Methods and apparatus for identifying mountain lakes based on environmental characteristics

Abstract

The application discloses a mountainous lake recognition method and device based on environmental features, and the method comprises the following steps: acquiring a plurality of remote sensing images of a target water body; calculating a comprehensive water body index based on the remote sensing images to enhance the spectral characteristics of the water body and obtain a spectral enhancement image; calculating terrain data based on the spectral enhancement image, calculating a buffer zone of the target water body according to the terrain data, and extracting remote sensing images of the buffer zone; inputting the remote sensing images of the target water body, the remote sensing images of the buffer zone and the terrain data into a pre-trained water body classification model to obtain a class recognition result output by the water body classification model. The technical problem that the existing method is limited by the information representation ability of traditional local spectral and texture features and is difficult to realize accurate identification of small water bodies in alpine regions is solved, so that the remote sensing recognition precision of lakes in alpine regions is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing image processing, and particularly relates to a mountainous lake recognition method and device based on environmental features. BACKGROUND

[0002] Remote sensing technology has become the dominant means for lake monitoring in alpine regions. However, remote sensing recognition of lakes in alpine regions still faces great difficulties. This difficulty mainly comes from the following aspects:

[0003] Firstly, the unique geographical environment of lakes in alpine regions brings great challenges to data acquisition. The lakes in these regions are covered by ice and snow for most of the annual cycle, and the high-altitude regions generally have more clouds, which greatly limits the availability of high-quality remote sensing images. The widespread existence of clouds and snow not only reduces the opportunity for data acquisition, but also significantly increases the complexity of image processing, making more stringent requirements for accurate remote sensing recognition of lakes in alpine regions.

[0004] Secondly, the complex topography and water body development characteristics in alpine regions further exacerbate the difficulty of recognition. The terrain in these regions is highly undulating, resulting in the existence of a large number of shadows. Due to the similarity in spectral characteristics between water bodies and shadows, especially in the visible and infrared bands, it is difficult to effectively distinguish water bodies and shadows in remote sensing images relying solely on spectral features. More critically, small water bodies are developed in alpine regions, which are often greatly affected by mixed pixels in remote sensing images and are easily overlooked or misclassified. This feature, combined with complex terrain and lighting conditions, makes high-precision recognition more difficult.

[0005] In order to overcome these challenges, existing technologies introduce local texture features to exclude shadow interference, but this method faces inherent limitations. Texture information is highly dependent on the lighting conditions in the image, and shadow areas are exactly due to insufficient lighting, which means that in shadow areas, texture features can be weakened or even disappear. In addition, water bodies also exhibit relatively smooth texture features due to their unique reflection characteristics, further blurring the boundaries between water bodies and shadows. For small water bodies, their weak feature information in the image makes it even more difficult to distinguish.

[0006] Currently, widely used methods in the field of lake extraction, such as index methods based on local spectral and texture information and machine learning methods, face serious limitations in alpine regions. These methods often struggle to accurately distinguish water bodies, shadows, and seasonal meltwater in complex terrain, leading to a large number of shadows and seasonal meltwater being incorrectly classified as lakes. This not only reduces the accuracy of the extraction results, but also often requires tedious post-processing work, greatly increasing the difficulty and workload of remote sensing lake monitoring in alpine regions.

[0007] In recent years, deep learning semantic segmentation methods, especially the state-of-the-art Transformer models, have shown considerable potential in eliminating shadow interference. Deep learning can extract multi-scale environmental information through hierarchical convolution, while the Transformer's self-attention mechanism enables it to capture long-range dependencies, i.e., broader contextual information. These characteristics make deep learning models perform well in recognizing large bodies of water with regular shapes and clear shorelines, and can effectively distinguish them from large areas of shadow. However, because deep learning methods focus more on global features than local details, the precision of lake boundary extraction is insufficient. For small lakes commonly found in high-altitude and cold regions, which occupy fewer pixels and have less prominent features in remote sensing images, the recognition performance of deep learning models is often unsatisfactory.

[0008] In summary, despite significant progress made in water body identification using machine learning and deep learning technologies, accurate identification of lakes in high-altitude and cold regions remains one of the most challenging tasks in remote sensing science. The limitations of existing technologies primarily lie in their inadequacies in feature extraction and representation. Therefore, improving the accuracy of feature inputs and enhancing the representation of water body information are key issues in overcoming the bottleneck in lake identification accuracy in high-altitude and cold regions.

[0009] In view of this, a method and device for identifying mountain lakes based on environmental characteristics are provided to solve the technical problem that existing methods are limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in high-altitude and cold regions. Summary of the Invention

[0010] To address this, embodiments of the present invention provide a method and apparatus for identifying lakes in mountainous areas based on environmental features, aiming to solve the technical problem that existing methods are limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in high-altitude and cold regions, thereby improving the remote sensing identification accuracy of lakes in high-altitude and cold regions.

[0011] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0012] This invention provides a method for identifying mountain lakes based on environmental features, the method comprising:

[0013] Acquire multiple remote sensing images of the target water body;

[0014] Based on the remote sensing image, a comprehensive water index is calculated to enhance the spectral characteristics of the water body and obtain a spectrally enhanced image.

[0015] Topographic data is calculated based on the spectral enhanced image, the buffer zone of the target water body is calculated based on the topographic data, and remote sensing image of the buffer zone is extracted.

[0016] The remote sensing images of the target water body, the remote sensing images of the buffer zone, and the terrain data are input into a pre-trained water body classification model to obtain the category recognition result output by the water body classification model.

