CN117456378A - Water conservancy digital twin base element realization method and system based on satellite remote sensing - Google Patents

Abstract

The invention relates to a method and system for realizing water conservancy digital twin base elements based on satellite remote sensing, and belongs to the field of remote sensing technology. This invention uses the supercomputing power of the GEE platform to directly find the required remote sensing images on the platform and perform preprocessing operations, and then constructs vegetation index feature values that can improve the accuracy of ground object classification, and then uses recognition software to convert the vegetation index feature values into Organically integrated with the multi-scale segmentation method, the remote sensing images are classified into ground objects, and the accuracy of the classification results is tested. Then the water conservancy elements are distinguished based on the ground object classification results. A proposal is made to use the aspect ratio and area of the water body shape. Index is a method to determine whether a water body is a river, reservoir or lake. The invention greatly improves the efficiency of water conservancy element discrimination and provides strong support for the construction of data base in the digital twin base.

Description

Method and system for implementing water conservancy digital twin base elements based on satellite remote sensing

Technical field

The invention relates to a method and system for realizing water conservancy digital twin base elements based on satellite remote sensing, and belongs to the field of remote sensing technology.

Background technique

Digital twin, the English name is Digital Twin (digital twin), also known as digital mapping and digital mirroring. It makes full use of data such as physical models, sensor updates, and operation history, integrates multi-disciplinary, multi-physical quantities, multi-scale, and multi-probability simulation processes, and completes mapping in virtual space to reflect the full life cycle process of the corresponding physical equipment. . It can be seen that applying digital twin technology to the field of water conservancy management, based on digital clones of rivers and lakes, can grasp the latest status of rivers and lakes, and then conduct real-time monitoring and scientific management. With the continuous development of satellite remote sensing technology, remote sensing technology has become the "clairvoyance" of the water conservancy industry and plays a very important role in the construction of digital twins. Remote sensing technology can provide multi-dimensional, multi-spatial and spatial resolution, and high-timeliness data. These data can be used to build a data base platform for digital twin bases, including dynamic monitoring of geographical entities such as rivers, lakes, and water conservancy projects. In particular, remote sensing technology can Real-time monitoring and identification of water conservancy elements on remote sensing images can provide comprehensive location information and changing trends of water conservancy elements.

The traditional method of monitoring water conservancy elements is to download satellite images, then extract feature types, and then distinguish water body elements and water conservancy project elements based on the extracted feature types. This method not only requires downloading satellite images locally, which consumes a lot of time and energy, is inefficient, and takes up device storage space. Subsequently, the classified features are classified into water body categories. This process also requires a lot of manual participation and inspection one by one, resulting in high labor costs.

Contents of the invention

The purpose of the present invention is to overcome the above shortcomings and provide a method for implementing water conservancy digital twin base elements based on satellite remote sensing, which greatly improves the efficiency of remote sensing image processing.

The technical solutions adopted by the present invention are:

The implementation method of water conservancy digital twin base elements based on satellite remote sensing includes the following steps:

S1. Obtain the required remote sensing images of the study area on the GEE platform and perform preprocessing;

S2. Construct vegetation index characteristic values, including normalized vegetation index NDVI, soil-adjusted vegetation index SAVI and normalized difference water index NDWI;

S3. Use recognition software to organically integrate vegetation index feature values and multi-scale segmentation methods to classify remote sensing images;

S4. Perform an accuracy test on the classification results of ground objects that incorporate vegetation feature index. If the classification results are good, proceed to the next step;

S5. Distinguish water conservancy elements based on the classification results of ground objects, including hydraulic structures, rivers, reservoirs, and lakes. Use the aspect ratio and area index of the water body shape to determine whether the water body is a river, reservoir, or lake.

In the above method, the preprocessing described in step S1 is to perform orthorectification and atmospheric correction on the image, then perform shadow removal and cloud mask operations, and use the median filter method to obtain inter-annual composite images, thereby combining all remote sensing images in the study year. Combined into one, a remote sensing image that can clearly and completely display the surface coverage information of the study area is obtained.

