes such as "manual summary → on-site verification → report writing → problem feedback", and the average time-consuming reaches 7-15 days. Taking the water environment monitoring of a certain basin as an example, it often takes more than 2 weeks from discovering that the pollutant exceeds the standard to completing the source tracing, and at this time the pollution may have spread.

[0003] In summary, the traditional ecological environment protection supervision has the disadvantages of slow processing speed, low efficiency and poor scientificity. Summary of the Invention

[0004] In order to overcome the defects of slow processing speed, low efficiency and poor scientificity existing in the above-mentioned traditional ecological environment protection supervision, the present invention provides an ecological environment protection supervision method, including:

[0005] Obtain multi-source data of the ecological environment to be protected and supervised;

[0006] Based on the multi-source data, use a pollution identification model to identify environmental problems, and obtain an environmental problem identification result and a target area with environmental problems;

[0007] Based on the historical ecological data of the obtained target area, use a dynamic prediction model to detect pollution dynamics, and obtain the pollution diffusion path of the target area;

[0008] According to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, then take this area as a key area for supervision.

[0009] Optionally, the identifying environmental problems in the set area based on the multi-source data by using a pollution identification model to obtain an environmental problem identification result includes:

[0010] Use the dual-stream attention mechanism and gating mechanism in the pollution identification model to extract and fuse features of the multi-source data to obtain data features;

[0011] Based on the data features, use the dynamic target detection network in the pollution identification model to identify, and obtain an environmental problem identification result and a target area with environmental problems.

[0012] Optionally, the backbone network of the pollution identification model includes the MobileNetv3 and depthwise separable convolutional network, and a lightweight Transformer layer is embedded in the Neck layer of the pollution identification model. The lightweight Transformer layer includes a 4-head attention mechanism and a residual connection structure;

[0013] The dynamic prediction model is composed of LSTM combined with an attention mechanism.

[0014] Optionally, the step of taking the area as a key area for supervision includes:

[0015] Sending a supervision task to the supervision terminal of the supervisor, where the supervision task is used to conduct ecological environment protection supervision on the key area;

[0016] Receiving the on-site evidence collection data and heat map of the key area returned by the supervision terminal;

[0017] Sending the on-site evidence collection data and heat map of the key area to the rectification terminal of the rectification personnel in the key area.

[0018] Optionally, after identifying that there is a pollution diffusion trend in the target area according to the pollution diffusion path and taking the area as a key area for supervision, the method further includes:

[0019] After the rectification period ends, obtaining the rectification data of the key area;

[0020] Based on the rectification data, using the pollution identification model to identify environmental problems, and obtaining a rectification problem identification result;

[0021] Based on the rectification problem identification result and the environmental problem identification result, using a rectification effect evaluation formula to obtain the effect score of the key area;

[0022] If the effect score reaches a preset threshold, it is determined that the key area has been effectively rectified.

[0023] Optionally, the step of obtaining the effect score of the key area based on the rectification problem identification result and the environmental problem identification result by using the rectification effect evaluation formula includes:

[0024] Substituting the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula;

[0025] Based on the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result; based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result; use the peak signal-to-noise ratio as the effect score of the key area.

[0026] Optionally, the rectification effect evaluation formula satisfies the following formula:

[0027]

[0028] where MSE is the mean square error, I orig represents the environmental problem identification result, I comp represents the rectification problem identification result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, MAX I is the maximum possible range of the rectification data, M and N represent the width and height of the rectification data, (i,j) represents the (i,j)-th data point in the rectification data, and the value range of i is from 1 to M, and the value range of j is from 1 to N.

[0029] On the other hand, the present invention also provides an ecological environment protection supervision system, including:

[0030] An acquisition module, configured to acquire multi-source data of the ecological environment to be protected and supervised;

[0031] An environmental problem identification module, configured to identify environmental problems based on the multi-source data by using a pollution identification model, to obtain an environmental problem identification result and a target area with environmental problems;

[0032] A pollution dynamic prediction module, configured to perform pollution dynamic detection based on the historical ecological data of the obtained target area by using a dynamic prediction model, to obtain the pollution diffusion path of the target area;

[0033] A supervision module, configured to, according to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, use the area as a key area for supervision.

