A forest stand tree species identification method and system based on multi-source remote sensing data fusion

Abstract

The application discloses a kind of based on multi-source remote sensing data fusion's forest stand tree species identification method and system, the method includes the following steps: obtaining data, the data includes forest remote sensing image data and ground data;The data obtained is preprocessed;Double-path CBAM-UNet network model is constructed;The data after pre-processing is input into the double-path CBAM-UNet network model to carry out tree species classification prediction and obtain tree species classification chart;Image mosaic is carried out to the tree species classification chart and tree species classification grid chart is obtained;The system includes data acquisition module, data preprocessing module, prediction module and image processing module.The application provides a kind of based on multi-source remote sensing data fusion's forest stand tree species identification method and system, and multi-scale feature capture and the feature fusion of multi-source heterogeneous data can be realized by the double-path CBAM-UNet network model constructed, and high-precision identification can be carried out to forest stand tree species under complex link, provide high-quality data for forest management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of forestry resource monitoring, and particularly relates to a forest stand tree species identification method and system based on multi-source remote sensing data fusion. BACKGROUND

[0002] At present, high-precision forest tree species identification mainly relies on artificial field investigation. Professional personnel need to enter the forest area and visually interpret remote sensing images to record the tree species composition and boundary information of each sample plot. The workload is large, the cost is high, and it is difficult to implement in complex terrain areas. It is difficult to obtain large-scale continuous forest resource data, and it is difficult to provide timely and accurate decision support for forestry management.

[0003] The automatic tree species identification method based on remote sensing technology mainly includes pixel-based supervised classification (Pixel-based Supervised Classification, PSC) (Franklin et al., 2000; Dalponte et al., 2012), object-based image analysis (Object-based Image Analysis, OBIA) (Blaschke, 2010; Immitzer et al., 2016), machine learning classification (Machine Learning Classification, MLC) (Maxwell et al., 2018; Liu et al., 2018) and deep learning classification (Deep Learning Classification, DLC) developed in recent years (Hamraz et al., 2019; Onishi and Ise, 2021).

[0004] The PSC method compares the spectral characteristics of each pixel in the remote sensing image with the known tree species samples, and determines the tree species category to which the pixel belongs through maximum likelihood method, minimum distance method and other algorithms. This method is simple to implement and has high computational efficiency, but it does not consider spatial context information, resulting in "salt and pepper noise" in the classification results, and insufficient ability to distinguish between tree species with similar spectral characteristics.

[0005] The OBIA method first segments the image into homogeneous objects, and then classifies based on the spectral, shape, texture and other characteristics of the objects. This method can utilize the spatial structure information of tree crowns to reduce "salt and pepper noise", but the adjustment of segmentation parameters is complex, and it is difficult to determine the optimal segmentation scale in heterogeneous forest stands, resulting in over-segmentation or under-segmentation.

[0006] The MLC method mainly adopts algorithms such as random forest, support vector machine and artificial neural network, and establishes a classification model by extracting spectral and texture features. This method can process high-dimensional data and has strong modeling capability for nonlinear relationships between features, but it is heavily dependent on manual feature engineering, has poor scalability, and has limited generalization ability between different regions or different sensor data.

[0007] The DLC method has developed rapidly in recent years and mainly adopts convolutional neural network (CNN) to automatically extract hierarchical features. The U-Net architecture and its variants retain spatial detail information through the encoder-decoder structure and skip connection, and are widely used in forest remote sensing segmentation tasks. However, most existing deep learning methods only use single optical remote sensing data, and do not fully exploit vertical structure and seasonal variation information.

[0008] The above methods are proposed under specific research areas and specific forest conditions, and it is difficult to effectively distinguish tree species with similar spectral characteristics, and it is difficult to support large-scale operations. In the case of high canopy density, complex forest structure, and mixed natural forest and artificial forest, it is difficult for existing methods to achieve high-precision tree species recognition.

[0009] Therefore, developing a deep learning method using high-resolution optical remote sensing and canopy height model to achieve high-precision tree species recognition in complex forest conditions has become a technical problem to be solved in modern forest fine management. SUMMARY

[0010] In view of the above-mentioned deficiencies in the existing forest tree species recognition technology, the purpose of the present application is to provide a forest stand tree species recognition technology for solving the problem of high-precision recognition of forest stand tree species in complex forest conditions.

[0011] To achieve the above-mentioned purpose, the first aspect of the present application provides a forest stand tree species recognition method based on multi-source remote sensing data fusion, which comprises the following steps:

[0012] Obtaining data, the data comprising forest remote sensing image data and ground data;

[0013] Preprocessing the obtained data;

[0014] Constructing a double-path CBAM-UNet network model;

[0015] Inputting the preprocessed data into the double-path CBAM-UNet network model for tree species classification prediction and obtaining a tree species classification map;

[0016] Image stitching is performed on the tree species classification map to obtain a tree species classification raster map.

[0017] According to the first aspect of the present application, the forest stand tree species identification method based on multi-source remote sensing data fusion comprises: preparing deep learning training samples according to the obtained data to train the dual-path CBAM-UNet network model.

[0018] According to the first aspect of the present application, the preparation of deep learning training samples according to the obtained data to train the dual-path CBAM-UNet network model comprises:

[0019] Preparing training samples, including image slicing and label preparation, and normalizing, feature engineering and data enhancement processing of the samples;

[0020] Model initialization configuration, including using AdamW optimizer, cosine annealing scheduling dynamic learning rate strategy and weighted cross-entropy loss function as the main loss function;

[0021] Model training, using mixed precision training strategy and gradient clipping method to improve the training efficiency of the model and applying random dropout strategy to enhance the generalization ability of the model.

[0022] According to the first aspect of the present application, the dual-path CBAM-UNet network model comprises an input layer and an encoder, the input layer comprises a left path and a right path, and the encoder comprises a left path encoder and a right path encoder, the input layer is used to receive original image data and simultaneously input the original image data to the left path encoder and the right path encoder for parallel processing through the left path and the right path;

[0023] The left path encoder comprises:

[0024] The first level feature extraction layer comprises a BasicConv module;

[0025] The second level feature extraction layer comprises a max-pooling layer for down-sampling and a BasicConv module;

[0026] The third level feature extraction layer comprises a max-pooling layer for down-sampling and a BasicConv module, and Dropout processing is performed after max-pooling;

[0027] The fourth level feature extraction layer comprises a max-pooling layer for down-sampling and a BasicConv module, and Dropout processing is performed after max-pooling;

[0028] The right path encoder comprises:

[0029] The first level feature extraction layer comprises a BasicConv module;

[0030] The second-level feature extraction layer includes a max-pooling layer for down-sampling and a BasicConv module, and the max-pooling layer is subjected to a Dropout process after max-pooling;

[0031] The third-level feature extraction layer includes a max-pooling layer for down-sampling and a BasicConv module, and the max-pooling layer is subjected to a Dropout process after max-pooling;

[0032] The fourth-level feature extraction layer includes a max-pooling layer for down-sampling and a BasicConv module, and the max-pooling layer is subjected to a Dropout process after max-pooling.

