A method, system and computer program product for detecting changes in vegetation types in forest areas

Abstract

The present invention discloses a forest vegetation type change detection method, system and computer program product, including: obtaining multi-phase high spatial resolution remote sensing image data of forest area; constructing a multi-task full convolutional neural network model based on a convolutional neural network model and multi-task learning according to the multi-phase high spatial resolution remote sensing image data of forest area; constructing a sample data set based on the multi-phase high spatial resolution remote sensing image data of forest area, training the multi-task full convolutional neural network model, and obtaining a forest vegetation type change detection model; inputting the forest vegetation type change detection model into the forest vegetation type change detection model, and obtaining a forest vegetation type change detection result. The technical solution provided by the present invention can use the multi-phase high spatial resolution remote sensing image data of forest area to construct a multi-task full convolutional neural network model, detect the change of forest vegetation type, and realize the accurate detection of vegetation coverage change.

Description

Technical Field

[0001] The present invention relates to the field of forest area dynamic change detection, and in particular to a forest area vegetation type change detection method. Background Art

[0002] Dynamic change detection in forest areas can timely understand the quantity and quality of forest resources, grasp the laws and trends of the growth and decline of forest resources, analyze the natural, economic, and social objective conditions that affect and restrict forest growth, and establish or update forest resource archives. It is of great significance to revealing ecosystem environmental changes, vegetation restoration, and reconstruction layout.

[0003] Remote sensing technology provides a new means for detecting dynamic changes in forest areas. Compared with traditional forest resource inventory, remote sensing can obtain a large amount of forest image data with different spatial resolutions and multiple temporal phases, providing a rich data source for forest change detection. Through high-resolution imaging spectral equipment, multi-temporal and high-spatial resolution remote sensing image data of forest areas can be obtained to detect changes in vegetation types in forest areas. Summary of the invention

[0004] The present invention provides a method for detecting changes in vegetation types in forest areas, which can utilize multi-temporal and high-spatial-resolution remote sensing image data of forest areas to construct a multi-task fully convolutional neural network model to detect changes in vegetation types in forest areas.

[0005] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0006] A method for detecting changes in forest vegetation types, comprising the following steps:

[0007] Acquire multi-temporal and high spatial resolution remote sensing image data of forest areas;

[0008] According to the multi-temporal high-spatial resolution remote sensing image data of the forest area, a multi-task full convolutional neural network model was constructed based on the convolutional neural network model and multi-task learning;

[0009] Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a sample data set was constructed, and a multi-task fully convolutional neural network model was trained to obtain a forest vegetation type change detection model.

[0010] The multi-temporal and high-spatial-resolution remote sensing image data of the forest area are input into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results.

[0011] According to one aspect of the present invention, the method for detecting changes in forest vegetation types further includes: preprocessing multi-temporal and high spatial resolution remote sensing image data of the forest area.

[0012] According to one aspect of the present invention, the preprocessing of multi-temporal and high spatial resolution remote sensing image data of forest areas includes: radiation calibration, atmospheric correction and image registration.

[0013] According to one aspect of the present invention, the multi-task full convolutional neural network model is constructed based on the convolutional neural network model and multi-task learning, including:

[0014] A hierarchical feature extraction network is constructed as an encoder to conduct in-depth analysis of the complex spatial information contained in high spatial resolution remote sensing image data, integrate the spatiotemporal attention mechanism, integrate the change information of the time dimension, and extract comprehensive multi-scale spatiotemporal features from multi-phase high spatial resolution remote sensing image data.

[0015] According to one aspect of the present invention, the multi-task full convolutional neural network model constructed based on the convolutional neural network model and multi-task learning also includes:

[0016] A converter is built based on the Transformer architecture, and a spatiotemporal feature integration module is designed. The self-attention mechanism is used, and a multi-task learning strategy for change detection and semantic segmentation is adopted to refine and strengthen the multi-scale spatial features of the encoder output and identify changes in vegetation cover over time.