The calculation method of the vegetation index characteristic value described in step S2 is as follows:

,

,

,

Among them, P _NIR : reflectivity in the near-infrared band,

P _RED : Reflectivity in the infrared band,

P _GREEN : reflectivity of green light band,

L: Soil adjustment coefficient.

In step S3, the multi-scale segmentation method mainly divides the pixels in the image into cultivated land, woodland, grassland, buildings, water bodies, and other land; by setting the threshold of feature values, woodland, water bodies, and buildings are mainly distinguished, in which NDVI The threshold is set to 0.46 to distinguish forest land from non-forest areas. NDVI greater than 0.46 is represented as forest land; NDWI is set to 0.2 to extract water bodies. NDWI greater than 0.2 is represented as water bodies, including water bodies in urban areas and shallow river areas; SAVI is set to 0.2 to distinguish buildings from non-buildings, where SAVI less than 0.2 is represented as a building.

In step S4, the confusion matrix accuracy test method is used, and the overall accuracy OA and Kappa coefficient are used to verify the accuracy of the classification results, where:

,

,

,

TP: Prediction is positive, actual is positive,

FN: Prediction is negative, actual is positive,

FP: Prediction is positive, actual is negative,

TN: Prediction is negative, actual is negative,

AA: average accuracy,

The Kappa coefficient after accuracy verification is greater than 0.6, indicating a good classification result.

In step S5, the method for identifying hydraulic buildings is: extract the buildings from the feature classification. If the building is a regular rectangle and is beside the reservoir, it indicates that the building is a reservoir dam; there are obvious structures in the middle of the river. The building that cuts off the water body is represented as a barrage. When identifying water bodies, the smallest enclosing rectangle of an irregular-shaped water body is regarded as the boundary of the water body. The length and width of the rectangle are regarded as the length and width of the water body. When the aspect ratio K1 of the water body is greater than 44, it means that the water body is a river; when the water body If the aspect ratio K2 of the water body is greater than 2.5 and the area S2 is greater than 0.37 hectares, and there are engineering dam buildings around the water body, the water body is a reservoir; when the aspect ratio K3 of the water body is greater than 2 and the area S3 is greater than 10.37 hectares, there are no engineering projects around the water body. The dam building is a lake.

A water conservancy digital twin base element implementation system based on satellite remote sensing includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the satellite remote sensing as described above. Implementation method of water conservancy digital twin base elements.

The beneficial effects of the present invention are:

This invention utilizes the powerful supercomputing capabilities of the GEE platform to perform batch processing and calculation of data directly on the platform, completely changing the traditional mode of downloading remote sensing data to the device for processing and analysis, and greatly improving the efficiency of remote sensing image processing. . In addition, the present invention has developed a new method for identifying rivers, lakes and reservoirs in water conservancy elements. By calculating the aspect ratio and area of the water body shape, it is determined whether the water body is a river, a reservoir or a lake. This method can realize automatic calculation, and only need to perform manual screening after calculation to obtain accurate river and lake element information, which greatly improves the efficiency of water conservancy element identification and provides a strong basis for the construction of the data base in the digital twin base. support.

Description of the drawings

Figure 1 is a flow chart of the method of the present invention;

Figure 2 is a remote sensing image of a river basin obtained by the embodiment of the present invention;

Figure 3 is a schematic diagram of the aspect ratios of rivers, reservoirs, and lakes according to the embodiment of the present invention; (a) is a river, (b) is a reservoir, and (c) is a lake;

Figure 4 shows the identification results of rivers, lakes and reservoirs in the embodiment of the present invention.

Detailed ways

The present invention will be further described below in conjunction with specific embodiments.