[0033] According to the first aspect of the present application, the left path encoder is used for processing optical remote sensing image data, the right path encoder is used for processing CHM canopy height model data, the optical remote sensing image data includes RGB bands or multispectral data, and the CHM canopy height model data provides forest height information, and the complementary fusion of spectral features and structural features is realized through double-path feature extraction.

[0034] According to the first aspect of the present application, the double-path CBAM-UNet network model sets a mode switch to control the working mode of the BasicConv module, and the working mode includes a standard mode and an ASPP enhanced mode.

[0035] When in the standard mode, the input features are output after being processed by the basic convolution unit only.

[0036] When in the ASPP enhanced mode, the input features are processed in parallel by the basic convolution unit and the ASPP sub-module, the processing results of the two paths are fused by element-wise addition, the fused features are subjected to batch normalization and ReLU activation function processing, and are output after applying Dropout regularization.

[0037] According to the first aspect of the present application, the double-path CBAM-UNet network model includes a decoder, an output layer and a CBAM module.

[0038] The decoder is a four-layer structure, each layer of which includes a transposed convolution module for up-sampling and a DoubleConv module for processing fused features.

[0039] The output layer includes a convolution layer, which maps the output features of the decoder to the pre-defined tree species categories and applies a Softmax activation function to generate the probability distribution of each tree species category.

[0040] The CBAM module is used for enhancing the fused features after feature extraction of the left path encoder and the right path encoder each time, and the enhanced features are fused with the output features of the transpose convolution module of the corresponding layer in the decoder, and the fused features are input into the DoubleConv module for processing.

[0041] According to the first aspect of the application, the double-path CBAM-UNet network model comprises a bottleneck layer, and the bottleneck layer comprises:

[0042] a left path DoubleConv module for processing the feature map output by the left path encoder;

[0043] a right path DoubleConv module for processing the feature map output by the right path encoder;

[0044] a bottleneck fusion layer for concatenating the feature maps output by the left path DoubleConv module and the right path DoubleConv module in the channel dimension to generate a concatenated feature map;

[0045] a terminal enhancement module for converting the concatenated feature map into a highly nonlinear representation;

[0046] a CBAM enhancement module for applying attention mechanism to the output of the bottleneck layer for enhancement processing.

[0047] According to the first aspect of the application, the forest remote sensing image data comprises airborne digital images, satellite images and CHM canopy height model data, and the preprocessing of the obtained data comprises:

[0048] radiometric correction, atmospheric correction, orthorectification and panchromatic fusion are performed on the airborne digital images and the satellite images;

[0049] noise suppression, edge preservation and data normalization are performed on the CHM data, and the processed CHM data is image-registered with the optical image to realize the unified alignment of the spatial coordinate system.

[0050] To achieve the above-mentioned purposes, the second aspect of the application provides a forest stand tree species identification system based on multi-source remote sensing data fusion, comprising:

[0051] a data acquisition module for acquiring forest remote sensing image data and ground data;

[0052] a data preprocessing module for preprocessing the forest remote sensing image data and ground data acquired by the data acquisition module;

[0053] A prediction module is configured to perform tree species classification prediction according to the input image slices and obtain a tree species classification map;

[0054] An image processing module is configured to perform splicing processing on the tree species classification map to obtain a complete tree species classification grid map

[0055] Advantages of the embodiment of the present application: first, the dual-path CBAM-UNet network model of the present application captures multi-scale features, fuses GF-7 multispectral images and high-resolution CHM data, and obtains multi-scale features such as spectrum, spatial texture and vertical height by means of a deep learning model, thereby improving the ability to distinguish similar tree species and the classification accuracy. Secondly, in the area with mixed coniferous and broad-leaved trees and large terrain undulations, the classification accuracy is more than 81%, and the tree species can be stably identified in artificial pure forest and natural mixed forest, and the adaptability to complex forest land is strong. Thirdly, a fully automatic processing flow is constructed, manual intervention and field operation are reduced, a classification map is automatically output, large-scale forest mapping update is quickly completed, and an automatic and efficient process is realized. Finally, through multi-source data feature fusion and algorithm improvement, the classification accuracy is better than that of the traditional method, and the Kappa coefficient reaches 0.78, thereby providing high-quality data for forest management. At the same time, under the condition of NVIDIA RTX4090 system and pytorch2.2 framework, it only takes 2 to 3 hours to process 12,300 hectares of test area, the calculation performance is excellent, and it is suitable for automatic processing of large-scale forest information. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0057] Figure 1 A flowchart of the forest stand tree species identification method based on multi-source remote sensing data fusion according to the present application;

[0058] Figure 2 A flowchart of the deep learning model sample making process according to the present application;

[0059] Figure 3 A schematic diagram of the dual-channel CBAM-Unet network structure according to the present application;

[0060] Figure 4 A schematic diagram of the BasicConv module structure according to the present application;

[0061] Figure 5 A schematic diagram of the ASPP module structure according to the present application;

[0062] Figure 6A schematic diagram of the CBAM module structure described in the present application;

[0063] Figure 7 A schematic diagram of the model training process described in the present application;

[0064] Figure 8 A schematic diagram of the tree species classification prediction process of the present application;

[0065] Figure 9 A schematic diagram of the structure of the forest stand tree species identification system based on multi-source remote sensing data fusion of the present application. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0067] Before the present application is further described, the terms and phrases involved in the embodiments of the present application are explained, which are applicable to the following explanations:

[0068] Canopy Height Model (CHM): is a digital model reflecting the relative height of trees above the ground. It is calculated by subtracting the Digital Elevation Model (DEM) from the Digital Surface Model (DSM) obtained by LiDAR or photogrammetry technology, and is used to provide pixel-level tree height information. CHM has important applications in forestry remote sensing, ecological research and resource management, especially in tree species classification, forest boundary extraction and biomass estimation.

[0069] ArcMap: is a core component of GIS system developed by Environmental Systems Research Institute (ESRI), and is one of the main application programs of ArcGIS Desktop suite. It has powerful map making, spatial analysis and data management functions, and can display, edit and analyze map data. It provides data view and layout view, supports import and export of multiple data formats. Users can perform geographic processing tasks, publish map services, customize user experience through it, and is widely used in urban planning, environmental protection, resource management and other fields.

[0070] ASPP (Atrous Spatial Pyramid Pooling): is a module for deep learning (especially computer vision), which aims to enhance the model's perception of different scale targets through multi-scale feature extraction and atrous convolution technology.

[0071] Figure 1 A flowchart of a forest stand tree species identification method based on multi-source remote sensing data fusion of the present application is shown, which includes the following steps:

[0072] Step S1: Obtain data.

[0073] The data includes forest remote sensing image data and ground data.

[0074] The forest remote sensing image data used in the present application is sub-meter high spatial resolution multispectral data, which includes airborne digital images, satellite images (high resolution satellite images) and canopy height model (CHM) data. Among them, airborne digital images are used to make labels required for deep learning model training, and satellite images are used for tree species classification and forest boundary delineation.