[0017] According to one aspect of the present invention, the multi-task full convolutional neural network model constructed based on the convolutional neural network model and multi-task learning also includes:

[0018] The decoder is built based on multi-task processing to perform change detection and semantic segmentation of vegetation cover. The self-attention mechanism is embedded to conduct in-depth analysis of spatiotemporal features, optimize the feature fusion process, accurately analyze feature maps based on rich details, fuse the multi-scale spatiotemporal features output by the encoder and the composite features of vegetation cover changes over time output by the converter, and accurately detect vegetation cover changes.

[0019] According to one aspect of the present invention, when constructing a sample data set, data enhancement is performed on image data, and the sample data set is constructed based on a time window sliding technology.

[0020] According to one aspect of the present invention, the training of the multi-task fully convolutional neural network model includes:

[0021] A fusion loss function is used to integrate the cross entropy loss of semantic segmentation and the binary cross entropy loss of change detection. An adjustable weight coefficient is introduced to dynamically adjust the weight according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance.

[0022] Use the Adam optimizer with adaptive learning rate to improve training efficiency and model convergence speed;

[0023] Integrate dynamic loss weight adjustment and learning rate scheduling strategies to help the model maintain a balance in task learning between training stages and optimize the overall performance of the model;

[0024] The early stopping mechanism and model regularization measures are adopted to prevent overfitting, enhance the generalization ability of the model, and ensure the stability of the training process.

[0025] According to one aspect of the present invention, the training of the multi-task fully convolutional neural network model further comprises:

[0026] Use multiple indicators, including but not limited to precision, recall, F1 score, and overall accuracy, to comprehensively evaluate model performance;

[0027] Based on the comprehensive evaluation results, the network model is fine-tuned.

[0028] According to one aspect of the present invention, the forest vegetation type change detection method further includes the following steps:

[0029] Based on the detection results of forest vegetation type changes, a spatiotemporal distribution map of forest vegetation type changes was constructed.

[0030] A forest vegetation type change detection system, based on the forest vegetation type change detection method as described above, comprises:

[0031] Data acquisition module, used to obtain multi-temporal high spatial resolution remote sensing image data of forest areas;

[0032] The model building module is used to build a multi-task full convolutional neural network model based on the convolutional neural network model and multi-task learning according to the multi-temporal high-spatial resolution remote sensing image data of the forest area;

[0033] The training module is used to construct a sample data set based on multi-temporal and high-spatial-resolution remote sensing image data of forest areas, train a multi-task fully convolutional neural network model, and obtain a forest vegetation type change detection model;

[0034] The detection module is used to input the multi-temporal and high-spatial-resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results.

[0035] A computer program product comprises a computer program, wherein when the computer program is executed, the steps of the above-mentioned method for detecting changes in forest vegetation types are implemented.

[0036] Advantages of the present invention:

[0037] The present invention provides a method for detecting changes in vegetation types in forest areas, which can utilize multi-temporal and high-spatial-resolution remote sensing image data of forest areas to construct a multi-task fully convolutional neural network model, detect changes in vegetation types in forest areas, and achieve accurate detection of changes in vegetation coverage.

[0038] This method adopts an encoder-converter-decoder architecture combined with a multi-task learning strategy to achieve collaborative processing of semantic segmentation and vegetation cover change detection; by fusing multi-temporal high-spatial resolution remote sensing data with a multi-task fully convolutional neural network, accurate monitoring of vegetation cover changes can be achieved.

[0039] This method can also conduct a comprehensive evaluation of the network model based on multiple indicators, and fine-tune and improve the network model according to the evaluation results; according to the detection results of forest vegetation type changes, the data is integrated to construct a spatiotemporal distribution map of forest vegetation type changes, providing important information for environmental monitoring and decision-making analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is a flow chart of a method for detecting changes in vegetation types in a forest area according to Embodiment 1 of the present invention;

[0042] Figure 2 This is a flow chart of a method for detecting changes in forest vegetation types described in Example 2 of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Embodiment 1

[0045] like Figure 1 As shown, a method for detecting changes in forest vegetation types includes the following steps:

[0046] S1: Acquire multi-temporal and high spatial resolution remote sensing image data of forest areas.