[0034] Optionally, the environmental problem identification module is specifically configured to use the dual-stream attention mechanism and the gating mechanism in the pollution identification model to perform feature extraction and fusion on the multi-source data to obtain data features; based on the data features, use the dynamic target detection network in the pollution identification model to perform identification, to obtain an environmental problem identification result and a target area with environmental problems.

[0035] Optionally, the backbone network of the pollution identification model includes the MobileNetv3 and depthwise separable convolutional network, and a lightweight Transformer layer is embedded in the Neck layer of the pollution identification model. The lightweight Transformer layer includes a 4-head attention mechanism and a residual connection structure;

[0036] The dynamic prediction model is composed of a long short-term memory network (LSTM) combined with an attention mechanism.

[0037] Optionally, the supervision module is specifically configured to send a supervision task to the supervision terminal of the supervisor. The supervision task is used to conduct ecological environment protection supervision on the key area;

[0038] Receive the on-site evidence collection data and heat map of the key area returned by the supervision terminal;

[0039] Send the on-site evidence collection data and heat map of the key area to the rectification terminal of the rectification personnel in the key area.

[0040] Optionally, the ecological environment protection supervision system further includes:

[0041] An effect scoring module, which is configured to, after the rectification period ends, obtain the rectification data of the key area; based on the rectification data, use the pollution identification model to identify environmental problems and obtain a rectification problem identification result; based on the rectification problem identification result and the environmental problem identification result, use a rectification effect evaluation formula to obtain the effect score of the key area; if the effect score reaches a preset threshold, it is determined that the key area has been effectively rectified.

[0042] Optionally, the effect scoring module is specifically configured to substitute the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula; based on the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result; based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result; use the peak signal-to-noise ratio as the effect score of the key area.

[0043] Optionally, the rectification effect evaluation formula satisfies the following formula:

[0044]

[0045] where MSE is the mean square error, I orig represents the environmental problem identification result, I comp represents the rectification problem identification result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, and MAX IFor the maximum possible range of the rectified data, M and N represent the width and height of the rectified data, and (i, j) represents the (i, j)-th data point in the rectified data. The value range of i is from 1 to M, and the value range of j is from 1 to N.

[0046] On the other hand, the present invention also provides an electronic device, including: at least one processor and a memory; the memory and the processor are connected through a bus;

[0047] The memory is used to store one or more programs;

[0048] When the one or more programs are executed by the at least one processor, the ecological environment protection supervision method described in any one of the above is implemented.

[0049] On the other hand, the present invention also provides a readable storage medium, on which an execution program is stored. When the execution program is executed, the ecological environment protection supervision method described in any one of the above is implemented.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] The present invention provides an ecological environment protection supervision method, system, device and medium. The method includes: obtaining multi-source data of the ecological environment to be protected and supervised; based on the multi-source data, using a pollution identification model to identify environmental problems, obtaining an environmental problem identification result and a target area with environmental problems; based on the historical ecological data of the obtained target area, using a dynamic prediction model to perform pollution dynamic detection, obtaining a pollution diffusion path of the target area; according to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, the area is taken as a key area for supervision. The present invention can quickly identify environmental problems, quickly locate the target area with environmental problems, improve the efficiency of environmental supervision, and perform pollution dynamic detection on the target area. When there is a pollution diffusion trend in the target area, supervision can be carried out in a timely manner, improving the supervision efficiency, thereby constructing a comprehensive and efficient environmental management system to quickly and effectively respond to and prevent and control environmental pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of the ecological environment protection supervision method of the present invention;

[0053] Figure 2 It is a structural schematic diagram of the ecological environment protection supervision system of the present invention;

[0054] Figure 3 It is a structural schematic diagram of the electronic device of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0055] The following further elaborates on the specific implementation manners of the present invention in conjunction with the accompanying drawings.