[0075] In the embodiment of the present application, the main data source is the GF-7 satellite multispectral optical image and the canopy height model (CHM) data:

[0076] GF-7 satellite image: acquired in the vegetation growth period on May 23, 2023, the original multispectral band spatial resolution is 2.6 meters, after fusion of the panchromatic band, the spatial resolution is improved to about 0.5 meters. The image is processed through radiation calibration, atmospheric correction, orthorectification and panchromatic fusion, etc. to generate high-quality multispectral orthoimage, and resampled to 0.5 meter resolution.

[0077] CHM data: collected by airborne laser radar on September 3, 2023, with a resolution of 0.5 meters. By accurately registering and superimposing CHM data and GF-7 spectral image, canopy height information can be provided for each pixel, realizing the fusion of spectral features and spatial structure features.

[0078] The ground data of the present application covers forest tree species sample point data collected by using real-time kinematic differential (RTK) technology, and national forest resources second-class survey subcompartment data. Among them, the forest tree species sample point data records the tree species type and spatial coordinate position in detail, which is mainly used for model checking and verification; the national forest resources second-class survey subcompartment data can be used as a reference basis for label making.

[0079] In the embodiment of the present application, field sample plot investigation was carried out in Mulan test area in September 2024, 100 ground sample points were collected for verification. At the same time, the forest stand subcompartment division data formed by the national forest resources second-class survey in 2021 was obtained, which can be used for precision verification and auxiliary annotation. Based on the above data, the tree species classification system of the test area was finally determined, and the target categories included 7 categories, namely larch, Chinese pine, scots pine, white birch, oak, bare land and background.

[0080] The remote sensing image data and the ground data are preprocessed to serve as training data of a deep learning neural network model.

[0081] Step S2: preprocessing the acquired data.

[0082] In the embodiment of the present application, different preprocessing methods are adopted according to the data type difference for the acquired original remote sensing image data and ground data.

[0083] For the airborne and satellite data, the preprocessing procedure is as follows:

[0084] Radiation correction: for the panchromatic and multispectral images of GF-7 satellite, the absolute radiation correction method is adopted, the correction coefficient is selected according to the sensor model, and the DN value is converted into apparent reflectivity;

[0085] Atmospheric correction: for the GF-7 multispectral image, the FLAASH model is used to eliminate the atmospheric influence;

[0086] Ortho rectification: the panchromatic and multispectral images are geometrically corrected based on the 0.5m resolution DEM obtained by LiDAR and RPC parameters, and the control point accuracy is better than 0.5 pixels;

[0087] Panchromatic fusion: the panchromatic image is fused by the Gram-Schmidt sharpening algorithm, the spectral characteristics are maintained, and the spatial resolution is improved to 0.5m.

[0088] The CHM crown height model data and the optical image are subjected to image registration processing, and the registration processing process includes:

[0089] Image registration: the nearest neighbor resampling method is adopted to unify the GF-7 image and the CHM data to 0.5m resolution grid;

[0090] Noise suppression: the 7x7 window and the local adaptive Frost filter with K coefficient of 2.0 are applied to the GF-7 image and the CHM data to suppress abnormal high value noise;

[0091] Edge preservation: the edge preservation parameter a=0.4 is set during filtering to retain the forest boundary characteristics;

[0092] Data normalization: the CHM height value and the GF-7 multispectral value are linearly normalized to the range of 0-1.

[0093] Through the above preprocessing, the multi-source data fused with multispectral and height information are generated to provide a basis for subsequent analysis.

[0094] Step S3: preparing deep learning training samples according to the preprocessed data.

[0095] Figure 2The application embodiment shows a manufacturing process of a deep learning training sample and a schematic diagram of a data set division manner.

[0096] Firstly, representative areas in the test area range are selected by using ArcMap software, and in the room, 0.1-meter resolution aerial images are used to visually interpret and draw the initial draft of the forest boundary of different tree species by referring to the forest resource class II survey subcompartment data.

[0097] Then, the forest and tree species sample points are collected by using RTK in the field, the tree species types and patch boundary coordinates are recorded, the forest vector map patches are verified and corrected, and high-quality forest boundary vector maps are obtained. The rasterization is performed to the pixel grid (grid length is 0.5 m) of the GF-7 image, and a tree species classification label map is obtained.

[0098] At the same time, in order to adapt to the input requirements of the deep learning model, the labels and corresponding images are cropped into small slices. In the application embodiment, the size of the small slice image picture is 256*256 pixels, and 14,159 labeled image samples are made for model training and evaluation.

[0099] It can be understood that, in order to enhance the training effect and generalization ability of the model, the training sample data is normalized and feature engineered in the application embodiment.

[0100] In the normalization processing, first, the mean value and the standard deviation of each feature dimension are calculated based on the entire data set, and then the linear transformation of the formula is performed on each data point . The operation maps the original data to the standard normal distribution space with a mean value of 0 and a standard deviation of 1, which can eliminate the dimensional differences between different features, retain the distribution form of the data, make the model training based on gradient descent more efficient and avoid optimization deviation caused by feature scale difference.

[0101] In the feature engineering processing, for the spatial structure and semantic information of the image data, texture features, shape features and color histogram features can be extracted to enhance the representation ability of the data and provide more discriminative inputs for the model. In addition, combined with the automatic feature extraction advantage of deep learning, the high-level semantic features of the image are abstracted layer by layer through the convolutional neural network structure, which can reduce the subjectivity of manual feature design, mine the complex correlations hidden in the data, and further improve the understanding and generalization ability of the model to the image scene.

[0102] In addition, in order to increase the diversity of the training data, avoid overfitting and enhance the generalization ability of the model, a data enhancement technique is used to expand the training data set. In the application embodiment, the data enhancement techniques include:

[0103] Geometric augmentation: random horizontal / vertical flip (probability 0.5 each), random rotation (rotation magnitude ±15°, probability 0.3).

[0104] Spectral augmentation: random brightness adjustment (adjustment magnitude ±10%, probability 0.3), random contrast adjustment (adjustment magnitude ±10%, probability 0.3).

[0105] Spatial augmentation: random cropping, cropped image size 256x256 and more than 80% valid pixels.

[0106] Finally, the prepared sample data is divided into a training set, a validation set and a test set according to a certain ratio to ensure the fairness and reliability of the model training process and the accuracy evaluation. In the embodiment of the present application, the ratio of the training set, the validation set and the test set is 6:2:2.

[0107] Step S4: constructing a double-path CBAM-UNet network model.

[0108] In terms of neural network design, the improved double-path CBAM-UNet deep convolutional neural network architecture is adopted in the embodiment of the present application, which can accurately identify different tree species by using multispectral and canopy height data.

[0109] Figure 3 The overall structure schematic diagram of the double-path CBAM-UNet network model of the present application is shown. As can be seen from the figure, the neural network includes an input layer, an encoder, a feature fusion layer, a bottleneck layer, a decoder and an output layer, and each module of the neural network will be described in detail below in combination with the schematic diagram.

[0110] Input layer: the input layer provides double-path input channels. Among them, the left path input channel is used to receive GF-7 multispectral data, and the right path input channel is used to receive CHM canopy height data.

[0111] In the embodiment of the present application, the channel number of the GF-7 multispectral data is 4, and the size is 256x256 pixels; the channel number of the CHM canopy height data is 1, and the size is also 256x256 pixels.