Example 1 Implementation method of water conservancy digital twin base elements based on satellite remote sensing, including the steps (see Figure 1) as follows:

S1. Obtain the required remote sensing images of the study area on the GEE platform and perform preprocessing:

Select the Gaofen-1 satellite images that meet the requirements on the Google Earth Engine (GEE) platform, and preprocess the remote sensing images of the study area. The specific steps are: on the GEE platform (https://developers.google.cn/earth- engine/) After selecting the satellite remote sensing data of the corresponding phase, perform orthorectification and atmospheric correction on it, then use the cloud mask tool to perform shadow removal and cloud mask operations, and use the median filter method to obtain the interannual composite image, thereby All remote sensing images in the study year were combined into one to obtain a remote sensing image that can clearly and completely display the surface coverage information of the study area.

This embodiment acquires remote sensing images of a river basin in Shandong Province, as shown in Figure 2.

S2. Construct vegetation index characteristic values, including normalized vegetation index NDVI, soil-adjusted vegetation index SAVI and normalized difference water index NDWI:

The application of vegetation index as a feature value in an object-oriented classification algorithm can effectively make up for the lack of spectral value of the remote sensing image itself, and can amplify the feature value to achieve accurate extraction of ground objects. This invention mainly adds three characteristic values: Normalized Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI) and Normalized Difference Water Index (NDWI) to improve the accuracy of ground object classification. details as follows:

,

,

,

P _NIR : reflectivity in the near infrared band,

P _RED : Reflectivity in the infrared band,

P _GREEN : reflectivity of green light band,

L: soil adjustment coefficient, which is set to 0.6 in the present invention.

S3. Use recognition software to organically integrate vegetation index feature values and multi-scale segmentation methods to classify remote sensing images:

ecognition software (Yikang Software) is an object-oriented classification system and an image processing software based on machine learning and deep learning algorithms. The multi-scale segmentation method is a bottom-up segmentation method. On the premise of minimizing the average heterogeneity between targets and maximizing the uniformity among internal pixels, it achieves a segmentation-based segmentation method by merging adjacent pixels or small segmented objects. Image segmentation using regional fusion technology. In this invention, the pixels in the image are mainly divided into cultivated land, woodland, grassland, buildings, water bodies, and other land. The specific steps are: import image data, write the segmentation process in the [Process Tree] dialog box, insert [Append New], name it "Segmentation", select the algorithm as multi-scale segmentation [multiresolution segmentation], and set the segmentation parameter [Scale parameter] ] is 0.4, the shape factor parameter (shape) is 0.2, and the compactness parameter (compactness) is 0.5, and then click [Execute] to perform the operation.

The vegetation index feature value is integrated into the multi-scale segmentation method to improve the accuracy of ground object classification. By setting the threshold of the feature value, woodlands, water bodies and buildings are distinguished. Among them, setting the NDVI threshold to 0.46 can clearly distinguish forest land from non-forest areas. NDVI greater than 0.46 is represented as forest land; setting NDWI to 0.2 can better extract water bodies. NDWI greater than 0.2 is represented as water bodies, including Water bodies and shallow water areas of rivers in urban areas; setting SAVI to 0.2 can distinguish buildings from non-buildings, where SAVI less than 0.2 is represented as a building. The specific steps are: right-click in the [ProcessTree] dialog box and insert [Append New], change the name to [Fusion Vegetation Index Eigenvalue], then press [ok], then right-click and insert [Insert child], select [multiresolution segmentation], on the right In the [Activeclasses] of the table, check the categories to be configured, that is, check the categories to be classified as cultivated land, woodland, grassland, buildings, water bodies, and other land. Then select [Layer Values] under [Object features] in [Features], double-click [Mean] and [Standard deviation], then select [Create new 'ArithmeticFeature] under [Customized], enter the algorithm formula of NDVI, and click Execute . Repeat the above steps for the executed results until the input algorithm formula is modified to the SAVI and NDWI formulas, and the final output result is the ground object classification result that incorporates the vegetation feature index. Then perform an accuracy test on the classification results.