[0017] The water body classification model is obtained by training a pre-built convolutional neural network using topographic data, historical image data, and category labels of water body samples.

[0018] In some embodiments, a comprehensive water body index is calculated based on the remote sensing image to enhance the spectral characteristics of the water body, specifically including:

[0019] Select images from the acquired remote sensing images during periods without snow or ice cover to obtain the target image;

[0020] Calculate the water index in multiple dimensions of the target image;

[0021] By converting the RGB to HSV color space, the water body indices of different dimensions are combined to generate a comprehensive water body index, thereby enhancing the spectral characteristics of the water body.

[0022] In some embodiments, enhancing the spectral characteristics of the water body further includes:

[0023] Water areas are extracted from the spectral enhanced image using a remote sensing index algorithm or a mask generation algorithm.

[0024] In some embodiments, topographic data is calculated based on the spectral enhancement image, a buffer zone of the target water body is calculated based on the topographic data, and remote sensing imagery of the buffer zone is extracted, specifically including:

[0025] Water bodies are extracted from the spectral enhanced image, and water units are identified and marked within the water bodies.

[0026] The spectral enhanced images from different times are stitched together to obtain a composite image;

[0027] Calculate the terrain data of the synthetic image and fuse the obtained terrain data with the data of the synthetic image;

[0028] Construct buffer zones for each of the aforementioned water body units;

[0029] Extract the water environment image of the buffer zone and resample the water environment image of the buffer zone to obtain water environment feature parameters.

[0030] In some embodiments, calculating the terrain data of the synthetic image specifically includes:

[0031] Calculate the slope, elevation, and azimuth of each pixel in the synthesized image.

[0032] In some embodiments, a pre-constructed convolutional neural network is trained using topographic data, historical image data, and category labels of water body samples to obtain the water body classification model, specifically including:

[0033] Obtain topographic data, historical imagery data, and category labels for water samples to construct a dataset;

[0034] The dataset is divided into a training set, a validation set, and a test set;

[0035] Based on a pre-defined loss function, the pre-built convolutional neural network is trained using the training set to obtain an initial model;

[0036] The initial model is iteratively trained using the validation set to obtain the optimized water body classification model.

[0037] In some embodiments, the pre-built network architecture of the convolutional neural network includes:

[0038] Convolutional layers are used to extract spatial features from the input image and enhance the network’s ability to distinguish water bodies, rivers and non-water bodies through layer-by-layer feature extraction.

[0039] Pooling layers are used to reduce the size of feature maps while preserving the main features of the input image;

[0040] Activation functions are used after each convolutional layer to capture additional features;

[0041] The classification layer, located at the end of the network, uses a fully connected layer to fuse features, and finally outputs the classification probabilities for lakes, rivers, and non-aquatic bodies through the softmax function.

[0042] The present invention also provides a mountain lake identification device based on environmental characteristics, the device comprising:

[0043] The remote sensing image acquisition unit is used to acquire multiple remote sensing images of the target water body.

[0044] The spectral information enhancement unit is used to calculate the comprehensive water body index based on the remote sensing image to enhance the spectral characteristics of the water body and obtain the spectral enhanced image.

[0045] An environmental information extraction unit is used to calculate terrain data based on the spectral enhancement image, calculate the buffer zone of the target water body according to the terrain data, and extract remote sensing images of the buffer zone.

[0046] The identification result generation unit is used to input the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the terrain data into a pre-trained water body classification model to obtain the category identification result output by the water body classification model.

[0047] The water body classification model is obtained by training a pre-built convolutional neural network using topographic data, historical image data, and category labels of water body samples.

[0048] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method described above.

[0049] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0050] The present invention provides a method and apparatus for identifying mountain lakes based on environmental features. This method acquires multiple remote sensing images of the target water body; calculates a comprehensive water body index based on the remote sensing images to enhance the spectral characteristics of the water body, resulting in a spectrally enhanced image; calculates topographic data based on the spectrally enhanced image; calculates a buffer zone for the target water body based on the topographic data; and extracts remote sensing images of the buffer zone. The remote sensing images of the target water body, the buffer zone, and the topographic data are then input into a pre-trained water classification model to obtain the category identification result output by the model. This solves the technical problem of existing methods being limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in high-altitude and cold regions, thereby improving the remote sensing identification accuracy of lakes in high-altitude and cold regions. Attached Figure Description

[0051] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0052] The structures, proportions, sizes, etc. illustrated in this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0053] Figure 1 This is one of the flowcharts for the mountain lake identification method based on environmental features provided by the present invention;

[0054] Figure 2 The second flowchart illustrates the mountain lake identification method based on environmental features provided by this invention.

[0055] Figure 3 The third flowchart of the mountain lake identification method based on environmental features provided by the present invention;

[0056] Figure 4 The fourth flowchart of the mountain lake identification method based on environmental features provided by the present invention;

[0057] Figure 5 A schematic diagram illustrating the principle of water body unit extraction;

[0058] Figure 6 A schematic diagram illustrating the principle affecting resolution calculation;

[0059] Figure 7 Diagram of ConvNet architecture;

[0060] Figure 8 A flowchart of the water body classification model training process;

[0061] Figure 9 A schematic diagram of the structure of the mountain lake identification device based on environmental features provided by the present invention;

[0062] Figure 10 This is a structural block diagram of a computer device provided by the present invention. Detailed Implementation

[0063] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Currently, the main methods widely used for lake identification in high-altitude and cold regions include the exponential method, machine learning, and deep learning. However, these methods are limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in these regions. The core solution lies in how to effectively capture and utilize the unique characteristics of water bodies in complex, cold environments. Specifically, the following aspects need to be addressed:

[0065] Refinement of Feature Representation: Traditional methods often directly use raw spectra or simple indices as input, which is insufficient to fully express the characteristics of water bodies in the complex environment of high-altitude areas. This method aims to enhance the spectral characteristics of water bodies through multi-index synthesis, preserving richer texture information to distinguish water bodies from similar land features such as shadows;

[0066] Effective use of contextual information: While conventional deep learning models can learn complex features, they often lack explicit consideration of the local environment. This method aims to establish a water body buffer zone to explicitly introduce environmental information surrounding the water body, providing the model with more valuable contextual cues to aid in the identification of small water bodies;

[0067] Scale adaptability: Traditional machine learning methods perform inconsistently when dealing with water bodies of different scales, especially exhibiting significant uncertainty in the identification of small water bodies. This method aims to improve the model's ability to identify water bodies of various scales, particularly enhancing the detection accuracy of small water bodies, by combining enhanced spectral information and local environmental features, and designing the level of detail of input features according to the size of the water body.

[0068] Anti-interference capability: The complex terrain and lighting conditions in high-altitude areas often lead traditional methods to mistakenly identify shadows or seasonal meltwater as lakes. This method improves the model's anti-interference capability and recognition accuracy by comprehensively considering multi-source features and environmental information (multi-environmental features: shape features, shoreline features, slope around the lake, etc.; multi-spectral features: water volume index, vegetation index, etc.).

[0069] In view of this, the present invention provides a method for identifying mountain lakes based on environmental characteristics. By enhancing the water body's spectral and environmental information, the method improves the water body's characterization ability, achieves high-precision water body identification, and promotes the application of remote sensing science in extreme environments. This provides more reliable data support for water resource monitoring, ecological environment assessment, and climate change research in high-altitude and cold regions.

[0070] In one specific implementation, such as Figure 1 As shown, the mountain lake identification method based on environmental features provided by this invention includes the following steps:

[0071] S110: Acquire multiple remote sensing images of the target water body;

[0072] S120: Based on the remote sensing image, calculate the comprehensive water index to enhance the spectral characteristics of the water body and obtain a spectrally enhanced image; in a specific application scenario, such as Figure 2 As shown, when enhancing the spectral information of water bodies, images from periods without ice or snow cover are first selected, and water body indices such as MNDWI, NDVI, and NDWI are calculated. Then, by converting the RGB to HSV color space, different water body indices are combined to synthesize multiple new comprehensive water body indices to enhance the spectral characteristics of the water bodies.

[0073] S130: Calculate topographic data based on the spectral enhanced image, calculate the buffer zone of the target water body based on the topographic data, and extract remote sensing imagery of the buffer zone; still using Figure 2 Taking the scenario shown as an example, in the process of determining the range of water body environmental information, the water body range is first extracted based on the spectral information of remote sensing images, and a water body buffer is established; at the same time, multi-temporal images are synthesized and terrain parameters are calculated to construct multi-source feature parameters; then, features within the water body buffer range are extracted, and resampling is performed according to the size of the water body to ensure that the level of detail of the feature information is appropriate.

[0074] S140: Input the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the topographic data into a pre-trained water body classification model to obtain the category recognition result output by the water body classification model; wherein, the water body classification model is obtained by training a pre-constructed convolutional neural network using topographic data, historical image data, and category labels of the water body samples. Figure 2 As shown, in this specific application scenario, during the training process of the water body classification model, three types of labeled samples—lakes, rivers, and non-water bodies—are first constructed. Based on the ConvNet classification method and synthesized high-quality multispectral images, comprehensive water body indices, and terrain parameters, a water body classification model is constructed to classify water body units, ultimately achieving accurate identification of alpine lakes.

[0075] Thus, the method provided by this invention breaks through the limitations of traditional machine learning and deep learning methods that rely solely on enhancing spectral information to improve water body representation capabilities. By establishing a water body buffer and integrating multiple water body indices, it enhances spectral information and local environmental features, improves the utilization efficiency of contextual information, and is more suitable for small water body surface tasks in complex environments in high-altitude and cold regions.

[0076] In some embodiments, such as Figure 3 As shown, based on the remote sensing image, a comprehensive water index is calculated to enhance the spectral characteristics of the water body, specifically including the following steps:

[0077] S310: Select images from the acquired remote sensing images during periods without snow or ice cover to obtain the target image;

[0078] S320: Calculate water quality indices in multiple dimensions of the target image; specifically, this includes calculating the following indices:

[0079] a) MNDWI (Modified Normalized Difference Water Index):

[0080] MNDWI=(Green-SWI R) / (Green+SWIR)

[0081] Among them, Green refers to the green light band, and SWI R refers to the short-wave infrared band;

[0082] b) NDVI (Normalized Difference Vegetation Index):

[0083] NDVI = (NIR - Red) / (NIR + Red)

[0084] Wherein, NIR stands for near-infrared band and Red stands for red band;

[0085] c) NDWI (Normalized Difference Water Index):

[0086] NDWI=(Green-NIR) / (Green+NIR)

[0087] S330: By converting from RGB to HSV color space, the water body indices of different dimensions are combined to generate a comprehensive water body index, thereby enhancing the spectral characteristics of the water body. Specifically, the calculated MNDWI, NDVI, and NDWI are assigned to the R, G, and B channels respectively, resulting in a false-color composite image in which the water body displays a specific color. Then, the RGB color space is converted to the HSV (Hue, Saturation, Value) color space to form a new index. The HSV color space is closer to how humans perceive color, helping to better distinguish and enhance specific features. This comprehensive index (Hue, Saturation, Value) integrates the information from MNDWI, NDVI, and NDWI, while utilizing the enhancement effect brought by the color space conversion.