[0112] Encoder: the encoder includes parallel left path encoder and right path encoder, which respectively extracts features from the feature data input by the left path input channel and the right path input channel. Among them, the left path encoder and the right path encoder are both multi-level structures to realize multi-layer step-by-step feature extraction. In the embodiment of the present application, the left path encoder and the right path encoder are both four-level structures. The input layer simultaneously inputs the original image data to the left path encoder and the right path encoder for parallel processing.

[0113] The left path encoder is used for processing left path input channel inputted processing optical remote sensing image data (including GF-7 multi-spectral image data), and the optical remote sensing image data includes RGB bands or multi-spectral data, and the network structure is as follows:

[0114] The first feature extraction layer includes a BasicConv module. In the embodiment of the application, a 96-channel feature map with a size of 256x256 pixels is output by the BasicConv module.

[0115] The second feature extraction layer includes a maximum pooling layer for down sampling and a BasicConv module, and the maximum pooling can be subjected to Dropout processing. In the embodiment of the application, the size of the pooling kernel is 2x2, and a 128-channel feature map with a size of 128x128 pixels is output by the BasicConv module.

[0116] The third feature extraction layer includes a maximum pooling layer for down sampling and a BasicConv module, and the maximum pooling can be subjected to Dropout processing. In the embodiment of the application, the size of the pooling kernel is 2x2, and a 256-channel feature map with a size of 64x64 pixels is output by the BasicConv module.

[0117] The fourth feature extraction layer includes a maximum pooling layer for down sampling and a BasicConv module, and the maximum pooling can be subjected to Dropout processing. In the embodiment of the application, the size of the pooling kernel is 2x2, and a 512-channel feature map with a size of 32x32 pixels is output by the BasicConv module.

[0118] The right path encoder is used for processing CHM canopy height data inputted by the right path input channel, and the CHM canopy height model data provides forest height information, and the network structure is as follows:

[0119] The first feature extraction layer includes a BasicConv module. In the embodiment of the application, a 64-channel feature map with a size of 256x256 is output by the BasicConv module.

[0120] The second feature extraction layer includes a maximum pooling layer for down sampling and a BasicConv module, and the maximum pooling can be subjected to Dropout processing. In the embodiment of the application, a 128-channel feature map with a size of 128x128 pixels is output by the BasicConv module.

[0121] The third feature extraction layer includes a maximum pooling layer for down sampling and a BasicConv module, and the maximum pooling can be subjected to Dropout processing. In the embodiment of the application, a 256-channel feature map with a size of 64x64 pixels is output by the BasicConv module.

[0122] The fourth-level feature extraction layer includes a max-pooling layer for down-sampling and a BasicConv module, and the max-pooling layer can be subjected to a Dropout process. In the embodiment of the application, the BasicConv module outputs a 512-channel feature map with a size of 32*32 pixels.

[0123] In the embodiment of the application, the feature extraction and fusion of multi-source heterogeneous data, including complementary fusion of spectral features and structural features, are achieved through double-path multi-level parallel feature extraction.

[0124] The Dropout process in each feature extraction layer in the left-path encoder and the right-path encoder is an optional operation. Dropout can randomly discard neurons to prevent overfitting of the neural network and improve the generalization ability of the model. In the embodiment of the application, the Dropout Rate is 0.3.

[0125] In the embodiment of the application, the BasicConv module used in each feature extraction layer in the left-path encoder and the right-path encoder is a basic convolutional layer for feature extraction. Figure 4 The network structure diagram of the BasicConv module is shown. As shown in the diagram, the BasicConv module of the application has two working modes, Figure 4 The left side shows the network structure diagram of the BasicConv module in the standard mode; Figure 4 The right side shows the network structure diagram of the BasicConv module in the ASPP enhanced mode.

[0126] As shown in the left side, Figure 4 As shown in the left side, the BasicConv module in the standard mode includes two convolutional layers connected in series, each followed by a BatchNorm and a ReLU activation function. To prevent overfitting of the model, a Dropout layer can be connected after each layer for regularization. At this time, the input features are only processed by the basic convolution unit and then output. In the embodiment of the application, the kernel size of the two convolutional layers is 3*3, and the Dropout Rate is 0.5.

[0127] As shown in the right side, Figure 4 As shown in the right side, the BasicConv module in the ASPP enhanced mode integrates an ASPP module in the network structure. At this time, the input features are processed in parallel by the basic convolution unit and the ASPP submodule, the processing results of the two paths are fused by element-wise addition, and the fused features are processed by batch normalization and ReLU activation function, and then output after applying Dropout regularization. In this way, the BasicConv module has better multi-scale feature extraction capability.

[0128] In the embodiment of the present application, the dual-path CBAM-UNet network model has a preset aspp mode switch, which is used to determine the working mode of the BasicConv module. In a specific application scenario, when the model needs to balance the calculation efficiency and the basic feature extraction capability (such as processing small size targets or low complexity scenes), the aspp mode switch parameter can be set to the standard mode; and when the task requirement emphasizes more on multi-scale context information capture (such as large size target segmentation or complex scene classification), the aspp mode switch parameter can be set to the ASPP enhanced mode to improve the feature expression capability of the model.

[0129] Figure 5 The network structure diagram of the ASPP module is shown, and as can be seen from the diagram, the input layer of the ASPP module receives a feature map as the input data processed by the module. Five parallel branches are connected after the input layer, and the five specific branches include:

[0130] The first convolution branch: in the embodiment of the present application, a 1x1 convolution kernel is used to perform convolution operation on the input feature map, which is used to extract local detail features and capture the correlation between feature map channels.

[0131] The second dilated convolution branch: in the embodiment of the present application, a 3x3 convolution kernel is used and the dilated rate of dilated convolution is 6, which is used to expand the receptive field and capture medium scale features.

[0132] The third dilated convolution branch: in the embodiment of the present application, a 3x3 convolution kernel is used and the dilated rate of dilated convolution is 12, which is used to further expand the receptive field and obtain larger scale features.

[0133] The fourth dilated convolution branch: in the embodiment of the present application, a 3x3 convolution kernel is used and the dilated rate of dilated convolution is 18, which is used to make the model have a larger receptive field and capture global scale features.

[0134] The fifth pooling and convolution branch: first, a pooling layer with a 1x1 pooling kernel is used to perform global average pooling operation on the feature map; then, a 1x1 convolution is performed; then, BatchNorm and ReLU activation function processing are performed; finally, the feature map size is adjusted by Bilinear Interpolate.

[0135] After the feature map is processed by the five parallel branches respectively, the fusion layer is used to perform concatenation (Concat) operation on the feature maps processed by the above parallel branches in the channel dimension, so as to fuse multi-scale feature information. Then, a 1x1 convolution kernel is used to perform convolution operation on the concatenated feature map, and the channel number of the feature map is adjusted.

[0136] Feature fusion layer: the dual-path CBAM-UNet network of the application has a plurality of feature fusion layers, which provide a multi-scale feature fusion mechanism, and implement a progressive cascaded feature fusion strategy at each encoding level to fuse the features of the left and right paths.