[0056] Embodiment 1:

[0057] An ecological environment protection supervision method provided by the present invention has a process schematic diagram as Figure 1 shown, and includes:

[0058] Step 101: Obtain multi-source data of the ecological environment to be subject to protection supervision;

[0059] Step 102: Based on the multi-source data, use a pollution identification model to identify environmental problems, obtaining an environmental problem identification result and a target area with environmental problems;

[0060] Step 103: Based on the historical ecological data of the obtained target area, use a dynamic prediction model to perform pollution dynamic detection, obtaining the pollution diffusion path of the target area;

[0061] Step 104: According to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, then the area is taken as a key area for supervision.

[0062] An ecological environment protection supervision method provided by an embodiment of the present invention is applied to an electronic device, and the electronic device can be a personal computer (PC), a server, etc.

[0063] In the present invention, the electronic device collects multi-source data of the ecological environment to be subject to protection supervision and conducts ecological environment protection supervision based on the multi-source data. Among them, the multi-source data includes but is not limited to satellite data, sensor data, and drone data, enabling comprehensive environmental monitoring through a trinity sensing network of space-based, air-based, and ground-based.

[0064] Exemplarily, the satellite monitoring module uses an S-band antenna to receive Sentinel-2 and Gaofen (GF) series satellite data, and updates the 2-meter resolution images covering the monitoring area daily. The ground monitoring module deploys intelligent sensor nodes in a 1 km × 1 km grid, and each node integrates water quality, atmosphere, and noise sensors, and uploads sensor data every 10 minutes through the Long Range Wide Area Network (LoRaWAN) protocol. The drone inspection module is equipped with a 6-rotor drone and a multi-spectral camera, performs a preset route scan at an altitude of 100 m - 300 m, covers 20 square kilometers for a single task, and can autonomously plan a supplementary flight path when encountering cloud cover.

[0065] Among them, after the electronic device obtains satellite data, sensor data and drone data, it can pre-process these data. For example, the satellite data is subjected to radiation correction and atmospheric correction to eliminate the influence of illumination and atmospheric scattering. The drone data is restored by the Structure from Motion (SFM) algorithm to generate a three-dimensional point cloud model with centimeter-level accuracy. The 3σ criterion is used to eliminate outliers in the ground sensor data. The sensor data, drone data and satellite data are calibrated in time and space through the spatiotemporal alignment algorithm to ensure the temporal and spatial consistency of data from different sources, and the data conflicts between data from different sources are eliminated through weighted Kalman filtering.

[0066] In the present invention, the automated collection and standardized processing of multi-scale and multi-temporal environmental data are realized, providing a high-quality data basis for subsequent intelligent analysis and solving the problems of limited coverage and poor timeliness of traditional manual sampling.

[0067] After acquiring multi-source data, the electronic device can use a pollution recognition model to identify environmental problems based on the multi-source data, thereby obtaining environmental problem recognition results and target areas with environmental problems. For example, the pollution recognition model can be a modified You Only Look Once version 5 (YOLOv5) object detection model.

[0068] Specifically, transfer learning is introduced into the YOLOv5 model as a pollution identification model, which can output environmental problem identification results. During identification, preprocessed multi-source data is input into the pollution identification model, which extracts and fuses features from the multi-source data. Based on the fused features, the model performs dynamic target detection and outputs environmental problem identification results and target areas with environmental problems.

[0069] After obtaining the environmental problem identification results and the target area with the environmental problem, the system retrieves the saved historical ecological data for the target area and uses the dynamic prediction model to perform dynamic pollution detection on the target area to determine the pollution diffusion path of the target area. For example, the historical ecological data can be daily average water quality values, etc., but this is not limited here.

[0070] In the present invention, a dynamic prediction model can be used to predict the pollution diffusion path within a preset time period in the future, and the preset time period can be 72 hours, etc.

[0071] In the present invention, if the electronic device identifies a pollution diffusion trend in a target area based on the pollution diffusion path, the target area is inspected as a key area. For example, the pollution diffusion trend in the target area includes but is not limited to an increase in pollution diffusion paths and an increase in the pollution area.

