Hazardous waste dumping high-risk area identification early warning method based on multi-dimensional data

Abstract

The invention provides a dangerous waste dumping high-risk area identification early warning method based on multi-dimensional data, and belongs to the technical field of image identification. According to the technical scheme, firstly, a first coverage distribution model is obtained through pre-training based on first set data, then image data including hyperspectral imaging information and pixel point information are obtained, and different weights are distributed; on this basis, second set data of the target coverage in the target area is obtained, feature training is carried out to obtain a second coverage distribution model, and then the high-risk area range of the target coverage is identified from the current image data; finally, remote sensing data is adopted for correction, and exact area boundary data is obtained. The dangerous waste dumping high-risk area is accurately and efficiently recognized based on the image recognition method, and then monitoring and early warning are achieved. The method is obviously superior to the currently adopted on-site investigation and sampling detection means, and the recognition accuracy and efficiency are comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, specifically to a method for identifying and issuing early warnings of high-risk areas for hazardous waste dumping based on multi-dimensional data. Background Technology

[0002] Hazardous waste refers to solid waste that is listed in the National Hazardous Waste Inventory or identified as having hazardous characteristics according to the national hazardous waste identification standards and methods. Hazardous waste mainly includes solid waste (including liquid waste) with hazardous characteristics such as corrosivity, toxicity, flammability, reactivity, and infectivity, such as medical waste, chemical waste, and waste containing heavy metals.

[0003] Hazardous waste dumping refers to the act of arbitrarily discharging, stockpiling, burying, or disposing of hazardous waste without permission, in violation of national regulations. Harmful substances in hazardous waste can flow into rivers, lakes, and seas with natural precipitation runoff, causing surface water pollution; leachate seeps into the soil, further polluting groundwater. Direct discharge into water systems or well discharge into groundwater systems will result in more severe and irreversible pollution. The evaporation, sublimation, and microbial decomposition of hazardous waste itself produce harmful gases that directly pollute the atmosphere; fine particles and dust are dispersed by wind, exacerbating atmospheric dust pollution; harmful gases and dust generated during improper transportation, storage, and disposal also indirectly pollute the air. Hazardous waste dust and particles drift to the soil surface and enter the soil, causing direct pollution; liquid and semi-solid waste spilled on the ground or seeping into the soil can cause harmful components to migrate with the soil, polluting groundwater and entering the food chain through bioaccumulation. Therefore, combating and preventing hazardous waste dumping is crucial.

[0004] Despite the high-pressure crackdown by national environmental protection departments on hazardous waste dumping, some criminals still risk illegally dumping hazardous waste indiscriminately or covertly to reduce production or operating costs. For dumping that has already occurred, minimizing its harm is a crucial remedial measure, which first requires delineating the dumping area. However, relying solely on on-site identification by suspects is not only inaccurate but also time-consuming. Therefore, using technological means for identification and delineation is particularly important. Currently, conventional identification and delineation methods mainly rely on on-site investigation and sampling tests, which are labor-intensive, time-consuming, and their accuracy is not entirely satisfactory. Summary of the Invention

[0005] This invention aims to address the technical deficiencies of existing technologies by providing a method for identifying and warning of high-risk areas for hazardous waste dumping based on multi-dimensional data. This method solves the technical problems of current identification methods that rely on on-site investigation and sampling, such as high workload, long time consumption, and insufficient accuracy.

[0006] To achieve the above technical objectives, the present invention adopts the following technical solution: A method for identifying and issuing early warnings of high-risk areas for hazardous waste dumping based on multi-dimensional data includes: 1) Obtain a first overlay distribution model, which is a model pre-trained based on a first set of data; the first set of data is a set of past image data of target overlays in several regions over several time periods; the image data includes hyperspectral imaging information and pixel information; the hyperspectral imaging information and pixel information are given different weights in different time periods; 2) Obtain the second set of data of the target cover in the target area, wherein the second set of data is the set of data obtained by aggregating the previous image data of the target cover in the target area over a certain period of time; 3) Based on the first set of data and the second set of data, perform feature training on the first cover distribution model to obtain the second cover distribution model, and use the second cover distribution model to identify the high-risk area range of the target cover from the current image data of the target area; 4) For the high-risk area, acquire the primary resolution remote sensing data at the first time point; perform image segmentation on the primary resolution remote sensing data to obtain a multi-region image set; the multi-region image set includes image information from several different locations; assign the cover type to the multi-region image set to determine the cover type and cover assignment target corresponding to the image information at different locations; based on the cover type, cover assignment target, primary resolution remote sensing data at the first time point, and secondary resolution remote sensing data at the first time point, correct the secondary resolution remote sensing data at the second time point to the regional boundary data corresponding to the image information at the different locations at the second time point, thus obtaining the high-risk area.