[0137] It can be understood that, in the multi-level feature fusion process, the feature map sizes of the left path and the right path need to be consistent to meet the conditions of feature fusion. At the same time, the feature maps are stacked to increase the number of channels of the feature maps, thereby enriching the expression of the features and avoiding information loss. The specific process of multi-level feature fusion is as follows:

[0138] First-level fusion: cascade the left path features and the right path features to form an enhanced representation of the features. In the embodiment of the application, the channel number of the feature map input by the left path is 96, the channel number of the feature data input by the right path is 64, and the size of the feature map is 256x256 pixels. After fusion, a 256x256 feature map with 160 channels is obtained.

[0139] Second-level fusion: cascade the left path features and the right path features to form an enhanced representation of the features. In the embodiment of the application, the channel number of the feature map input by the left path and the right path is 128, and the size of the feature map is 128x128 pixels. After fusion, a 128x128 feature map with 256 channels is obtained.

[0140] Third-level fusion: cascade the left path features and the right path features to form an enhanced representation of the features. In the embodiment of the application, the channel number of the feature map input by the left path and the right path is 256, and the size of the feature map is 64x64 pixels. After fusion, a 64x64 feature map with 512 channels is obtained.

[0141] Fourth-level fusion: cascade the left path features and the right path features to form an enhanced representation of the features. In the embodiment of the application, the channel number of the feature map input by the left path and the right path is 512, and the size of the feature map is 32x32 pixels. After fusion, a 32x32 feature map with 1024 channels is obtained.

[0142] Bottleneck layer: the dual-path CBAM-UNet network model of the application includes a bottleneck layer, which is a key component connecting the encoder and the decoder. Through the bottleneck layer, the fused high-level features can be further enhanced. In the embodiment of the application, the bottleneck layer includes the following modules:

[0143] DoubleConv module: including a left DoubleConv module and a right DoubleConv module. The left and right DoubleConv modules process the feature maps output by the left and right encoders, respectively. In the embodiment of the application, as known from the above-described encoder, the left and right encoders output 512-channel feature maps with a size of 32x32 pixels.

[0144] Bottleneck layer fusion module: used for concatenating features of the left and right paths to generate a 1024-channel feature map with a size of 32x32 pixels.

[0145] Terminal enhancement module: used for converting the multi-channel feature map output by the bottleneck layer fusion module into a highly nonlinear representation through an additional DoubleConv module.

[0146] CBAM enhancement module: used for applying an attention mechanism to the output of the bottleneck layer to highlight key features. It can be understood that the output of the bottleneck layer is the output of the terminal enhancement module.

[0147] The bottleneck layer can effectively solve the feature redundancy and boundary ambiguity problems in multi-source data integration through the implementation of multiple strategies such as cross-modal fusion, attention enhancement, and parameter compression. The model can greatly improve the boundary segmentation accuracy while maintaining efficiency compared to the traditional UNet network model.

[0148] Decoder: the dual-path CBAM-UNet network model of the present application includes a decoder, which is used for restoring the spatial resolution of the feature map. In the embodiment of the present application, the decoder also has a four-layer structure, and the difference from the encoder is that each layer of the decoder includes a transposed convolution module for upsampling and a DoubleConv module for processing fused features. In the embodiment of the present application, the convolution kernel size of the transposed convolution module is 2x2.

[0149] Meanwhile, the enhanced features obtained by the CBAM module enhancement of each layer of the left and right path encoders are connected to the corresponding layers of the decoder through a skip connection. The upsampled features and the enhanced features corresponding to the skip connection are merged through a concatenation operation. In the decoding process of the decoder, the channels of the feature map are gradually reduced (512→256→128→64) while the size of the feature map is gradually increased (32→64→128→256).

[0150] Output layer: the dual-path CBAM-UNet network model of the present application also includes an output layer, which maps the output features of the decoder to the predefined classes of forest tree species through a convolution layer with a convolution kernel size of 1x1, specifically including 7 predefined classes. Finally, the probability distribution of each class is generated by applying a Softmax activation function. Here, the size of the output feature map of the output layer is 256x256 pixels.

[0151] In conclusion, the double-path CBAM-UNet network of the present application contains two parallel encoder paths, which respectively process spectral information (from GF-7 four-channel multispectral data) and height information (from single-channel CHM data), through the "separation-fusion" mechanism, the neural network can independently extract features of different data types, avoid mutual interference, and perform feature fusion at four resolution levels, fully exerting the complementarity of spectral and height information. Through multi-level feature extraction and feature fusion, high-precision recognition of different tree species is realized.

[0152] Firstly, the network architecture combines CBAM attention module and ASPP multi-scale feature extraction technology. CBAM enhances the key features of regions such as tree species boundaries through dual attention mechanisms of channels and spaces. At the same time, the ASPP module uses parallel convolution with multiple dilation rates (6, 12, 18), which can capture tree crown features of different scales, thereby improving the recognition ability of different scale tree species.

[0153] Secondly, the network is optimized in the design of skip connection. On the basis of the traditional U-Net network model, the double-path feature fusion after CBAM enhancement is connected with the decoder, thereby improving the recognition ability of the network model in high canopy density and mixed forest area boundary fuzzy area.

[0154] Thirdly, through path differentiation design and parameter sharing, the total parameter quantity of the model is controlled within 126 million, effectively improving the training and application efficiency.

[0155] In the embodiment of the present application, the model input is 4 multispectral bands of GF-7 and 1 CHM height band, and the output is the tree species category corresponding to the image element. The classification categories cover the main dominant tree species (such as larch, Chinese pine, Korean pine, white birch, oak, etc.) and other secondary tree species or non-forest land (marked as "other" category), a total of 7 categories.

[0156] As shown in Figure 6 The CBAM module (convolution block attention mechanism) of the present application contains channel attention mechanism and spatial attention mechanism. Among them, the channel attention is used to process the maximum pooling and average pooling features in parallel and apply shared MLP to extract channel attention weights; the spatial attention uses the maximum and average values of the channel dimension to generate a feature map, and applies a 7x7 convolution to generate a spatial attention map. In the embodiment of the present application, the CBAM module is applied after the bottleneck layer and at each skip connection, which can significantly enhance the important feature representation.

[0157] The DoubleConv module of the present application realizes feature layer-by-layer abstraction, non-linear modeling and channel adaptation through the combination of "two convolutions + non-linear activation + normalization", and is a basic component for improving feature expression capability in the double-path CBAM-UNet model. In the embodiment of the present application, the module is combined with ASPP, CBAM and the like, which significantly improves the distinguishing ability of similar tree species in complex forest stands (such as the classification accuracy of larch and oil pine is improved by about 5%), and optimizes the delicacy of boundary extraction. The DoubleConv module includes two 3x3 convolution layers, each of which is followed by a BatchNorm and a ReLU activation function. This design can extract more complex features while keeping the spatial size of the feature map unchanged (by setting padding=1) by stacking convolution layers. In specific implementation, the input and output channel numbers of DoubleConv can be adjusted according to network requirements, and the output channel number of the second convolution layer is usually consistent with the output channel number of the module.