[0047] Acquire multi-temporal and high-spatial-resolution remote sensing data to ensure that the data used can effectively reflect changes in vegetation cover and provide a basis for model construction and analysis.

[0048] The GF-2 and GF-1 PMS images are selected as the main data sources, which can provide high spatial resolution surface feature information and high accessibility. Among them, GF-2 images have high spatial resolution and can provide detailed vegetation coverage features; while GF-1 PMS images effectively supplement the problem of insufficient coverage of GF-2 in specific areas and enhance the data foundation.

[0049] The method for detecting changes in forest vegetation types also includes: preprocessing multi-temporal and high spatial resolution remote sensing image data of the forest area.

[0050] The preprocessing of multi-temporal high-spatial-resolution remote sensing image data of forest areas includes: radiation calibration, atmospheric correction and image registration.

[0051] Radiometric calibration is to convert the original digital image values ​​of remote sensing images into surface reflectivity, correct variations caused by differences in sensor performance and imaging angles, and improve data accuracy and consistency.

[0052] Atmospheric correction is to eliminate the impact of atmospheric components on remote sensing data and ensure that the data can accurately reflect the actual situation on the surface.

[0053] Image registration is to achieve spatial consistency of multi-temporal remote sensing data, so that multi-temporal remote sensing data can achieve spatial correspondence and provide a data basis for subsequent change detection.

[0054] S2: Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a multi-task fully convolutional neural network model is constructed based on the convolutional neural network model and multi-task learning.

[0055] The multi-task fully convolutional neural network model mainly consists of three parts: encoder, converter and decoder.

[0056] The multi-task full convolutional neural network model is constructed based on the convolutional neural network model and multi-task learning, including:

[0057] S21: Construct a hierarchical feature extraction network as an encoder to perform in-depth analysis of the complex spatial information contained in high spatial resolution remote sensing image data, integrate the spatiotemporal attention mechanism, integrate the change information of the time dimension, and extract comprehensive multi-scale spatiotemporal features from multi-phase high spatial resolution remote sensing image data.

[0058] The encoder uses a deep learning framework, such as U-Net, a model designed specifically for semantic segmentation tasks. By applying hierarchical feature extraction technology, the complex spatial information contained in high-resolution remote sensing data is deeply analyzed, thereby achieving overall recognition of vegetation coverage areas from local to global. In order to adapt to the variability of remote sensing images in spatial scale, the encoder extracts features at multiple spatial scales to ensure accurate recognition of vegetation coverage features under different spatial resolution conditions. Considering the multi-temporal properties of the input data, the encoder further designs a fusion spatiotemporal attention mechanism to effectively integrate the change information in the time dimension in order to accurately reveal the dynamic process of vegetation changes over time.

[0059] During the implementation process, the preprocessed multi-temporal high-spatial resolution remote sensing data is first subdivided into small blocks, and a convolutional neural network architecture such as EfficientNet or U-Net is used as the basic framework to build a hierarchical feature extraction network. This network design can extract multi-scale features at different levels, and then build a hierarchical feature system. The integration of features at different levels is achieved through the feature pyramid network technology. At the same time, by integrating the spatiotemporal attention mechanism, such as Transformer or a specially designed spatiotemporal fusion module, the significant changes in multi-temporal remote sensing data are focused on to more accurately identify vegetation features that change over time. With the help of this deep learning architecture, the encoder can effectively extract comprehensive multi-scale and spatiotemporal features from multi-temporal high-spatial resolution remote sensing data, providing a solid feature foundation for vegetation cover change detection and classification tasks.

[0060] S22: Build a converter based on the Transformer architecture, design a spatiotemporal feature integration module, use the self-attention mechanism, and adopt a multi-task learning strategy for change detection and semantic segmentation to refine and strengthen the multi-scale spatial features of the encoder output to identify changes in vegetation cover over time.