[0088] Furthermore, enhancing the spectral characteristics of the water body, and subsequently including:

[0089] Water bodies are extracted from the spectral enhanced image using remote sensing index algorithms or mask generation algorithms. That is, water bodies are initially identified and extracted through pixel-level analysis, including:

[0090] Remote sensing index calculation: Commonly used water body indices (such as Normalized Difference Water Index (NDWI), modified NDWI, ratio indices, etc.) are calculated to enhance the contrast between water bodies and surrounding features. Index methods distinguish between water bodies and non-water bodies through reflectance or reflectance calculations.

[0091] Mask generation: Water body masks can be generated using either exponential or machine learning methods, the choice depending on the size and complexity of the region. Exponential methods are suitable for small regions with distinct spectral characteristics, offering high efficiency and ease of implementation; while machine learning methods are applicable to large-scale or complex environments, handling more features and background interference, providing higher extraction accuracy, generating binarized water body masks, and marking the water area, thereby further helping to filter out background noise or irrelevant information.

[0092] It should be understood that in step S310, selecting suitable remote sensing imagery is fundamental to ensuring the accuracy of water body identification. Image sources can include images acquired from various remote sensing satellites, such as Landsat and Sentinel. Selecting higher resolution imagery helps to more accurately identify the boundaries of small and complex lakes. Image types can be selected based on requirements, such as optical remote sensing imagery and synthetic aperture radar (SAR) imagery. Optical imagery is suitable for spectral feature analysis of water bodies, while SAR imagery is insensitive to cloud cover and weather conditions and can be used to supplement the shortcomings of optical imagery. When selecting the time phase, imagery from months without ice or snow interference should be chosen, as ice and snow cover can obscure the true boundaries of water bodies and affect their spectral characteristics.

[0093] After selecting an image, preprocessing can be performed to improve image quality, eliminate interference that could affect analysis, and ensure data consistency and accuracy. Preprocessing can include the following aspects:

[0094] Geometric correction: Geometric correction is performed on the image to eliminate geometric distortions caused by factors such as the image acquisition angle or the curvature of the earth. Geometric correction ensures that the ground features in the image are consistent with the actual ground position.

[0095] Atmospheric correction: Atmospheric correction is applied to images to reduce the impact of atmospheric scattering and absorption on spectral information. Atmospheric correction can improve the recognizability of water bodies and other land features.

[0096] Radiometric correction: Radiometric correction is performed on images to standardize the radiometric characteristics of images acquired by different sensors or at different times. Radiometric correction helps ensure data consistency between different images.

[0097] In some embodiments, such as Figure 4 As shown, the method for calculating topographic data based on the spectral enhanced image, calculating the buffer zone of the target water body based on the topographic data, and extracting remote sensing images of the buffer zone specifically includes the following steps:

[0098] S410: Extract water body regions from the spectral enhanced image, and identify and mark water body units within the water body regions;

[0099] S420: Stitch together spectral enhanced images from different times to obtain a composite image;

[0100] S430: Calculate the terrain data of the synthetic image (e.g., calculate the slope information, elevation and azimuth of each pixel in the synthetic image), and fuse the obtained terrain data with the data of the synthetic image;

[0101] S440: Construct buffer zones for each of the aforementioned water body units;

[0102] S450: Extract the water environment image of the buffer zone and resample the water environment image of the buffer zone to obtain water environment characteristic parameters.

[0103] exist Figure 2 In the specific application scenario shown, the extraction of water environment information in alpine lake identification involves several key steps to accurately extract the water body and analyze its environmental characteristics. These steps include multi-temporal image synthesis, terrain parameter calculation, and water environment image extraction and image resolution calculation. Specifically, water environment information extraction includes the following steps:

[0104] S1: Water body unit extraction, such as Figure 5 As shown, the initially extracted water body regions are further processed to identify and separate independent water body units. Water body unit extraction specifically includes connectivity analysis, morphological processing, and water body unit labeling. During connectivity analysis, the binarized water body mask is subjected to connectivity analysis to identify continuous water body regions. This step helps identify interconnected water body regions and eliminate small noise areas. During morphological processing, morphological operations (such as dilation, erosion, opening and closing operations) are applied to optimize water body boundaries and repair missing or noisy boundaries. Morphological processing helps improve the smoothness and continuity of water body boundaries. During water body unit labeling, connected water body regions are labeled to distinguish different water body units (such as lakes, rivers, non-water bodies, etc.). Each water body unit can be assigned a unique identifier to form training samples for subsequent model building.

[0105] S2: Multi-temporal image synthesis improves image clarity and accuracy by synthesizing multi-temporal images, reduces interference from clouds, snow, ice, and shadows, and provides a better foundation for water body extraction and environmental analysis. Multi-temporal image compositing includes image selection, temporal selection (selecting image periods without cloud or snow coverage to improve image sharpness), cloud filtering and cloud masking, cloud detection (using cloud detection algorithms such as band-based cloud detection and cloud mask generation methods to identify cloud and cloud shadow regions in the image; common methods include threshold-based cloud detection, machine learning methods, or existing cloud mask data), cloud masking (generating cloud masks to mask cloud and cloud shadow regions in the image for subsequent processing. Cloud masks can be synthesized from the cloud detection results of multi-temporal images to ensure the most comprehensive masking of cloud areas), image compositing, image stitching (stitching images from different times into a complete large-scale image. Image fusion techniques (such as pixel-based fusion and feature-based fusion) are used to synthesize a clear image and reduce interference from clouds, snow, and ice), and sharpness enhancement (enhancing the synthesized image, such as contrast adjustment and color enhancement, to improve the visibility of environmental details).