[0158] In the embodiment of the present application, DoubleConv can enhance the feature expression capability of the network. Through two convolution operations, the network can learn more abstract features, and the introduction of BatchNorm and ReLU activation function helps to speed up the training and alleviate the gradient vanishing problem. This modular design makes DoubleConv can be flexibly embedded into the encoder and decoder of UNet, becoming an important part of the network structure.

[0159] Step S5: training the double-path CBAM-UNet network model using the preprocessed data.

[0160] Figure 7 The specific process of training the double-path CBAM-UNet network model of the present application is shown, which specifically includes the following steps:

[0161] Step S51: model initialization configuration.

[0162] The double-path CBAM-UNet network model constructed in the embodiment of the present application aims to realize effective semantic segmentation of multi-channel remote sensing images. The model is designed as a double-path structure, in which the left path inputs 4-channel multispectral image data and the right path inputs 1-channel CHM image data, and the overall input is 5-channel data. The final output of the model is an 8-channel classification map, which corresponds to different tree species categories and background areas.

[0163] In the model training initialization stage, the optimizer and learning rate scheduling strategy are configured. Specifically, the AdamW optimizer is used for parameter update, the momentum parameter β1 is set to 0.9, β2 is set to 0.999, and weight decay = 0.01 is set to prevent overfitting. In terms of learning rate, the initial learning rate is set to 0.001, the minimum learning rate limit is 1e-6, and the cosine annealing scheduling algorithm (CosineAnnealingLR) is used to dynamically adjust the learning rate, thereby accelerating the model convergence speed and improving the training stability.

[0164] In order to improve the adaptability of the model to the unbalanced class distribution, the loss function design adopts the combination of main loss and auxiliary loss:

[0165] Main loss function: the weighted cross-entropy loss function is used, and the weight of each class is calculated according to the inverse ratio of its sample number to alleviate the class imbalance problem.

[0166] Auxiliary loss function: combine Dice loss and cross-entropy loss, weighted combination according to the proportion of 0.4:0.6, enhance the model's ability to identify the boundary of the target region.

[0167] Specific class balance enhancement mechanism: for the target class with fewer samples (such as oak trees), introduce FocalLoss (γ = 2.0) to further improve the model's recognition effect on minority classes.

[0168] The organic combination of the above model structure and training strategy makes the present application have stronger expression ability and higher classification accuracy in complex vegetation classification tasks.

[0169] Step S52: use the prepared training sample data to train the model.

[0170] During the model training process, the data loader loads the training data and validation data with appropriate batch size. Among them, the training data is randomly shuffled to improve the generalization ability of the model, while the validation data maintains a fixed order to stably evaluate the model performance, and the data set size is recorded to provide a basis for subsequent performance index calculation. In the embodiment of the present application, the batch size in the training stage is set to 16, and the prediction stage is 30.

[0171] It can be understood that model training needs multiple cycles to be continuously optimized, and each cycle is mainly divided into two stages of training and verification. In the training stage, the model parameters are adjusted through the loss function to enable the model to learn the features in the data; in the verification stage, the model performance is evaluated by using the verification data that does not participate in the training to monitor the generalization ability of the model. In order to improve the training efficiency and stability, the mixed precision training strategy and the gradient clipping method are adopted in the embodiment of the application, and the random dropout (Dropout) is applied to enhance the generalization ability of the model, and the Dropout Rate can be set according to actual needs, and the application does not make special limitation.

[0172] In terms of model performance evaluation, the model performance is comprehensively measured by calculating the loss, pixel-level accuracy and intersection over union (IoU, based on the overlap ratio of prediction and true label), and the model weight with the highest accuracy on the verification set is saved. In addition, the training progress will be recorded during the training process, and the computing resources will be regularly cleaned to optimize the memory usage and ensure the smooth progress of the training process. In the embodiment of the application, the total training cycle is set to 100 epochs, and if the verification set accuracy does not improve for 10 consecutive epochs during the training process, the training will be terminated in advance to avoid waste of resources.

[0173] Step S53: model evaluation and optimization.

[0174] In the model training stage, the training set images are input into the double-path CBAM-UNet network for iterative training, and the verification set is used to evaluate the model performance and adjust the hyperparameters to monitor the performance of the model. Specifically, the loss value, accuracy and intersection over union (IoU) are output for each epoch to evaluate the classification and segmentation ability of the model. At the same time, the time consumption of the entire training process is recorded to analyze the efficiency and effect of the model.

[0175] Once the evaluation result of the model reaches the expected standard, we will end the training and export the model; if it does not reach the expected, the parameters will be adjusted according to the evaluation feedback and the training will continue. In the training stage, the training set images are input into the network model for iterative training, and the verification set is used to guide the adjustment of the hyperparameters.

[0176] In the embodiment of the application, after multiple rounds of training and parameter optimization, the model achieves an accuracy of 87.66% on the training set and an accuracy of 80.07% on the verification set, which indicates that the model not only has good convergence, but also exhibits strong generalization ability. Finally, the best-performing model is selected for the test set, achieving an overall classification accuracy of 81.83%, an average intersection over union (mIoU) of 70.39%, and a Kappa coefficient of 0.7803, which proves the effectiveness of the model.

[0177] When the model training meets the expected performance, the model is exported and deployed to the target device for subsequent tree species classification tasks.

[0178] Step S6: Tree species classification prediction is performed through the deployed double-path CBAM-UNet network model, and a tree species classification map is obtained.

[0179] In the embodiment of the present application, the trained double-path CBAM-UNet model is applied to the whole remote sensing image of the Magnolia test area to realize pixel-by-pixel tree species classification prediction in a large area. Due to the large area of the research area, the GF-7 and CHM images can be cut into continuous and non-overlapping small image blocks of 256x256 pixels during implementation, and the double-path CBAM-UNet model is input block by block for forward prediction to obtain the corresponding small slice classification results.

[0180] As shown in Figure 8 , tree species classification prediction and obtaining a tree species classification map specifically includes the following steps:

[0181] Step S61: Data preparation.

[0182] Collect the GF-7 and CHM images of the test area, divide the collected image data of the test area into several continuous and non-overlapping image slices, and then pre-process the sliced images. The pre-processing process has been described in detail in step S2, which will not be repeated here. At the same time, the pre-processed image data needs to be sliced, which is the same as the processing method in the training sample. The size of the image slice is also 256x256 pixels.

[0183] Step S62: Load the trained model.

[0184] Read the weight file with the highest validation set accuracy (such as.pth or.h5 format) from the model save path, and load the weight parameters into the double-path CBAM-UNet network model through a deep learning framework (such as PyTorch / TensorFlow). To improve prediction efficiency, deploy the model to a computing device equipped with a high-performance GPU such as NVIDIA RTX4090, and use CUDA to accelerate parallel computing.

[0185] After loading, set the model to evaluation mode through the model.eval() interface, and disable gradient calculation (such as using the torch.no_grad() context manager in PyTorch) to avoid unnecessary parameter update operations, reduce memory occupation, and speed up the forward propagation process.

[0186] Step S63: Slice prediction.