[0007] Preferably, in step 4), the remote sensing data is first subjected to median filtering to obtain denoised data, and then the missing data is filled in the denoised data.

[0008] Preferably, the first overlay distribution model includes an encoding layer, a feature compression layer, a decoding layer, and a feature connection layer.

[0009] Preferably, the second cover distribution model is shown in the following equation: ; in, M , C , K These are image resolution, bit depth, and saturation coefficient, respectively. For noise displacement, parameters , It takes into account the feature parameters of the location where the image acquisition device is located.

[0010] As a preferred embodiment, the first cover distribution model is shown in the following equation: ; in, f 1(t) and f 2( t J1 and J2 are the sequence and pixel datasets of the image data, respectively; J1 and J2 are the location and type data of the target object to be identified. m For pixel coordinate labels, k For correction factor, b These are feature values ​​for a fixed region within the image acquisition area; y This represents the change in displacement of the covering.

[0011] Preferably, the hyperspectral imaging information in step 1) includes spectral data of each point on the image and image information of any spectral band.

[0012] Preferably, the pixel information in step 1) includes the pixel's color value, brightness value, and coordinate value.

[0013] Preferably, the high-risk area range in step 3) includes at least two levels of boundaries, with a transition zone between adjacent two-level boundaries, and the edge pixels of the transition zone are represented by coordinates.

[0014] Preferably, the different weights are calculated using the following formula: ; in, This represents the correlation coefficient between hyperspectral imaging information and pixel information.

[0015] Preferably, the type of cover and the cover allocation target in step 4) are preset based on previous data.

[0016] This invention provides a method for identifying and issuing early warnings of high-risk areas for hazardous waste dumping based on multi-dimensional data. The technical solution first pre-trains a first cover distribution model based on a first set of data. Then, it acquires image data, including hyperspectral imaging information and pixel information, and assigns different weights to each. Based on this, it acquires a second set of data on the target cover in the target area, and uses this data to train the first cover distribution model, resulting in a second cover distribution model. This second cover distribution model is then used to identify the high-risk area range of the target cover from the current image data of the target area. Finally, remote sensing data is used to correct the high-risk area range, obtaining precise area boundary data. This invention accurately and efficiently identifies high-risk areas for hazardous waste dumping based on image recognition methods, thereby achieving monitoring and early warning. This method is significantly superior to current on-site investigation and sampling detection methods, achieving a comprehensive improvement in both accuracy and efficiency. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, those skilled in the art will understand that with technological development and the emergence of new scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0019] In the description of this application, it should be understood that, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. Furthermore, any terminology used is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0020] Furthermore, to better illustrate this application, numerous specific details are provided in the following detailed embodiments. Those skilled in the art should understand that this application can be implemented without certain specific details. In some instances, methods, means, components, and circuits well-known to those skilled in the art have not been described in detail in order to highlight the main points of this application.

[0021] A method for identifying and warning of high-risk areas for hazardous waste dumping based on multi-dimensional data, such as Figure 1 As shown, it includes: 1) Obtain a first overlay distribution model, which is a model pre-trained based on a first set of data; the first set of data is a set of past image data of target overlays in several regions over several time periods; the image data includes hyperspectral imaging information and pixel information; the hyperspectral imaging information and pixel information are given different weights in different time periods; 2) Obtain the second set of data of the target cover in the target area, wherein the second set of data is the set of data obtained by aggregating the previous image data of the target cover in the target area over a certain period of time; 3) Based on the first set of data and the second set of data, perform feature training on the first cover distribution model to obtain the second cover distribution model, and use the second cover distribution model to identify the high-risk area range of the target cover from the current image data of the target area; 4) For the high-risk area, acquire the primary resolution remote sensing data at the first time point; perform image segmentation on the primary resolution remote sensing data to obtain a multi-region image set; the multi-region image set includes image information from several different locations; assign the cover type to the multi-region image set to determine the cover type and cover assignment target corresponding to the image information at different locations; based on the cover type, cover assignment target, primary resolution remote sensing data at the first time point, and secondary resolution remote sensing data at the first time point, correct the secondary resolution remote sensing data at the second time point to the regional boundary data corresponding to the image information at the different locations at the second time point, thus obtaining the high-risk area.

[0022] In step 4), the remote sensing data is first subjected to median filtering to obtain denoised data, and then the missing data is filled in using the denoised data. The first cover distribution model includes an encoding layer, a feature compression layer, a decoding layer, and a feature connection layer.

[0023] The second cover distribution model is shown in the following equation: ; in, M , C , K These are image resolution, bit depth, and saturation coefficient, respectively. For noise displacement, parameters , It takes into account the feature parameters of the location where the image acquisition device is located.