[0072] To improve the efficiency and accuracy of environmental inspection, based on the above multi-source data, a pollution recognition model is used to identify environmental problems in a set area, and the environmental problem recognition results include:

[0073] Using the dual-stream attention mechanism and gating mechanism in the pollution recognition model, feature extraction and fusion are performed on the multi-source data to obtain data features;

[0074] Based on the data features, a dynamic object detection network in the pollution recognition model is used for identification to obtain environmental problem recognition results and the target areas with environmental problems.

[0075] In the present invention, the pollution recognition model includes a dual-stream attention mechanism, a gating mechanism, and a dynamic object detection network. Exemplarily, the dynamic object detection network can be YOLOv5.

[0076] Among them, the dual-stream attention mechanism and the gating mechanism are used for feature extraction and fusion of multi-source data, and the dynamic object detection network is used for environmental problem recognition.

[0077] Specifically, the dual-stream attention mechanism and the gating mechanism in the pollution recognition model perform feature extraction and fusion on the multi-source data to obtain data features. The dynamic object detection network in the pollution recognition model performs environmental problem recognition based on the data features to obtain environmental problem recognition results and the target areas with environmental problems.

[0078] Among them, the training process of the pollution recognition model can be divided into two stages: first, pre-training is performed based on a general dataset, and then transfer learning is performed using the labeled environmental data of a specific area. The data augmentation strategy includes random rotation, light change, and simulated cloud occlusion to improve the robustness of the model. The output of the pollution recognition model includes pollution type, position coordinates, and confidence score, and can identify various typical environmental problems including water eutrophication, industrial waste gas emission, etc.

[0079] To improve the efficiency and accuracy of environmental inspection, the backbone network of the above pollution recognition model includes a mobile network MobileNetv3 and a depthwise separable convolutional network, and a Transformer lightweight layer is embedded in the Neck layer of the pollution recognition model. The Transformer lightweight layer includes a 4-head attention mechanism and a residual connection structure;

[0080] The dynamic prediction model is composed of a long short-term memory network LSTM combined with an attention mechanism.

[0081] In the present invention, the backbone network of the pollution recognition model includes a mobile network (MobileNetv3) and a depthwise separable convolutional network, and a lightweight Transformer layer is embedded in the Neck layer of the pollution recognition model. The lightweight Transformer layer includes a 4-head attention mechanism and a residual connection structure; the dynamic prediction model is composed of a Long Short-Term Memory (LSTM) network combined with an attention mechanism.

[0082] Specifically, the backbone network of the model uses MobileNetv3 combined with depthwise separable convolution, which reduces the number of parameters while maintaining accuracy and has a wider range of applications, such as being suitable for deployment on edge computing devices. The Neck layer generally refers to the middle part between the backbone network and the output head. In this application, a lightweight Transformer module is innovatively embedded, which contains 4 attention heads and a residual connection structure, enhancing the model's ability to capture global context information. The input layer supports multimodal data fusion and can simultaneously process satellite data, drone data, and sensor data.

[0083] The dynamic prediction model uses a bidirectional LSTM network as the basic architecture, which contains multiple hidden units. The dynamic prediction model automatically focuses on key time nodes through the attention mechanism. The input data is the time series of historical ecological data, including water quality indicators, meteorological data, and terrain information for the past 30 days. The dynamic prediction model first learns the spatio-temporal patterns of pollutant diffusion and then generates prediction results by combining real-time monitoring data.

[0084] To improve the efficiency and accuracy of environmental supervision, the above-mentioned supervision of key areas includes:

[0085] Sending a supervision task to the supervision terminal of the supervisor, where the supervision task is used to conduct ecological environment protection supervision on key areas;

[0086] Receiving on-site evidence data and heat maps of key areas returned by the supervision terminal;

[0087] Sending the on-site evidence data and heat maps of key areas to the rectification terminal of the rectification personnel in key areas.

[0088] In the present invention, the process of supervising key areas can be achieved by supervisors.

[0089] Specifically, an inspection task for ecological environment protection inspection of key areas is sent to the inspection terminals of inspectors. After receiving the inspection task, the inspectors go to the key areas for on-site evidence collection, including but not limited to taking on-site photos, collecting on-site samples and conducting tests, etc. The inspectors return the on-site evidence collection data and heat maps of the key areas through the inspection terminals.