S4. Perform an accuracy test on the classification results of ground objects that incorporate the vegetation feature index. If the classification results are good, proceed to the next step:

Accuracy testing of ground object classification results that incorporate vegetation characteristic index is a necessary step to prove the reliability of the classification results. This invention randomly and evenly selects a total of 100 sample points of various land use types in the study area, and uses confusion matrix accuracy testing Method, use Overall Accuracy (OA) and Kappa coefficient to verify the accuracy of the classification results, where:

,

,

,

TP: Prediction is positive, actual is positive,

FN: Prediction is negative, actual is positive,

FP: Prediction is positive, actual is negative,

TN: Prediction is negative, actual is negative,

AA: average accuracy (average of accuracy for each category),

The Kappa coefficient after accuracy verification is greater than 0.6, which indicates that the classification result is good, and the feature classification results combined with the vegetation characteristic index can be used for the next step of water conservancy element identification.

S5. Distinguish water conservancy elements based on the classification results of ground objects, including hydraulic structures, rivers, reservoirs, and lakes. Use the aspect ratio and area index of the water body shape to determine whether the water body is a river, reservoir, or lake:

Based on the classification results of land objects integrated with vegetation characteristic index, the water body morphology is analyzed and water conservancy elements are identified, mainly including hydraulic structures, rivers, reservoirs, and lakes. The identification method for hydraulic buildings is as follows: extract the buildings from the feature classification. If the building is a regular rectangle and is beside the reservoir, it indicates that the building is a reservoir dam; there is an obvious water body in the middle of the river. The building is represented as a barrage. When identifying water bodies, the smallest enclosing rectangle of an irregular-shaped water body is regarded as the boundary of the water body. The length and width of the rectangle are regarded as the length and width of the water body. When the aspect ratio K1 of the water body is greater than 44, it means that the water body is a river; when the water body If the aspect ratio K2 of the water body is greater than 2.5 and the area S2 is greater than 0.37 hectares, and there are engineering dam buildings around the water body, the water body is a reservoir; when the aspect ratio K3 of the water body is greater than 2 and the area S3 is greater than 10.37 hectares, there are no engineering projects around the water body. The dam building is a lake. Specifically

K1=L1/B1,

K2=L2/B2,

K3=L3/B3,

S2=L2 B2,

S3=L3 B3,

In the formula, K1: the length-to-width ratio of the river,

K2: aspect ratio of the reservoir,

K3: Aspect ratio of the lake,

S2: area of reservoir (10 ⁴ m ² ),

S3: area of the lake (10 ⁴ m ² ),

L1: length of river (m),

L2: length of reservoir (m),

L3: length of lake (m),

B1: width of the river (m),

B2: width of reservoir (m),

B3: Width of lake (m).

The schematic diagram of the aspect ratio of rivers, reservoirs and lakes in this embodiment is shown in Figure 3, where L1=594m, B1=11m, K1=54; L2= 106m, B2=39m, K2=2.71, S2=0.41 10 ⁴ m ² ; L3=732m, B3=149m, K3=4.9, S3=10.9 10 ⁴ m ² . The results of water conservancy element discrimination based on the classification results of ground objects are shown in Figure 4.

Embodiment 2: A water conservancy digital twin base element implementation system based on satellite remote sensing, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the above embodiment is implemented. The implementation method of water conservancy digital twin base elements based on satellite remote sensing described in 1.

The above is a further description of the present invention in conjunction with the embodiments, and the protection scope of the present invention is not limited thereto.