[0061] The key to the design of the converter is to refine and strengthen the multi-scale spatial features obtained from the encoder based on the Transformer architecture, with the goal of accurately identifying changes in vegetation cover over time. This architecture uses the Transformer mechanism to achieve efficient integration of features at different spatial scales, ensuring that key vegetation cover change information can be identified. The built-in self-attention mechanism enhances the model's ability to interpret complex relationships between features, effectively reveals the spatial change pattern of vegetation cover, and is an indispensable component in the task of vegetation cover change detection.

[0062] In the implementation process, the converter accurately processes the multi-scale spatial feature maps output by the encoder by designing a spatiotemporal feature integration module to identify the temporal changes in vegetation cover. The spatiotemporal feature integration module uses the self-attention mechanism to fuse and analyze the features of the same geographical location at different time points, enhancing the perception of temporal changes, especially in identifying significant change areas, while retaining key spatial details and extracting key spatiotemporal change features. By implementing a multi-task learning strategy for change detection and semantic segmentation, the converter not only independently processes each detection target, but also optimizes the detection performance through the information exchange mechanism between tasks.

[0063] S23: Construct a decoder based on multi-task processing to perform change detection and semantic segmentation on vegetation cover, embed a self-attention mechanism, conduct in-depth analysis of spatiotemporal features, optimize the feature fusion process, accurately analyze feature maps based on rich details, fuse the multi-scale spatiotemporal features output by the encoder and the composite features of vegetation cover changes over time output by the converter, and accurately detect vegetation cover changes.

[0064] The decoder design is based on the principle of multi-tasking, aiming to accurately perform vegetation cover change detection and semantic segmentation. The design comprehensively applies the multi-layer spatial features and spatiotemporal features generated by the encoder and converter to enhance the accurate detection of vegetation cover changes. Through the embedded self-attention mechanism, the analysis of spatiotemporal features is deepened, and the feature fusion process is optimized to ensure accurate analysis based on detailed feature maps. At the same time, the decoder adopts a specific multi-task learning strategy to independently handle detection and classification tasks by customizing the task-related output layers, and establishes an information exchange mechanism to promote mutual learning between tasks.

[0065] In implementation, the decoder will fuse the composite features from the encoder and converter, and accurately analyze through the self-attention mechanism to highlight key changes and details. Applying the multi-task learning framework, the decoder is equipped with specific output layers to independently optimize each task for vegetation change detection and semantic segmentation. An information exchange mechanism is set up to promote mutual learning between tasks and improve the overall performance of the model. The performance of the decoder is continuously monitored and evaluated. Based on performance feedback, the parameters of the self-attention mechanism, the weight distribution of the multi-task learning strategy, and the details of feature fusion and information exchange are adjusted to maximize the efficiency and accuracy of the decoder.

[0066] S3: Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a sample dataset is constructed, and a multi-task fully convolutional neural network model is trained to obtain a forest vegetation type change detection model.

[0067] When constructing the sample data set, the image data is enhanced and the sample data set is constructed based on the time window sliding technology.

[0068] The sample data construction first generates a sample set reflecting the change spots based on the existing remote sensing images and spot data. The calibrated samples include the vegetation cover type before and after and the change type samples.

[0069] Using two periods of high spatial resolution remote sensing images from the same sensor in adjacent quarters, vegetation cover remote sensing image change calibration is carried out by province or region based on remote sensing image features and vegetation cover changes. According to the vegetation cover type reflected by the attributes of fixed sample plots or forest map patches, the remote sensing image features of the location are used for comparative analysis, and representative plots with relatively consistent remote sensing image features are selected. The vegetation cover type before and after, remote sensing data source, remote sensing data acquisition time, spatial resolution and other attribute information are calibrated according to a 512m×512m square vector frame.