[0090] Among them, the inspection task carries the location of the key area, such as longitude and latitude coordinates, problem types, recommended inspection contents and historical data references. The electronic device pushes it to the inspectors in the responsible area through a dedicated application (APP). This dedicated APP supports both Android and iOS platforms. When the inspectors conduct on-site verification, the dedicated APP provides functions such as navigation guidance, standard inspection procedures and automatic form filling.

[0091] The on-site evidence collection data (photos, videos, rapid test results) are transmitted back to the central platform in real time through 4G / 5G networks. Blockchain technology can also be used to ensure that the data cannot be tampered with, and all operation records have time stamps and geographical location information. For major environmental problems, multi-level early warnings can be automatically triggered, and relevant management departments and emergency response teams can be notified synchronously.

[0092] That is, the electronic device sends the on-site evidence collection data and heat maps of the key area to the rectification terminals of the rectification personnel in the key area. For example, dual-channel warnings of text messages / platforms can be triggered.

[0093] In order to improve the efficiency and accuracy of environmental inspections, after the above-mentioned method identifies that there is a pollution diffusion trend in the target area according to the pollution diffusion path and takes the area as a key area for inspection, the method further includes:

[0094] After the rectification cycle ends, obtain the rectification data of the key area;

[0095] Based on the rectification data, use the pollution identification model to identify environmental problems and obtain the rectification problem identification results;

[0096] Based on the rectification problem identification results and the environmental problem identification results, use the rectification effect evaluation formula to obtain the effect score of the key area;

[0097] If the effect score reaches the preset threshold, it is determined that the key area has been effectively rectified.

[0098] After the rectification cycle of the present invention ends, the rectification effect will be evaluated, and the rectification achievements will be objectively evaluated through quantitative indicators.

[0099] Specifically, after the rectification period ends, the electronic device obtains the rectification data of the key area, and based on the rectification data, uses the pollution identification model to identify environmental problems, obtaining the rectification problem identification result. The electronic device uses the rectification effect evaluation formula based on the rectification problem identification result and the environmental problem identification result before rectification to obtain the effect score of the key area. If the effect score reaches the preset threshold, it is determined that the key area has been effectively rectified.

[0100] Exemplarily, the rectification data includes satellite data, drone data, and sensor data.

[0101] In one example, a multi-dimensional index system can be adopted for rectification effect evaluation, including but not limited to: the image comparison dimension, calculating the difference in satellite data before and after rectification; the sensor data dimension, analyzing the improvement range of various environmental indicators; the on-site inspection dimension, verifying the implementation of rectification measures.

[0102] The electronic device will automatically generate a rectification evaluation report, including comparison charts of various indicators, compliance status, and improvement suggestions. For non-compliant items, a secondary rectification task will be automatically generated, and the subsequent monitoring frequency will be adjusted. All evaluation data is stored in the knowledge base for optimizing the pollution identification model.

[0103] To improve the efficiency and accuracy of environmental supervision, the above-mentioned obtaining the effect score of the key area based on the rectification problem identification result and the environmental problem identification result using the rectification effect evaluation formula includes:

[0104] Substitute the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula;

[0105] Based on the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result; based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result; use the peak signal-to-noise ratio as the effect score of the key area.

[0106] In the present invention, the electronic device calculates the peak signal-to-noise ratio (PSNR) of the rectification problem identification result and the environmental problem identification result through the rectification effect evaluation formula, and determines the peak signal-to-noise ratio as the effect score of the key area.

[0107] Specifically, substitute the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula. Through the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result, and based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result.

[0108] To improve the efficiency and accuracy of environmental inspections, the above rectification effect evaluation formula satisfies the following formula:

[0109]

[0110] Among them, MSE is the mean square error, I orig represents the environmental problem identification result, I comp represents the rectification problem identification result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, MAX I is the maximum possible range of the rectification data, M and N represent the width and height of the rectification data, and (i,j) represents the (i,j)-th data point in the rectification data. The value range of i is from 1 to M, and the value range of j is from 1 to N.