[0187] The pre-processed image slices are loaded into the model input queue in batches, and the batch size is set to 30 in this embodiment to balance GPU memory occupation and calculation efficiency. The model performs forward propagation on each slice, outputting a probability distribution tensor with dimensions of 256x256x7 (7 channels corresponding to the predicted probabilities of 6 target tree species and background). By applying the argmax maximum value operation, the maximum probability class index of each pixel is extracted along the channel dimension, and the continuous probability distribution is converted into a discrete pixel-level tree species classification map.

[0188] Finally, the predicted tree species classification map is saved according to the original slice position information for subsequent stitching processing.

[0189] Step S7: Image stitching of the tree species classification map to obtain a tree species classification raster map.

[0190] The prediction results of all slices are stitched and restored according to their geographical positions in the original image. During the stitching process, the results of adjacent blocks are smoothed to eliminate stitching marks and ensure the continuity and consistency of the classification map. Finally, a complete test area tree species distribution classification map is generated, and the result map is stored in raster format, with each pixel assigned a corresponding tree species category code. The specific process of image stitching includes the following steps:

[0191] Step S71: Geographical coordinate mapping and positioning.

[0192] When performing image stitching of the tree species classification map, first, the spatial reference information recorded during the slicing process, such as latitude and longitude range, row and column number, etc., is used to establish the position correspondence between each predicted slice and the original image. At the same time, a global grid index is constructed to determine the starting and ending coordinates of each slice in the complete grid map, and the overlapping area is marked to provide clear boundary information for subsequent fusion processing. This step is the basis for ensuring the accuracy of the stitching position.

[0193] Step S72: Feature fusion in overlapping areas.

[0194] When processing feature fusion in overlapping areas, there is usually a certain slice size overlap between adjacent slices. For this part of the area, a weighted average strategy can be used to achieve smooth transition. The weight is calculated by a spatial distance decay function, for example, the weight is equal to the product of the negative alpha of the natural exponential function and the square of the distance, where alpha is the adjustment parameter and distance refers to the distance from the pixel to the slice boundary. For pixels with inconsistent classification, the prediction result with higher confidence is preferred, such as the maximum probability value of the Softmax output, which makes the classification result of the overlapping area more reasonable.

[0195] Step S73: Boundary feathering and gradient smoothing.

[0196] In the process of boundary feathering and gradient smoothing, a Gaussian blur filter is applied to the edge of the slice, generally with a setting of 2-3 pixels for sigma, thereby generating a transition zone with a width of 5-10 pixels. For areas with sharp gradient changes, such as forest canopy edges, morphological dilation or erosion operations are introduced to optimize the continuity of the boundary. Moreover, using the terrain information in the CHM data, the areas with large terrain undulations are subjected to elevation-weighted fusion, thereby reducing the deviation caused by mountain shadows on the classification, and making the boundaries of the spliced images more natural and continuous.

[0197] Step S74: Raster encoding and metadata integration.

[0198] Finally, raster encoding and metadata integration need to be completed, the fused classification results are converted into standard GeoTIFF raster format, so that each pixel value corresponds to a specific tree species code, such as 1 for larch and 2 for oil pine. At the same time, spatial reference system information is embedded, including projection method and geographic coordinate system, etc., and a classification attribute table is also attached, recording the tree species name, confidence threshold and other metadata corresponding to each code, finally generating a tree species distribution raster map that is spatially continuous, has consistent attributes and has complete geographic positioning information.

[0199] Finally, the generated tree species distribution raster map is exported and saved, and relevant metadata information is attached, facilitating subsequent application and analysis.

[0200] Figure 9 An embodiment of the present application shows a structure diagram of a forest stand tree species recognition system based on multi-source remote sensing data fusion. The system 600 in the present application includes a graph data acquisition module 601, a data preprocessing module 602, a prediction module 603, and an image processing module 604.

[0201] The data acquisition module 601 is responsible for collecting forest remote sensing image data (including airborne digital images, satellite images, CHM canopy height model data) and ground data (RTK collected tree species sample points, national forest resources second-class survey plot data). Among them, the GF-7 satellite multispectral image and CHM data need to be obtained during the vegetation growth period and the laser radar collection period, and the ground data is used for model training and verification to ensure that the data covers the representative area of the test area.

[0202] The data preprocessing module 602 is responsible for radiation correction, atmospheric correction, orthorectification and panchromatic fusion of remote sensing images, and the resolution of GF-7 images is improved to 0.5 meters; CHM data is subjected to noise suppression, edge preservation and normalization processing, and is registered with optical images to unify the spatial coordinate system, and finally multi-source data fused with spectral and height information is generated, preparing for model input.

[0203] The prediction module 603 is constructed based on a double-path CBAM-UNet network model, a left path processes GF-7 multispectral data to extract spectral features, a right path processes CHM data to extract structural features, features are extracted in parallel through a four-level encoder, key features are enhanced by using a CBAM module, multiscale feature capture is optimized by combining an ASPP mode switch, spatial resolution is restored by a decoder after fusion through a bottleneck layer, and finally, a tree species classification probability distribution is output.

[0204] When working, the prediction module 603 slices the preprocessed image according to 256*256 pixels, loads the trained model weight to the GPU, performs forward propagation with a batch size of 30, converts the output probability tensor into a pixel-level classification map through argmax, records the geographic coordinates when saving the slice results, and provides position information for subsequent splicing.

[0205] The image processing module 604 performs geographic coordinate mapping based on the slice spatial reference information, fuses the overlapping areas by using distance weighted average and confidence priority strategy, smoothes the boundary through Gaussian blur and morphological operation, finally converts the classification results into GeoTIFF format, embeds the spatial reference system and integrates the metadata, and generates a complete tree species classification raster map.

[0206] In summary, the forest stand tree species identification method and system based on multi-source remote sensing data fusion has the following advantages:

[0207] First, multiscale feature capture: using GF-7 multispectral images and high-resolution CHM data, combined with attention mechanism and deep learning model of cavity pyramid pooling, multiscale feature information such as spectrum, spatial texture and vertical height of forest stand can be obtained at the same time. This enables the model to identify subtle differences between different tree species, even in tree species with similar spectral characteristics (such as pine and oil pine), which can be distinguished by crown height or texture, thereby improving classification accuracy.

[0208] Second, adaptability to complex forest land: the method achieves a classification accuracy of more than 81% in the complex and large-terrain Mulan test area, indicating its good adaptability to complex and heterogeneous forest land. The results show that whether it is a large-area continuous artificial pure forest or a natural mixed-age forest land, this method can realize stable tree species identification.

[0209] Third, automatic and efficient process: from data preprocessing, model training to result generation, a complete automatic processing flow is formed, which greatly reduces manual intervention and field work. The model automatically outputs tree species classification map and forest stand boundary through remote sensing data, without the time-consuming manual delineation of traditional methods. Compared with manual investigation, this method can complete the mapping update of large-scale forest stand in a short time, which is helpful to obtain timely forest resource change information.

[0210] Fourth, precision improvement: thanks to the feature fusion of multi-source data and the improved model algorithm, the method of the present application has obvious improvement in the recognition accuracy of tree species classification compared with the traditional method. This high-precision data product can be used as high-quality basic data for forest resource management, and provides strong support for forest management decision-making and ecological research.