[0106] S3: Topographic parameter calculation. This involves calculating topographic parameters (such as slope) to aid in distinguishing between water bodies and mountain shadows, and to provide crucial information for identifying lakes and rivers. Topographic parameter calculation mainly includes the following steps:

[0107] Digital Elevation Model (DEM) Processing:

[0108] DEM Acquisition: Acquire high-resolution DEM data to provide surface elevation information. DEM data can be obtained from remote sensing satellites (such as SRTM and ASTER) or ground surveys.

[0109] DEM preprocessing: Correcting and filtering DEM data to eliminate noise and errors.

[0110] Terrain parameter calculation:

[0111] Slope calculation: Calculates the slope of each pixel in the image. Slope information helps distinguish between mountain shadows and bodies of water. Steeper slopes usually correspond to mountains, while flatter slopes may correspond to bodies of water.

[0112]

[0113] Its grid cell size is Δx and Δy, and its elevation value is z.

[0114] Other terrain features: Other terrain parameters (such as elevation z and azimuth) can be calculated as a supplement to improve the ability to distinguish between water bodies and mountains.

[0115]

[0116] Terrain parameter fusion:

[0117] The calculated terrain parameters (such as slope) are fused with remote sensing image data to enhance the identification of water bodies and mountains.

[0118] S4: Water Environment Image Extraction and Resolution Calculation. In high-altitude lake identification, water environment image extraction and resolution calculation at appropriate scales are crucial steps, such as... Figure 6 As shown, the core of this step is to collect and process aquatic environmental information based on the scale of the water body unit, ensuring that high-precision environmental features are extracted from small lakes while summarizing unnecessary details from large lakes. This differentiated processing optimizes the accuracy and efficiency of lake identification. Specifically, it includes the following steps:

[0119] Water body unit buffer establishment: First, the buffer refers to establishing a 20-pixel buffer for each water body unit. The buffer is used to extract environmental information around the water body unit and reduce boundary effects.

[0120] Buffer extraction: Use Geographic Information System (GIS) tools or programming tools (such as the geopandas library in Python) to create and extract buffers.

[0121] Image extraction and resampling: Extract remote sensing image data from the buffer, including water bodies and their surrounding environment to provide contextual information; extract DEM data from the buffer and calculate relevant topographic parameters (slope, azimuth).

[0122] Image resolution adjustment: The extracted buffer images and DEM data are resampled to a uniform resolution, typically 515x515 pixels. This step is to standardize the spatial resolution of the data and ensure consistency across different water body units. An appropriate resampling method (such as bilinear interpolation or cubic convolution interpolation) is selected to maintain the spatial accuracy of the image and DEM data.

[0123] In some embodiments, a pre-constructed convolutional neural network is trained using topographic data, historical image data, and category labels of water body samples to obtain the water body classification model, specifically including:

[0124] Obtain topographic data, historical imagery data, and category labels for water samples to construct a dataset;

[0125] The dataset is divided into a training set, a validation set, and a test set;

[0126] Based on a pre-defined loss function, the pre-built convolutional neural network is trained using the training set to obtain an initial model;

[0127] The initial model is iteratively trained using the validation set to obtain the optimized water body classification model.

[0128] The pre-built convolutional neural network architecture includes:

[0129] Convolutional layers are used to extract spatial features from the input image and enhance the network’s ability to distinguish water bodies, rivers and non-water bodies through layer-by-layer feature extraction.

[0130] Pooling layers are used to reduce the size of feature maps while preserving the main features of the input image;

[0131] Activation functions are used after each convolutional layer to capture additional features;

[0132] The classification layer, located at the end of the network, uses a fully connected layer to fuse features, and finally outputs the classification probabilities for lakes, rivers, and non-aquatic bodies through the softmax function.

[0133] exist Figure 2 In the specific application scenario shown, accurate lake identification is required. Specifically, in the process of identifying alpine lakes, accurate lake identification is the final and crucial step. Although pixel-level spectral water body identification can identify the fine contours of water bodies, it contains a large number of errors (shadows, snowmelt, etc.). This step constructs three types of labeled samples: lakes, rivers, and non-water bodies. It uses a convolutional neural network (ConvNet) classification method and combines multi-source data (such as synthetic optical remote sensing images, calculated comprehensive water body indices, and topographic parameters calculated from DEM) to achieve accurate classification of water body units, ultimately achieving the goal of accurate identification of alpine lakes.

[0134] During model training, the first step is to construct labeled samples to generate a sample dataset for training and validating the classification model, including three categories: lakes, rivers, and non-water bodies. Specifically, in sample collection, data sources include lake, river, and non-water body samples extracted from known geographic information data, historical imagery, and manually labeled data. Then, class balancing is performed to ensure a relative balance in the number of samples across the three categories, avoiding biases caused by class imbalance during model training. Samples are manually labeled, with clearly marked lake areas designated as lake samples, as lakes typically have smooth boundaries and stable water surface features. When selecting river samples, attention is paid to distinguishing between the linear morphology and water body characteristics of rivers. Non-water body samples include categories such as mountains, vegetation, and bare land; these areas are labeled as non-water body samples to help the model distinguish water bodies from other surface features. After sample collection, the samples are expanded using data augmentation techniques such as rotation, scaling, and flipping to improve the model's robustness and generalization ability. Finally, multi-temporal samples are integrated using multi-temporal imagery samples to ensure the model can adapt to water body identification under different seasons and lighting conditions.