[0111] In the present invention, the rectification effect evaluation formula satisfies the following formula:

[0112]

[0113] Among them, MSE is the mean square error, I orig represents the environmental problem identification result, I comp represents the rectification problem identification result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, MAX I is the maximum possible range of the rectification data.

[0114] Based on this, the electronic device calculates the mean square error between the rectification problem identification result and the environmental problem identification result based on Formula 1. The electronic device calculates the peak signal-to-noise ratio between the rectification problem identification result and the environmental problem identification result based on Formula 2 and the mean square error.

[0115] In the present invention, if the PSNR value between the rectification problem identification result and the environmental problem identification result is greater than 28 dB, it is considered that the current rectification is an effective rectification, and a comprehensive score can be generated by combining the trend analysis of sensor data.

[0116] Compared with the traditional ecological environment protection inspection method, the time taken from the discovery of a pollution incident to the task assignment in the inspection process of the present invention is less than 15 minutes. Compared with the 6 - 48 hours required for traditional manual inspections, the inspection efficiency of the present invention is significantly improved. And in the inspection process of the present invention, the pollution source is located by the cooperation of satellite data and UAV data, so that the pollution source location error does not exceed 50 meters, and the monitoring integrity rate of the key areas with pollution diffusion is higher than 97%.

[0117] The following uses a specific embodiment to illustrate the embodiments of the present invention. The following steps are included in this embodiment:

[0118] (1) Obtain multi-source data of the ecological environment to be protected and supervised.

[0119] (2) Based on the multi-source data, use a pollution identification model to identify environmental problems, and obtain environmental problem identification results and target areas with environmental problems.

[0120] (3) Based on the historical ecological data of the obtained target area, use a dynamic prediction model to perform pollution dynamic detection, and obtain the pollution diffusion path of the target area.

[0121] (4) According to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, then the area is regarded as a key area for supervision.

[0122] (5) Send a supervision task to the supervision terminal of the supervisor. The supervision task is used to conduct ecological environment protection supervision on the key area; receive the on-site evidence-taking data and heat map of the key area returned by the supervision terminal; send the on-site evidence-taking data and heat map of the key area to the rectification terminal of the rectification personnel in the key area.

[0123] (6) After the rectification period ends, obtain the rectification data of the key area; based on the rectification data, use a pollution identification model to identify environmental problems, and obtain rectification problem identification results; based on the rectification problem identification results and environmental problem identification results, use a rectification effect evaluation formula to obtain the effect score of the key area; if the effect score reaches the preset threshold, it is determined that the key area has been effectively rectified.

[0124] The present invention can quickly identify environmental problems, quickly locate the target areas with environmental problems, improve the efficiency of environmental supervision, and conduct pollution dynamic detection on the target areas. When there is a pollution diffusion trend in the target area, supervision can be carried out in a timely manner, improving the supervision efficiency, thereby constructing a comprehensive and efficient environmental management system to quickly and effectively respond to and prevent and control environmental pollution.

[0125] The present invention discloses an ecological environment protection supervision service system and method based on multi-source data fusion. The system integrates a Geographic Information System (GIS), a big data analysis platform and an artificial intelligence model to realize real-time discovery, intelligent analysis and efficient supervision of ecological environment protection problems. The system includes an environmental problem automatic identification module, a key area dynamic monitoring module, an intelligent task assignment module and a rectification effect evaluation module. By integrating satellite remote sensing, unmanned aerial vehicle monitoring and ground sensor data, a full-scale and multi-scale ecological environment supervision network is constructed, which can quickly lock the pollution source and provide accurate decision-making support. The present invention significantly improves the efficiency and scientificity of ecological environment protection supervision.

