SAR satellite remote sensing surface water extraction method for complex climate regions in africa

Abstract

The application discloses a SAR satellite remote sensing surface water extraction method for complex climate regions in Africa, and comprises the following steps: pre-processing satellite image data, and making sample point label data; making a shadow mask file to remove mountain shadows on the satellite image data; calculating a gray level co-occurrence matrix to generate texture features; establishing a multi-dimensional feature space to perform feature space screening; using two random forest classifiers to build a semi-supervised collaborative training model, and introducing unlabeled data to assist model training; and predicting the global surface water distribution of a study area. The application combines SAR remote sensing images with a semi-supervised collaborative random forest model, greatly reduces the labor and time cost of actual sample labeling, improves the surface water monitoring efficiency, and overcomes the problem of visible light remote sensing data loss in complex climate regions by using microwave remote sensing data, thereby providing more timely and accurate information for water resource management in African regions.

Description

Technical Field

[0001] The present invention relates to the technical field of surface water identification using microwave remote sensing technology, and in particular to a surface water extraction method using SAR satellite remote sensing in African regions with complex climates. Background Art

[0002] Satellite remote sensing technology has played a vital role in water resource monitoring due to its wide coverage, spatial continuity, and short revisit periods. However, optical sensors are susceptible to cloud cover, rainfall, fog, and other factors, making it difficult to provide continuous, high-quality, low-cloud imagery. In contrast, synthetic aperture radar (SAR) actively transmits microwave signals and receives reflected waves, enabling it to acquire high-resolution surface information under various meteorological conditions. Its all-weather, all-day, and high-penetration capabilities make it a crucial tool for routine water resource monitoring.

[0003] Common methods for extracting water bodies from SAR images include threshold segmentation, machine learning, and deep learning. Threshold segmentation primarily uses the difference in reflectance characteristics between the water body and the background to find the optimal threshold. However, due to the complex variations in water bodies across large areas, it is difficult to determine a uniform threshold. Furthermore, this method ignores factors such as object texture, shape, and water body proportion, which can easily lead to incomplete water body extraction results and misclassification. With the rapid development of machine learning and deep learning-based methods, many image classification methods have been effectively applied to water body extraction and mapping. While deep learning methods can effectively extract semantic features from remote sensing imagery, they require a large amount of labeled data. Collecting labeled data for large areas requires significant time and effort, making it difficult to obtain sample data sufficient for complex deep learning models. Machine learning methods can extract surface water using a small amount of sample data, using manually designed classification features tailored to the characteristics of individual objects. To address these issues, many researchers have proposed semi-supervised learning methods, which use a small amount of labeled data and a large amount of unlabeled data for training. This approach can both compensate for the lack of real sample data and effectively utilize unlabeled data to improve classification accuracy. Among them, the divergence-based semi-supervised learning method can obtain results that are superior to other semi-supervised classification algorithms by using multiple classifiers and making full use of unlabeled data.

[0004] By combining SAR satellite imagery with a semi-supervised collaborative random forest model and incorporating environmental features such as geography and texture specific to Africa's complex climate, surface water resource monitoring and analysis can be performed more efficiently, providing a scientific basis for local ecological protection, management, and sustainable development. Summary of the Invention

[0005] The purpose of the present invention is to provide a SAR satellite remote sensing surface water extraction method for African regions with complex climates. The method combines SAR satellite imagery with a semi-supervised collaborative random forest model to achieve accurate extraction of surface water in African regions with complex climates.

[0006] To achieve the above functions, the present invention designs a SAR satellite remote sensing surface water extraction method for the African climate complex region. For the target African climate complex region, the following steps A to G are performed to complete the surface water distribution prediction:

[0007] Step A: Acquire GRD image data from the Sentinel-1 satellite for regions with complex climates in Africa, preprocess the Sentinel-1 satellite image data, and convert it into a 10m resolution backscatter coefficient map before proceeding to Step B.

[0008] Step B: Use ArcGIS software to visually interpret the image data obtained in step A, generate sample point label data for water bodies and non-water bodies, and proceed to step C;

[0009] Step C: Use SRTM DEM software to create a shadow mask file with a slope greater than 5°, remove the mountain shadow on the image data obtained in step B, and proceed to step D;

[0010] Step D: Calculate the gray-level co-occurrence matrix using the image data obtained in step C to capture the texture characteristics of the African climate complex area, and then proceed to step E;

[0011] Step E: Combine the texture features obtained in step D, the polarization features of the Sentinel-1 satellite image data, and the water index and terrain features of the complex climate region in Africa to establish a multidimensional feature space. Use the Boruta algorithm to screen the feature space and use the shadow features to screen important features, and then proceed to step F.