Claims (8)

1. The implementation method of water conservancy digital twin base elements based on satellite remote sensing is characterized by including the following steps: S1. Obtain the required remote sensing images of the study area on the GEE platform and perform preprocessing; S2. Construct vegetation index characteristic values, including normalized vegetation index NDVI, soil-adjusted vegetation index SAVI and normalized difference water index NDWI; S3. Use recognition software to organically integrate vegetation index feature values and multi-scale segmentation methods to classify remote sensing images; S4. Perform an accuracy test on the classification results of ground objects that incorporate vegetation feature index. If the classification results are good, proceed to the next step; S5. Distinguish water conservancy elements based on the classification results of ground objects, including hydraulic structures, rivers, reservoirs, and lakes. Use the aspect ratio and area index of the water body shape to determine whether the water body is a river, reservoir, or lake. 2. The method for implementing water conservancy digital twin base elements based on satellite remote sensing according to claim 1, characterized in that the preprocessing in step S1 is to perform orthorectification and atmospheric correction on the image, and then perform shadow removal and cloud masking. Membrane operation, using the median filter method to obtain inter-annual composite images, thereby synthesizing all remote sensing images of the study year into one, and obtaining a remote sensing image that can clearly and completely display the surface coverage information of the study area. 3. The method for implementing water conservancy digital twin base elements based on satellite remote sensing according to claim 1, characterized in that the calculation method of the vegetation index characteristic value described in step S2 is as follows: , , , Among them, P _NIR : reflectivity in the near-infrared band, P _RED : Reflectivity in the infrared band, P _GREEN : reflectivity of green light band, L: Soil adjustment coefficient. 4. The method for implementing water conservancy digital twin base elements based on satellite remote sensing according to claim 1, characterized in that in step S3, the multi-scale segmentation method mainly divides the pixels in the image into cultivated land, woodland, grassland, buildings, Water bodies and other land uses; by setting the threshold of characteristic values, focus on distinguishing forest land, water bodies and buildings. The NDVI threshold is set to 0.46 to distinguish forest land from non-forest areas. NDVI greater than 0.46 is represented as forest land; NDWI is set When NDWI is 0.2, water bodies are extracted. Those with NDWI greater than 0.2 are represented as water bodies, including water bodies in urban areas and shallow water areas in rivers; SAVI is set to 0.2 to distinguish buildings and non-buildings, and those with SAVI less than 0.2 are represented as buildings. 5. The water conservancy digital twin base element implementation method based on satellite remote sensing according to claim 1, characterized in that, in step S4, the confusion matrix accuracy test method is used to verify the accuracy of the classification result using the overall accuracy OA and Kappa coefficient. in: , , , TP: Prediction is positive, actual is positive, FN: Prediction is negative, actual is positive, FP: Prediction is positive, actual is negative, TN: Prediction is negative, actual is negative, AA: average accuracy, The Kappa coefficient after accuracy verification is greater than 0.6, indicating a good classification result. 6. The method for realizing water conservancy digital twin base elements based on satellite remote sensing according to claim 1, characterized in that, in step S5, the method for identifying hydraulic buildings is: extracting the buildings in the classification of ground objects, If the building is a regular rectangle and is beside the reservoir, it indicates that the building is a reservoir dam; a building with an obvious water body in the middle of the river is indicated as a barrage. 7. The method for implementing water conservancy digital twin base elements based on satellite remote sensing according to claim 1, characterized in that, in step S5, when performing water body discrimination, the smallest enclosing rectangle of the irregular-shaped water body is regarded as the boundary of the water body, and the rectangle is The length and width are regarded as the length and width of the water body. When the aspect ratio K1 of the water body is greater than 44, it means that the water body is a river; when the aspect ratio K2 of the water body is greater than 2.5 and the area S2 is greater than 0.37 hectares, there is an engineering dam around the water body. The water body represented by the building is a reservoir; when the aspect ratio K3 of the water body is greater than 2 and the area S3 is greater than 10.37 hectares, the water body without engineering dam buildings around the water body is a lake. 8. A water conservancy digital twin base element implementation system based on satellite remote sensing, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The characteristic is that when the processor executes the program, it implements the following: The method for implementing the water conservancy digital twin base elements based on satellite remote sensing according to any one of claims 1 to 7.