[0211] Fifth, the calculation performance is improved: in the actual application test, the method of the present application runs on the NVIDIA RTX4090 system, uses the PyTorch 2.2 deep learning framework, and the complete processing time of 12,300 hectares of test area is about 2 to 3 hours, which has excellent calculation efficiency and economic benefit, and can support large-scale forest information automatic processing tasks.

[0212] In summary, the tree species identification method of the present application successfully realizes the high-precision and automatic acquisition of forest resource information based on high-resolution optical remote sensing and canopy height model. The method fully utilizes the advantages of multi-source remote sensing data, and realizes the accurate identification of complex forest tree species by combining deep learning, which has important practical significance for forest resource survey and monitoring.

[0213] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the scope of the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A forest stand tree species identification method based on multi-source remote sensing data fusion, characterized in that, The forest stand tree species recognition method based on multi-source remote sensing data fusion comprises the following steps: acquiring data, wherein the data comprises forest land remote sensing image data and ground data; preprocessing the acquired data; constructing a double-path CBAM-UNet network model; inputting the preprocessed data into the double-path CBAM-UNet network model for tree species classification prediction and obtaining a tree species classification map; performing image stitching on the tree species classification map to obtain a tree species classification raster map; the double-path CBAM-UNet network model comprises an input layer and an encoder, the input layer comprises a left path and a right path, the encoder comprises a left path encoder and a right path encoder, and the input layer is used to receive original image data and simultaneously input the original image data into the left path encoder and the right path encoder for parallel processing through the left path and the right path; wherein the left path encoder comprises: a first-level feature extraction layer comprising a BasicConv module; a second-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module; a third-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module, and performing Dropout processing after max-pooling; a fourth-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module, and performing Dropout processing after max-pooling; the right path encoder comprises: a first-level feature extraction layer comprising a BasicConv module; a second-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module, and performing Dropout processing after max-pooling; a third-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module, and performing Dropout processing after max-pooling; a fourth-level feature extraction layer comprising a max-pooling layer for down-sampling and a BasicConv module, and performing Dropout processing after max-pooling.

2. The forest stand tree species identification method based on multi-source remote sensing data fusion according to claim 1, characterized in that, The forest stand tree species recognition method based on multi-source remote sensing data fusion comprises: preparing a deep learning training sample according to the acquired data to train the double-path CBAM-UNet network model.

3. The forest stand tree species identification method based on multi-source remote sensing data fusion according to claim 2, characterized in that, The preparation of the deep learning training sample according to the acquired data to train the double-path CBAM-UNet network model comprises: preparing a training sample, including image slicing and label preparation, and performing normalization processing, feature engineering processing and data enhancement processing on the sample; model initialization configuration, including using an AdamW optimizer, a cosine annealing scheduling dynamic learning rate strategy and a weighted cross-entropy loss function as a main loss function; model training, wherein a mixed precision training strategy and a gradient clipping method are used in the process of the model training to improve the training efficiency of the model, and a random dropout strategy is applied to enhance the generalization ability of the model.

4. The forest stand tree species identification method based on multi-source remote sensing data fusion according to claim 1, characterized in that, The left path encoder is used for processing optical remote sensing image data, and the right path encoder is used for processing CHM canopy height model data, the optical remote sensing image data includes RGB band or multispectral data, and the CHM canopy height model data provides forest height information, and the complementary fusion of spectral features and structural features is realized through double-path feature extraction.

5. The forest stand tree species identification method based on multi-source remote sensing data fusion according to claim 1, characterized in that, The double-path CBAM-UNet network model sets a mode switch to control the working mode of the BasicConv module, and the working mode includes a standard mode and an ASPP enhanced mode. When in the standard mode, the input features are output after being processed by the basic convolution unit. When in the ASPP enhanced mode, the input features are processed in parallel by the basic convolution unit and the ASPP sub-module, and the two processing results are fused by element-wise addition, and the fused features are processed by batch normalization and ReLU activation function, and then output after applying Dropout regularization.

6. The forest stand tree species identification method based on multi-source remote sensing data fusion according to claim 1, characterized in that, The double-path CBAM-UNet network model includes a decoder, an output layer and a CBAM module. The decoder is a four-layer structure, each layer of which includes a transpose convolution module for upsampling and a DoubleConv module for processing fused features. The output layer includes a convolution layer that maps the output features of the decoder to predefined tree species categories and applies a Softmax activation function to generate a probability distribution for each tree species category. The CBAM module is used to enhance the fused features after each feature extraction of the left path encoder and the right path encoder, and the enhanced features are fused with the output features of the transpose convolution module of the corresponding layer of the decoder through a jump connection, and the fused features are input into the DoubleConv module for processing.

7. The forest stand tree species recognition method based on multi-source remote sensing data fusion according to claim 1, characterized in that, The double-path CBAM-UNet network model includes a bottleneck layer, which includes: a left path DoubleConv module for processing feature maps output by the left path encoder; a right path DoubleConv module for processing feature maps output by the right path encoder; a bottleneck fusion layer that concatenates the feature maps output by the left path DoubleConv module and the right path DoubleConv module in the channel dimension to generate concatenated feature maps; a terminal enhancement module that converts the concatenated features into a highly nonlinear representation; a CBAM enhancement module that applies attention mechanisms to the output of the bottleneck layer for enhanced processing.

8. The forest stand tree species recognition method based on multi-source remote sensing data fusion according to any one of claims 1 to 7, characterized in that, The forest land remote sensing image data includes airborne digital images, satellite images and CHM canopy height model data, and the preprocessing of the obtained data includes: radiometric correction, atmospheric correction, orthorectification and panchromatic fusion of the airborne digital images and the satellite images; noise suppression, edge preservation and data normalization processing of the CHM canopy height model data, and image registration processing of the processed CHM canopy height model data and optical images to realize the unified alignment of the spatial coordinate system.

9. A forest stand tree species identification system based on multi-source remote sensing data fusion, characterized in that, includes: The data acquisition module is configured to collect forest remote sensing image data and ground data. The data preprocessing module is configured to preprocess the forest remote sensing image data and ground data collected by the data acquisition module. The prediction module is configured to perform tree species classification prediction according to the input image slices and obtain a tree species classification map. The preprocessed data is input into a pre-constructed double-path CBAM-UNet network model to perform tree species classification prediction and obtain a tree species classification map. The image processing module is configured to perform stitching processing on the tree species classification map to obtain a complete tree species classification raster map. The double-path CBAM-UNet network model includes an input layer and an encoder. The input layer includes a left path and a right path. The encoder includes a left path encoder and a right path encoder. The input layer is configured to receive original image data and simultaneously input the original image data into the left path encoder and the right path encoder for parallel processing through the left path and the right path. The left path encoder includes: A first-level feature extraction layer including a BasicConv module. A second-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. A third-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. A fourth-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. The right path encoder includes: A first-level feature extraction layer including a BasicConv module. A second-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. A third-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. A fourth-level feature extraction layer including a max-pooling layer for down-sampling and a BasicConv module. The fourth-level feature extraction layer includes a max-pooling layer for down-sampling and a BasicConv module.

Images