[0012] Step F: Divide the filtered important features into two different feature subsets, input them into two random forest classifiers respectively, use the random forest classifier to build a semi-supervised collaborative training model, and introduce unlabeled sample data from the complex climate region of Africa to assist in model training, and then proceed to step G;

[0013] Step G: Train the semi-supervised collaborative training model constructed in step F to obtain a trained semi-supervised collaborative training model, and predict the surface water distribution in the African climate complex area based on the semi-supervised collaborative training model.

[0014] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0015] 1. This paper combines SAR satellite imagery with a semi-supervised collaborative random forest model to propose a surface water extraction method that integrates multidimensional feature space. This method addresses the challenges of surface water monitoring in Africa, which faces complex climate and geographical environments and a lack of visible light imagery.

[0016] 2. This paper uses Sentinel-1 satellite image data to quickly obtain surface water distribution in large-scale climate-complex areas, greatly reducing the manpower and time costs of actual sample labeling, overcoming the problem of small sample size, and improving the efficiency of surface water monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a method for extracting surface water from SAR satellite remote sensing in African regions with complex climates, provided according to an embodiment of the present invention;

[0018] Figure 2 2 is a structural diagram of a semi-supervised collaborative random forest model provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0020] The embodiment of the present invention provides a SAR satellite remote sensing surface water extraction method for African climate complex areas, referring to Figure 1 , for the target African region with complex climate, perform the following steps A to G to complete the prediction of surface water distribution:

[0021] Step A: Acquire Sentinel-1 IW-mode GRD (Ground Range Coded) imagery data for climate-complex regions in Africa, using VH+V polarization. Preprocess the Sentinel-1 imagery data and convert it into a 10m resolution backscatter coefficient map before proceeding to Step B.

[0022] In step A, the Sentinel-1 image data is filtered using a modified Lee filter for coherent wave filtering, further improving image quality and suppressing inherent speckle noise. Preprocessing of Sentinel-1 image data includes orbit correction, thermal noise removal, radiometric calibration, speckle filtering, terrain correction, and decibelization to remove inherent image noise.

[0023] Step B: Use ArcGIS software to visually interpret the image data obtained in step A, generate sample point label data for water bodies and non-water bodies, and proceed to step C;

[0024] Step C: Use SRTM DEM software to create a shadow mask file with a slope greater than 5°, remove the mountain shadow on the image data obtained in step B to eliminate the influence of radiation distortion of the SAR satellite image data, and then proceed to step D;

[0025] Step D: Calculate the gray-level co-occurrence matrix using the image data obtained in step C to capture the texture characteristics of the African climate complex area, and then proceed to step E;

[0026] Texture features include: mean, variance, correlation, homogeneity, contrast, entropy, dissimilarity and second moment.

[0027] Step E: Combine the texture features obtained in step D, the polarization features of the Sentinel-1 satellite image data, and the water index and terrain characteristics of the complex climate region in Africa to establish a multidimensional feature space, enrich the feature space of the SAR satellite image, and use the Boruta algorithm to screen the feature space. Use shadow features to screen important features, reduce model training costs, and improve training accuracy, and then proceed to step F.

[0028] Step F: Divide the selected important features into two different feature subsets to increase the differences between the features in the feature subsets and obtain a more complete water body distribution. The two different feature subsets are input into two random forest classifiers respectively. A semi-supervised collaborative training model is built using the random forest classifiers. Unlabeled sample data from the complex African climate region is introduced to assist in model training, and then proceed to step G.

[0029] Reference Figure 2 The semi-supervised collaborative training model in step F includes two random forest classifiers. The specific steps of building and training the semi-supervised collaborative training model are as follows:

[0030] Step F1: Randomly divide the training dataset into two groups of labeled sample sets L1 and L2 of equal size. Each group is used to train an independent random forest classifier, denoted as random forest classifier RF1 and random forest classifier RF2 respectively.

[0031] Step F2: Introduce a randomly generated unlabeled sample set U, label the elements in the unlabeled sample set U through the trained random forest classifier RF1, and expand the labeled data into the labeled sample set L2 as a new training set ; Use the trained random forest classifier RF2 to label the elements in the unlabeled sample set U, and expand the labeled data to the labeled sample set L1 as a new training set ;

[0032] Step F3: Use the newly constructed training set to retrain the random forest classifiers RF1 and RF2, and repeat this process until all unlabeled samples are consistently classified as water bodies or non-water bodies by the random forest classifiers RF1 and RF2. The unlabeled sample learning process ends.

[0033] Step F4: After the iterative training is completed, if there are still differences in the classification results of the random forest classifier RF1 and the random forest classifier RF2, the result with the larger weight ratio is selected as the final classification result of the sample based on the weights allocated in the random forest classifier RF1 and the random forest classifier RF2.