[0135] When constructing the ConvNet classification model, a classification model based on a convolutional neural network (ConvNet) is designed to classify water body units using imagery and topographic data as input. In the input data preparation stage, synthetic optical remote sensing imagery, topographic parameters (such as slope and elevation), and environmental information within the water body unit buffer are integrated into the input data. Data from different sources can be processed through multi-channel input to ensure effective fusion of topographic and imagery information. After data fusion, data standardization is performed to ensure that data from each channel (such as image bands and topographic parameters) are processed at the same scale, avoiding the impact of numerical differences on model training.

[0136] like Figure 7 As shown, the ConvNet architecture is a moderately deep convolutional neural network architecture, including multiple convolutional layers, pooling layers, fully connected layers, and a final softmax classification layer. The specific network depth and structure can be adjusted according to the amount of data and the complexity of the task. Specifically, convolutional layers extract spatial features from the image, and layer-by-layer feature extraction deepens the network's ability to distinguish between water bodies, rivers, and non-water bodies. Pooling layers (such as max pooling and average pooling) are used to reduce the size of the feature maps while retaining the main features, improving the model's computational efficiency and noise resistance. A non-linear activation function (such as ReLU) is used after each convolutional layer to increase the model's expressive power and capture more complex features. The classification layer is placed at the end of the network, using fully connected layers for feature fusion, and finally, the softmax function outputs the classification probabilities for lakes, rivers, and non-water bodies.

[0137] During model training, such as Figure 8As shown, a suitable loss function (such as cross-entropy loss) is selected to guide the training of the model and minimize the classification error; optimization algorithms (such as Adam and SGD) are used to iteratively train the model, and the network weights are continuously adjusted through backpropagation to improve the classification accuracy; hyperparameters such as learning rate, batch size, and number of training rounds are adjusted to optimize the convergence speed and final performance of the model.

[0138] During model validation and testing, the validation set is used to evaluate model performance in real time during training, and the model structure and parameters are adjusted to avoid overfitting. After the model training is completed, an independent test set is used to evaluate the model's generalization ability and verify its performance in practical applications.

[0139] In the water body unit classification process, a trained ConvNet classification model is used to classify each water body unit, achieving accurate identification of alpine lakes. For the classification of environmental information within the buffer zone, the process involves inputting the imagery and topographic data of each water body unit and its buffer zone into the trained ConvNet classification model to predict its category (lake, river, non-water body). The classification result is the model outputting the probability distribution of the three categories; the final category of the water body unit is determined based on the highest probability value. During the integration of classification results, the results are mapped back to geospatial data to create a complete water body classification layer, displaying the classification status of each water body unit. In the lake identification accuracy verification process, accuracy evaluation metrics such as overall accuracy, Kappa coefficient, and F1-score are used to assess the accuracy of lake identification, and the results are compared with the actual lake distribution. Finally, the sources of classification errors (such as image noise and shadow interference) are analyzed, and the model or input data is adjusted to further improve recognition accuracy.

[0140] In the above specific embodiments, the mountain lake identification method based on environmental features provided by the present invention acquires multiple remote sensing images of the target water body; calculates a comprehensive water body index based on the remote sensing images to enhance the spectral characteristics of the water body and obtains a spectrally enhanced image; calculates topographic data based on the spectrally enhanced image, calculates the buffer zone of the target water body based on the topographic data, and extracts the remote sensing image of the buffer zone; inputs the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the topographic data into a pre-trained water body classification model to obtain the category identification result output by the water body classification model. This solves the technical problem that existing methods are limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in high-altitude and cold regions, thereby improving the remote sensing identification accuracy of lakes in high-altitude and cold regions.

[0141] In some embodiments, the method provided by the present invention enhances water body information from multiple levels of spectral and environmental information through a multi-level water body information enhancement strategy. After multi-index fusion, a water body buffer is established, and rich multi-source feature parameters are constructed by combining multi-temporal image synthesis and terrain parameter calculation. This effectively incorporates the local environmental information of the water body into the scope of consideration, and improves the model's ability to identify complex terrain and small water bodies. This multi-level information enhancement strategy breaks through the limitation of traditional methods that rely only on single spectral information, and provides a more comprehensive and richer feature representation for water body identification.

[0142] In some embodiments, the method provided by this invention employs adaptive feature resampling, performing adaptive resampling processing on extracted features based on the size of the water body. This is particularly suitable for identifying small water bodies in high-altitude and cold regions. For small water bodies, high-resolution, fine features are preserved, while for large water bodies, detailed information is appropriately summarized to highlight overall structural features. This differentiated processing strategy ensures that the level of detail in the feature information matches the scale of the water body, especially enhancing the detailed description of small water bodies. In high-altitude and cold regions, small water bodies such as small glacial lakes are common but difficult to identify. This adaptive resampling method significantly improves the model's sensitivity to these small water bodies. By preserving high-precision features of the surrounding environment of small water bodies, this method effectively overcomes the limitations of traditional identification algorithms in handling small water bodies in complex terrains of high-altitude and cold regions, significantly improving the detection accuracy of small water bodies.

[0143] In some embodiments, the method provided by the present invention constructs a classification model that includes three categories: lakes, rivers, and non-water bodies, rather than a simple binary classification of lakes / non-lakes, through the design of a classification model for lakes in high-altitude and cold regions. This is more in line with the complex characteristics of lakes in high-altitude and cold regions and helps to improve the accuracy of identification. By using deep learning technology and combining enhanced multispectral imagery, comprehensive water index, and topographic parameters, a classification model that can make full use of multi-source information is constructed, which fully considers the special needs of lake identification in high-altitude and cold regions. Through adaptive feature processing and targeted model design, the adaptability and accuracy of the method in complex environments are improved.