[0126] Example 2:

[0127] Based on the same inventive concept, the present invention also provides an ecological environment protection supervision system, the structural schematic diagram of which is as Figure 2 shown, including:

[0128] An acquisition module 201, configured to acquire multi-source data of the ecological environment to be protected and supervised;

[0129] An environmental problem identification module 202, configured to identify environmental problems based on the multi-source data by using a pollution identification model, so as to obtain an environmental problem identification result and a target area with environmental problems;

[0130] A pollution dynamic prediction module 203, configured to perform pollution dynamic detection based on the historical ecological data of the target area obtained by using a dynamic prediction model, so as to obtain the pollution diffusion path of the target area;

[0131] A supervision module 204, configured to, according to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, use the area as a key area for supervision.

[0132] In a specific implementation manner, the environmental problem identification module 202 is specifically configured to use the dual-stream attention mechanism and the gating mechanism in the pollution identification model to perform feature extraction and fusion on the multi-source data to obtain data features; based on the data features, use the dynamic target detection network in the pollution identification model to perform identification, so as to obtain an environmental problem identification result and a target area with environmental problems.

[0133] In a specific implementation manner, the backbone network of the pollution identification model includes a mobile network MobileNetv3 and a depthwise separable convolutional network, and a Transformer lightweight layer is embedded in the Neck layer of the pollution identification model. The Transformer lightweight layer includes a 4-head attention mechanism and a residual connection structure;

[0134] The dynamic prediction model is composed of a long short-term memory network LSTM combined with an attention mechanism.

[0135] In a specific implementation manner, the supervision module 204 is specifically configured to send a supervision task to the supervision terminal of the supervisor, and the supervision task is used to conduct ecological environment protection supervision on the key area;

[0136] Receive the on-site evidence collection data and the heat map of the key area returned by the supervision terminal;

[0137] Send the on-site evidence collection data and the heat map of the key area to the rectification terminal of the rectification personnel of the key area.

[0138] In a specific implementation manner, the ecological environment protection supervision system further includes:

[0139] The effect scoring module 205 is configured to obtain the rectification data of the key area after the end of the rectification period; based on the rectification data, use the pollution identification model to identify environmental problems, and obtain the rectification problem identification result; based on the rectification problem identification result and the environmental problem identification result, use the rectification effect evaluation formula to obtain the effect score of the key area; if the effect score reaches the preset threshold, it is determined that the key area has been effectively rectified.

[0140] In a specific implementation manner, the effect scoring module 205 is specifically configured to substitute the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula; based on the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result; based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result; use the peak signal-to-noise ratio as the effect score of the key area.

[0141] Optionally, the rectification effect evaluation formula satisfies the following formula:

[0142]

[0143] where MSE is the mean square error, I orig represents the environmental problem identification result, I comp represents the rectification problem identification result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, MAX I is the maximum possible range of the rectification data, M and N represent the width and height of the rectification data, and (i,j) represents the (i,j)-th data point in the rectification data, where the value range of i is from 1 to M, and the value range of j is from 1 to N.

[0144] Embodiment 3:

[0145] As Figure 3 shown, the present invention further provides an electronic device, which may be a computer device, a single-chip microcomputer device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected by a bus; the memory can be used to store an execution program, and the exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, and the data can be called and / or modified when the instructions are executed.

[0146] The processor may be a Central Processing Unit (CPU), or it may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of an ecological environment protection supervision method in the above embodiments.

[0147] Embodiment 4:

[0148] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device-readable storage medium (Memory). The electronic device-readable storage medium is a memory device in the electronic device, used to store programs and data. It can be understood that the storage medium here can include both the built-in storage medium in the electronic device and, of course, the extended storage medium supported by the electronic device. The storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. By the processor loading and executing one or more instructions stored in the storage medium, the steps of an ecological environment protection supervision method in the above embodiments can be implemented.

[0149] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0150] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0151] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0152] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that after reading the present invention, various changes, modifications, or equivalent replacements can still be made to the specific implementation manners of the application. However, these changes, modifications, or equivalent replacements are all within the protection scope of the pending claims of the application.

Claims

1. An ecological environment protection supervision method, characterized in that, Including: Obtain multi-source data of the ecological environment to be protected and supervised; Based on the multi-source data, use a pollution identification model to identify environmental problems, and obtain an environmental problem identification result and a target area with environmental problems; Based on the historical ecological data of the obtained target area, use a dynamic prediction model to conduct pollution dynamic detection, and obtain the pollution diffusion path of the target area; According to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, then use this area as a key area for supervision.