[0024] The first cover distribution model is shown in the following equation: ; in, f 1(t) andf 2( t J1 and J2 are the sequence and pixel datasets of the image data, respectively; J1 and J2 are the location and type data of the target object to be identified. m For pixel coordinate labels, k For correction factor, b These are feature values ​​for a fixed region within the image acquisition area; y The change in displacement of the covering material. The hyperspectral imaging information in step 1) includes spectral data for each point on the image and image information for any spectral band. The pixel information in step 1) includes the color value, brightness value, and coordinate value of the pixel. The high-risk area in step 3) includes at least two levels of boundaries, with a transition zone between adjacent boundaries, and the edge pixels of the transition zone are represented by coordinates.

[0025] The different weights are calculated using the following formula: ; in, This represents the correlation coefficient between hyperspectral imaging information and pixel information.

[0026] The cover type and cover allocation target mentioned in step 4) are preset based on previous data.

[0027] In summary, after reading this detailed disclosure, those skilled in the art will understand that the foregoing detailed disclosure is presented by way of example only and is not restrictive. Although not explicitly stated herein, those skilled in the art will understand that this application is intended to encompass various reasonable changes, improvements, and modifications to the embodiments. These changes, improvements, and modifications are intended to be made by this application and are within the spirit and scope of the exemplary embodiments of this application.

[0028] It should be understood that in the foregoing description of the embodiments of this application, various features are combined in a single embodiment, drawing, or description for the purpose of simplifying the understanding of a feature. However, this does not mean that the combination of these features is necessary, and those skilled in the art can certainly extract some of the features as individual embodiments when reading this application.

[0029] It should be understood that the embodiments disclosed herein are illustrative of the principles of this application. Other modified embodiments are also within the scope of this application. The embodiments disclosed herein are merely examples and not limitations, and the embodiments of this application are not limited to the embodiments precisely described above.

Claims

1. A method for identifying and issuing early warnings of high-risk areas for hazardous waste dumping based on multi-dimensional data, characterized in that: include: 1) Obtain the first cover distribution model, which is a model pre-trained based on the first set of data; The first set of data is a set of image data of target coverage in several regions over several time periods; the image data includes hyperspectral imaging information and pixel information; the hyperspectral imaging information and pixel information are given different weights in different time periods; 2) Obtain the second set of data of the target cover in the target area, wherein the second set of data is the set of data obtained by aggregating the previous image data of the target cover in the target area over a certain period of time; 3) Based on the first set of data and the second set of data, perform feature training on the first cover distribution model to obtain the second cover distribution model, and use the second cover distribution model to identify the high-risk area range of the target cover from the current image data of the target area; 4) For the high-risk area, acquire the main resolution remote sensing data at the first time point; Image segmentation is performed on the main resolution remote sensing data to obtain a multi-region image set; the multi-region image set includes image information from several different locations. The multi-region image set is assigned an overlay type to determine the overlay type and overlay assignment target corresponding to image information at different locations; Based on the cover type, cover allocation target, primary resolution remote sensing data and secondary resolution remote sensing data at the first time point corresponding to the image information at different locations, the secondary resolution remote sensing data at the second time point is corrected to the regional boundary data corresponding to the image information at the different locations at the second time point, thus obtaining the high-risk area.

2. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, In step 4), the remote sensing data is first subjected to median filtering to obtain denoised data, and then the missing data is filled in using the denoised data.

3. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The first overlay distribution model includes an encoding layer, a feature compression layer, a decoding layer, and a feature connection layer.

4. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The second cover distribution model is shown in the following equation: ； in, M , C , K These are image resolution, bit depth, and saturation coefficient, respectively. For noise displacement, parameters , It takes into account the feature parameters of the location where the image acquisition device is located.

5. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The first cover distribution model is shown in the following equation: ; in, f 1(t) and f 2( t J1 and J2 are the sequence and pixel datasets of the image data, respectively; J1 and J2 are the location and type data of the target object to be identified. m For pixel coordinate labels, k For correction factor, b These are feature values ​​for a fixed region within the image acquisition area; y This represents the change in displacement of the covering.

6. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The hyperspectral imaging information mentioned in step 1) includes spectral data of each point on the image and image information of any spectral band.

7. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The pixel information mentioned in step 1) includes the pixel's color value, brightness value, and coordinate value.

8. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, Step 3) The high-risk area includes at least two levels of boundaries, with a transition zone between adjacent two levels of boundaries, and the edge pixels of the transition zone are represented by coordinates.

9. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The different weights are calculated using the following formula: ； in, This represents the correlation coefficient between hyperspectral imaging information and pixel information.

10. The method for identifying and issuing early warning of high-risk areas for hazardous waste dumping based on multi-dimensional data according to claim 1, characterized in that, The cover type and cover allocation target mentioned in step 4) are preset based on previous data.

Images