[0144] In addition to the methods described above, this invention also provides a mountain lake identification device based on environmental characteristics, such as... Figure 9 As shown, the device includes:

[0145] The remote sensing image acquisition unit 910 is used to acquire multiple remote sensing images of the target water body;

[0146] The spectral information enhancement unit 920 is used to calculate a comprehensive water body index based on the remote sensing image to enhance the spectral characteristics of the water body and obtain a spectrally enhanced image.

[0147] The environmental information extraction unit 930 is used to calculate terrain data based on the spectral enhancement image, calculate the buffer zone of the target water body according to the terrain data, and extract remote sensing images of the buffer zone.

[0148] The recognition result generation unit 940 is used to input the remote sensing image of the target water body, the remote sensing image of the buffer zone and the terrain data into a pre-trained water body classification model to obtain the category recognition result output by the water body classification model.

[0149] The water body classification model is obtained by training a pre-built convolutional neural network using topographic data, historical image data, and category labels of water body samples.

[0150] In some embodiments, a comprehensive water body index is calculated based on the remote sensing image to enhance the spectral characteristics of the water body, specifically including:

[0151] Select images from the acquired remote sensing images during periods without snow or ice cover to obtain the target image;

[0152] Calculate the water index in multiple dimensions of the target image;

[0153] By converting the RGB to HSV color space, the water body indices of different dimensions are combined to generate a comprehensive water body index, thereby enhancing the spectral characteristics of the water body.

[0154] In some embodiments, enhancing the spectral characteristics of the water body further includes:

[0155] Water areas are extracted from the spectral enhanced image using a remote sensing index algorithm or a mask generation algorithm.

[0156] In some embodiments, topographic data is calculated based on the spectral enhancement image, a buffer zone of the target water body is calculated based on the topographic data, and remote sensing imagery of the buffer zone is extracted, specifically including:

[0157] Water bodies are extracted from the spectral enhanced image, and water units are identified and marked within the water bodies.

[0158] The spectral enhanced images from different times are stitched together to obtain a composite image;

[0159] Calculate the terrain data of the synthetic image and fuse the obtained terrain data with the data of the synthetic image;

[0160] Construct buffer zones for each of the aforementioned water body units;

[0161] Extract the water environment image of the buffer zone and resample the water environment image of the buffer zone to obtain water environment feature parameters.

[0162] In some embodiments, calculating the terrain data of the synthetic image specifically includes:

[0163] Calculate the slope, elevation, and azimuth of each pixel in the synthesized image.

[0164] In some embodiments, a pre-constructed convolutional neural network is trained using topographic data, historical image data, and category labels of water body samples to obtain the water body classification model, specifically including:

[0165] Obtain topographic data, historical imagery data, and category labels for water samples to construct a dataset;

[0166] The dataset is divided into a training set, a validation set, and a test set;

[0167] Based on a pre-defined loss function, the pre-built convolutional neural network is trained using the training set to obtain an initial model;

[0168] The initial model is iteratively trained using the validation set to obtain the optimized water body classification model.

[0169] In some embodiments, the pre-built network architecture of the convolutional neural network includes:

[0170] Convolutional layers are used to extract spatial features from the input image and enhance the network’s ability to distinguish water bodies, rivers and non-water bodies through layer-by-layer feature extraction.

[0171] Pooling layers are used to reduce the size of feature maps while preserving the main features of the input image;

[0172] Activation functions are used after each convolutional layer to capture additional features;

[0173] The classification layer, located at the end of the network, uses a fully connected layer to fuse features, and finally outputs the classification probabilities for lakes, rivers, and non-aquatic bodies through the softmax function.

[0174] In the above specific embodiments, the mountain lake identification device based on environmental features provided by the present invention acquires multiple remote sensing images of the target water body; calculates a comprehensive water body index based on the remote sensing images to enhance the spectral characteristics of the water body and obtains a spectrally enhanced image; calculates topographic data based on the spectrally enhanced image, calculates the buffer zone of the target water body based on the topographic data, and extracts the remote sensing image of the buffer zone; inputs the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the topographic data into a pre-trained water body classification model to obtain the category identification result output by the water body classification model. This solves the technical problem that existing methods are limited by the information representation capabilities of traditional local spectral and texture features, making it difficult to accurately identify small water bodies in high-altitude and cold regions, thereby improving the remote sensing identification accuracy of lakes in high-altitude and cold regions.

[0175] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and model predictions. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The model predictions of the computer device store static and dynamic information data. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0176] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0177] Corresponding to the above embodiments, this invention also provides a computer storage medium containing one or more program instructions. These one or more program instructions are used to execute the method described above.

[0178] The present invention also provides a computer program product, the computer program product including a computer program, the computer program being stored on a non-transitory computer-readable storage medium, and the computer being able to perform the above-described method when the computer program is executed by a processor.

[0179] In this embodiment of the invention, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0180] The various methods, steps, and logic diagrams disclosed in the embodiments of this invention can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The processor reads information from the storage medium and, in conjunction with its hardware, completes the steps of the above methods.

[0181] The storage medium can be memory, such as volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0182] Among them, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory.