2. The method according to claim 1, wherein The using the pollution identification model to identify environmental problems in the set area based on the multi-source data and obtaining the environmental problem identification result includes: Use the dual-stream attention mechanism and gating mechanism in the pollution identification model to extract and fuse features of the multi-source data, and obtain data features; Based on the data features, use the dynamic target detection network in the pollution identification model to conduct identification, and obtain an environmental problem identification result and a target area with environmental problems.

3. The method according to claim 1 or 2, characterized in that, The backbone network of the pollution identification model includes a mobile network MobileNetv3 and a depthwise separable convolutional network, and a Transformer lightweight layer is embedded in the Neck layer of the pollution identification model. The Transformer lightweight layer includes a 4-head attention mechanism and a residual connection structure; The dynamic prediction model is composed of a long short-term memory network LSTM combined with an attention mechanism.

4. The method according to claim 1, wherein The using this area as a key area for supervision includes: Send a supervision task to the supervision terminal of the supervisor, and the supervision task is used to conduct ecological environment protection supervision on the key area; Receive the on-site evidence collection data and heat map of the key area returned by the supervision terminal; Send the on-site evidence collection data and heat map of the key area to the rectification terminal of the rectification personnel in the key area.

5. The method according to claim 1 or 4, characterized in that After the using the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, and then using this area as a key area for supervision, the method further includes: After the rectification period ends, obtain the rectification data of the key area; Based on the rectification data, use the pollution identification model to identify environmental problems, and obtain a rectification problem identification result; Based on the rectification problem identification result and the environmental problem identification result, use a rectification effect evaluation formula to obtain the effect score of the key area; If the effect score reaches a preset threshold, it is determined that effective rectification has been carried out on the key area.

6. The method according to claim 5, wherein The using the rectification problem identification result and the environmental problem identification result, and using a rectification effect evaluation formula to obtain the effect score of the key area includes: Substitute the rectification problem identification result and the environmental problem identification result into the rectification effect evaluation formula; Based on the rectification effect evaluation formula, calculate the mean square error of the rectification problem identification result and the environmental problem identification result; based on the mean square error, calculate the peak signal-to-noise ratio of the rectification problem identification result and the environmental problem identification result; use the peak signal-to-noise ratio as the effect score of the key area.

7. The method according to claim 6, characterized in that The rectification effect evaluation formula satisfies the following formula: where MSE is the mean square error, I orig represents the environmental problem recognition result, I comp represents the rectification problem recognition result, M×N represents the resolution of the rectification data, PSNR is the peak signal-to-noise ratio, MAX I is the maximum possible range of the rectification data, M and N represent the width and height of the rectification data, and (i,j) represents the (i,j)-th data point in the rectification data. The value range of i is from 1 to M, and the value range of j is from 1 to N.

8. An ecological environment protection supervision system, characterized in that, Including: An acquisition module, configured to acquire multi-source data of the ecological environment to be protected and supervised; An environmental problem identification module, configured to identify environmental problems based on the multi-source data by using a pollution identification model, so as to obtain an environmental problem identification result and a target area with environmental problems; A pollution dynamic prediction module, configured to perform pollution dynamic detection on the basis of the historical ecological data of the obtained target area by using a dynamic prediction model, so as to obtain a pollution diffusion path of the target area; An inspection module, configured to, according to the pollution diffusion path, if it is identified that there is a pollution diffusion trend in the target area, use the area as a key area for inspection.

9. An electronic device, characterized in that, Including: At least one processor and a memory; The memory and the processor are connected by a bus; The memory is configured to store one or more programs; When the one or more programs are executed by the at least one processor, the ecological environment protection inspection method according to any one of claims 1-7 is implemented.

10. A readable storage medium, characterized in that, There is an execution program stored thereon, and when the execution program is executed, the ecological environment protection inspection method according to any one of claims 1-7 is implemented.

Images