[0070] Based on the established remote sensing images and data labels of each vegetation cover type, combined with the remote sensing images and forest and grass resource data of the previous and next two periods, and referring to the attribute information of forest and grass resource spots, the characteristic change areas of the remote sensing images of the previous and next two periods are compared and determined, and the change type and the corresponding remote sensing data source, remote sensing data acquisition time, spatial resolution and other attribute information are calibrated according to the 512m×512m square vector frame. Specific change types include but are not limited to construction project use, logging, reclamation, disaster events, and ecological protection and restoration measures.

[0071] This step aims to establish a large-scale, diverse and spatially distributed sample database to support iterative training and parameter optimization of deep learning models. The goal is to increase the training accuracy of the model to more than 85%. Models that do not meet this accuracy standard will be inspected in detail, supplemented with labels for interpretation signs and change types, and training parameters will be adjusted and iterative training will be repeated to achieve higher accuracy.

[0072] The data annotation strategy focuses on the identification of changes in remote sensing images at different time points for each pixel and the accurate classification of vegetation types. By comparing image data at two or more time points, we can determine the change status of a specific pixel and the vegetation cover type before and after the change. This method aims to simplify the data annotation process while ensuring that the model can accurately learn to identify vegetation change patterns in multi-temporal data, providing effective data for subsequent accurate change detection and semantic segmentation tasks.

[0073] The training of the multi-task fully convolutional neural network model includes:

[0074] A fusion loss function is used to integrate the cross entropy loss of semantic segmentation and the binary cross entropy loss of change detection. An adjustable weight coefficient is introduced to dynamically adjust the weight according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance.

[0075] Use the Adam optimizer with adaptive learning rate to improve training efficiency and model convergence speed;

[0076] Integrate dynamic loss weight adjustment and learning rate scheduling strategies to help the model maintain a balance in task learning between training stages and optimize the overall performance of the model;

[0077] The early stopping mechanism and model regularization measures are adopted to prevent overfitting, enhance the generalization ability of the model, and ensure the stability of the training process.

[0078] Model training includes multiple steps such as loss function selection, optimizer selection and learning rate management.

[0079] In order to effectively handle the dual tasks of semantic segmentation and change detection, this method adopts a fusion loss function. This function combines the cross entropy loss of semantic segmentation with the binary cross entropy loss of change detection, and introduces adjustable weight coefficients. These coefficients allow the model to dynamically adjust the weights according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance.

[0080] When training the model, the Adam optimizer with adaptive learning rate is selected to accelerate model convergence and improve efficiency. By integrating dynamic loss weight adjustment and learning rate scheduling strategy, the model is assisted to maintain a balance of task learning between each training stage and optimize the overall model performance. To prevent overfitting, early stopping mechanism and model regularization measures such as Dropout and batch normalization are adopted to enhance the generalization ability of the model and ensure the stability of the training process.

[0081] S4: Input the multi-temporal and high spatial resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results.

[0082] This method focuses on two types of change detection in forest areas. First, according to the vegetation coverage type classification system of the "National Technical Regulations for Comprehensive Monitoring of Forest and Grass Ecology", combined with the specific vegetation coverage types of the project technology integration demonstration base, the change patches within the vegetation coverage are extracted; second, according to the remote sensing interpretation and verification change types in the "National Technical Regulations for Comprehensive Monitoring of Forest and Grass Ecology", the change types of the change patches obtained in the previous step are extracted, including construction project use, logging, reclamation, disasters, and ecological protection and restoration measures.

[0083] The output of the detection model of this method consists of two parts: one is the semantic segmentation result for each pixel, that is, identifying which vegetation type each pixel belongs to; the other is the change detection result, which identifies whether each pixel has changed between the time points under investigation and the nature of the change. These two parts of the output together constitute the model's comprehensive analysis of multi-temporal remote sensing data, which can provide rich information for subsequent application scenarios.