[0183] Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0184] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0185] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in this invention can be implemented using a combination of hardware and software. When applied as software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0186] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying mountain lakes based on environmental features, characterized in that, The method comprises: acquiring a plurality of remote sensing images of a target water body; calculating a comprehensive water body index based on the remote sensing images to enhance the spectral characteristics of the water body and obtain a spectral enhancement image; calculating terrain data based on the spectral enhancement image, calculating a buffer zone of the target water body according to the terrain data, and extracting a remote sensing image of the buffer zone; inputting the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the terrain data into a pre-trained water body classification model to obtain a class recognition result output by the water body classification model; wherein the water body classification model is obtained by training a pre-constructed convolutional neural network using terrain data, historical image data, and class labels of water body samples; wherein the water body classification model is obtained by training a pre-constructed convolutional neural network using terrain data, historical image data, and class labels of water body samples, and specifically comprising: acquiring terrain data, historical image data, and class labels of water body samples to construct a data set; dividing the data set into a training set, a validation set, and a test set; training the pre-constructed convolutional neural network using the training set based on a pre-set loss function to obtain an initial model; iteratively training the initial model using the validation set to obtain an optimized water body classification model; wherein the network architecture of the pre-constructed convolutional neural network comprises: a convolutional layer for extracting spatial features of input images and deepening the network's recognition ability of water bodies, rivers, and non-water bodies through layer-by-layer feature extraction; a pooling layer for reducing the size of the feature map and preserving the main features of the input image; an activation function using a nonlinear activation function after each convolution to capture additional features; a classification layer using a fully connected layer for feature fusion at the end of the network, and finally outputting the classification probability of lakes, rivers, and non-water bodies through a softmax function; During model training, first, label samples need to be constructed to generate a sample data set for training and validating the classification model, including three categories of lakes, rivers, and non-water bodies; in sample collection, the data is extracted from known geographic information data, historical image data, and manually annotated data to obtain samples of lakes, rivers, and non-water bodies; then, class balancing is performed to ensure that the number of samples of the three categories is relatively balanced, avoiding bias caused by class imbalance during model training; sample manual annotation, clearly annotated lake areas are used as lake samples, lakes usually have smooth boundaries and stable water surface characteristics; when selecting river samples, the linear morphology and water body characteristics of the river are distinguished; non-water body samples include mountains, vegetation, and bare land, which are annotated as non-water body samples to help the model distinguish water bodies from other surface features; after collecting samples, data augmentation techniques such as rotation, scaling, and flipping are used to expand the sample set; finally, multi-temporal samples are integrated, and multi-temporal image samples are used to ensure that the model can adapt to water body recognition under different seasons and lighting conditions; In constructing the ConvNet classification model, a classification model is designed based on a convolutional neural network (ConvNet), and image and terrain data are taken as inputs to classify water body units; in the input data preparation stage, the synthesized optical remote sensing image, terrain parameters, and environmental information within the buffer zone of the water body unit are integrated into the input data, and data from different sources can be processed through multi-channel input to ensure effective fusion of terrain and image information; after data fusion, data standardization is performed to standardize the input data and ensure that the data in each channel is processed on the same scale to avoid affecting model training due to value differences. 2.The mountain lake recognition method based on environmental features according to claim 1, characterized in that, Based on the remote sensing image, a comprehensive water body index is calculated to enhance the spectral characteristics of the water body, specifically including: Selecting an image of a period without ice and snow coverage from the obtained remote sensing images to obtain a target image; Calculating a plurality of dimensions of water body indexes of the target image; Through RGB to HSV color space conversion, combining water body indexes of different dimensions to generate a comprehensive water body index to enhance the spectral characteristics of the water body. 3.The mountain lake recognition method based on environmental features according to claim 2, characterized in that, After enhancing the spectral characteristics of the water body, it further includes: Extracting the water body region on the spectral enhancement image through a remote sensing index algorithm or a mask generation algorithm. 4.The mountain lake recognition method based on environmental features according to claim 1, characterized in that, Based on the spectral enhancement image, calculating terrain data, calculating the buffer zone of the target water body according to the terrain data, and extracting the remote sensing image of the buffer zone, specifically including: Extracting the water body region on the spectral enhancement image, identifying and marking the water body unit in the water body region; Splicing spectral enhancement images at different times to obtain a synthesized image; Calculating the terrain data of the synthesized image and fusing the obtained terrain data with the data of the synthesized image; Constructing a buffer zone of each water body unit; Extracting the water body environment image of the buffer zone and resampling the water body environment image of the buffer zone to obtain water body environment feature parameters. 5.The mountain lake recognition method based on environmental features according to claim 4, characterized in that, Calculating the terrain data of the synthesized image, specifically including: Calculating the slope information, elevation, and azimuth angle of each pixel in the synthesized image.

6. A device for the recognition of mountain lakes based on environmental features for implementing the method according to any one of claims 1 to 5, characterized in that, The device includes: A remote sensing image acquisition unit for acquiring a plurality of remote sensing images of a target water body; A spectral information enhancement unit for calculating a comprehensive water body index based on the remote sensing image to enhance the spectral characteristics of the water body and obtain a spectral enhancement image; An environmental information extraction unit for calculating terrain data based on the spectral enhancement image, calculating the buffer zone of the target water body according to the terrain data, and extracting the remote sensing image of the buffer zone; An identification result generation unit for inputting the remote sensing image of the target water body, the remote sensing image of the buffer zone, and the terrain data into a pre-trained water body classification model to obtain a class identification result output by the water body classification model; The water body classification model is obtained by training a pre-constructed convolutional neural network using terrain data, historical image data, and class labels of water body samples.

7. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the method of any one of claims 1-5.

8. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1-5.

Images