[0034] In step F, the randomly generated unlabeled sample set U is incremented by 1000 to find the optimal sample threshold in the constructed semi-supervised collaborative training model, and the training is repeated until a certain evaluation indicator converges.

[0035] In step F, a semi-supervised collaborative training model is constructed based on a random forest classifier. Each random forest classifier is set with 100 decision trees, and a data set for each tree is randomly generated by Bootstrap. During the construction of each tree, the number of features considered for each node is the square root of all numbers, which helps to reduce the correlation of the model and improve the generalization ability of the model. In one embodiment, 50 iterations are set, and the threshold is set to 0.7.

[0036] Step G: Train the semi-supervised collaborative training model constructed in step F to obtain a trained semi-supervised collaborative training model, and predict the surface water distribution in the African climate complex area based on the semi-supervised collaborative training model.

[0037] The embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the scope of knowledge possessed by ordinary technicians in this field without departing from the spirit of the present invention.

Claims

1. A SAR satellite remote sensing surface water extraction method for African regions with complex climates, characterized by: For the target African region with complex climate, perform the following steps A to G to complete the prediction of surface water distribution: Step A: Acquire GRD image data from the Sentinel-1 satellite for regions with complex climates in Africa, preprocess the Sentinel-1 satellite image data, and convert it into a 10m resolution backscatter coefficient map before proceeding to Step B. Step B: Use ArcGIS software to visually interpret the image data obtained in step A, generate sample point label data for water bodies and non-water bodies, and proceed to step C; Step C: Use SRTM DEM software to create a shadow mask file with a slope greater than 5°, remove the mountain shadow on the image data obtained in step B, and proceed to step D; Step D: Calculate the gray-level co-occurrence matrix using the image data obtained in step C to capture the texture characteristics of the African climate complex area, and then proceed to step E; Step E: Combine the texture features obtained in step D, the polarization features of the Sentinel-1 satellite image data, and the water index and terrain features of the complex climate region in Africa to establish a multidimensional feature space. Use the Boruta algorithm to screen the feature space and use the shadow features to screen important features, and then proceed to step F. Step F: Divide the filtered important features into two different feature subsets, input them into two random forest classifiers respectively, use the random forest classifier to build a semi-supervised collaborative training model, and introduce unlabeled sample data from the complex climate region of Africa to assist in model training, and then proceed to step G; Step G: Train the semi-supervised collaborative training model constructed in step F to obtain a trained semi-supervised collaborative training model, and predict the surface water distribution in the African climate complex area based on the semi-supervised collaborative training model.

2. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 1 is characterized in that: In step A, the improved Lee filter is used to perform coherent wave filtering on the image data of the Sentinel-1 satellite.

3. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 1 is characterized in that: The preprocessing of the Sentinel-1 satellite image data in step A includes: orbit correction, thermal noise removal, radiometric calibration, speckle filtering, terrain correction, and decibel conversion.

4. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 1 is characterized in that: The texture features in step D include: mean, variance, correlation, cooperativity, contrast, information entropy, dissimilarity and second-order moment.

5. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 1 is characterized in that: The semi-supervised collaborative training model in step F includes two random forest classifiers. The specific steps for building and training the semi-supervised collaborative training model are as follows: Step F1: Randomly divide the training dataset into two groups of labeled sample sets L1 and L2 of equal size. Each group is used to train an independent random forest classifier, denoted as random forest classifier RF1 and random forest classifier RF2 respectively. Step F2: Introduce a randomly generated unlabeled sample set U, label the elements in the unlabeled sample set U through the trained random forest classifier RF1, and expand the labeled data into the labeled sample set L2 as a new training set ; Use the trained random forest classifier RF2 to label the elements in the unlabeled sample set U, and expand the labeled data to the labeled sample set L1 as a new training set ; Step F3: Use the newly constructed training set to retrain the random forest classifiers RF1 and RF2, and repeat this process until all unlabeled samples are consistently classified as water bodies or non-water bodies by the random forest classifiers RF1 and RF2. The unlabeled sample learning process ends. Step F4: After the iterative training is completed, if there are still differences in the classification results of the random forest classifier RF1 and the random forest classifier RF2, the result with the larger weight ratio is selected as the final classification result of the sample based on the weights allocated in the random forest classifier RF1 and the random forest classifier RF2.

6. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 5 is characterized in that: In step F, the randomly generated unlabeled sample set U is incremented by 1000 to find the optimal sample threshold in the constructed semi-supervised collaborative training model.

7. The SAR satellite remote sensing surface water extraction method for African climate complex areas according to claim 5 is characterized in that: In step F, a semi-supervised collaborative training model is constructed based on the random forest classifier. Each random forest classifier sets 100 decision trees, and the data set of each tree is randomly generated by Bootstrap. During the construction of each tree, the number of features considered at each node is the square root of all numbers.

Images