[0084] The beneficial effects of this embodiment are as follows: this method adopts an encoder-converter-decoder architecture, combined with a multi-task learning strategy, to achieve collaborative processing of semantic segmentation and vegetation cover change detection; by fusing multi-temporal high-spatial resolution remote sensing data with a multi-task fully convolutional neural network, accurate monitoring of vegetation cover changes is achieved.

[0085] Embodiment 2

[0086] like Figure 2 As shown, a method for detecting changes in forest vegetation types includes the following steps:

[0087] S1: Acquire multi-temporal and high spatial resolution remote sensing image data of forest areas.

[0088] Acquire multi-temporal and high-spatial-resolution remote sensing data to ensure that the data used can effectively reflect changes in vegetation cover and provide a basis for model construction and analysis.

[0089] The GF-2 and GF-1 PMS images are selected as the main data sources, which can provide high spatial resolution surface feature information and high accessibility. Among them, GF-2 images have high spatial resolution and can provide detailed vegetation coverage features; while GF-1 PMS images effectively supplement the problem of insufficient coverage of GF-2 in specific areas and enhance the data foundation.

[0090] The method for detecting changes in forest vegetation types also includes: preprocessing multi-temporal and high spatial resolution remote sensing image data of the forest area.

[0091] The preprocessing of multi-temporal high-spatial-resolution remote sensing image data of forest areas includes: radiation calibration, atmospheric correction and image registration.

[0092] Radiometric calibration is to convert the original digital image values ​​of remote sensing images into surface reflectivity, correct variations caused by differences in sensor performance and imaging angles, and improve data accuracy and consistency.

[0093] Atmospheric correction is to eliminate the impact of atmospheric components on remote sensing data and ensure that the data can accurately reflect the actual situation on the surface.

[0094] Image registration is to achieve spatial consistency of multi-temporal remote sensing data, so that multi-temporal remote sensing data can achieve spatial correspondence and provide a data basis for subsequent change detection.

[0095] S2: Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a multi-task fully convolutional neural network model is constructed based on the convolutional neural network model and multi-task learning.

[0096] The multi-task fully convolutional neural network model mainly consists of three parts: encoder, converter and decoder.

[0097] The multi-task full convolutional neural network model is constructed based on the convolutional neural network model and multi-task learning, including:

[0098] S21: Construct a hierarchical feature extraction network as an encoder to perform in-depth analysis of the complex spatial information contained in high spatial resolution remote sensing image data, integrate the spatiotemporal attention mechanism, integrate the change information of the time dimension, and extract comprehensive multi-scale spatiotemporal features from multi-phase high spatial resolution remote sensing image data.

[0099] The encoder uses a deep learning framework, such as U-Net, a model designed specifically for semantic segmentation tasks. By applying hierarchical feature extraction technology, the complex spatial information contained in high-resolution remote sensing data is deeply analyzed, thereby achieving overall recognition of vegetation coverage areas from local to global. In order to adapt to the variability of remote sensing images in spatial scale, the encoder extracts features at multiple spatial scales to ensure accurate recognition of vegetation coverage features under different spatial resolution conditions. Considering the multi-temporal properties of the input data, the encoder further designs a fusion spatiotemporal attention mechanism to effectively integrate the change information in the time dimension in order to accurately reveal the dynamic process of vegetation changes over time.

[0100] During the implementation process, the preprocessed multi-temporal high-spatial resolution remote sensing data is first subdivided into small blocks, and a convolutional neural network architecture such as EfficientNet or U-Net is used as the basic framework to build a hierarchical feature extraction network. This network design can extract multi-scale features at different levels, and then build a hierarchical feature system. The integration of features at different levels is achieved through the feature pyramid network technology. At the same time, by integrating the spatiotemporal attention mechanism, such as Transformer or a specially designed spatiotemporal fusion module, the significant changes in multi-temporal remote sensing data are focused on to more accurately identify vegetation features that change over time. With the help of this deep learning architecture, the encoder can effectively extract comprehensive multi-scale and spatiotemporal features from multi-temporal high-spatial resolution remote sensing data, providing a solid feature foundation for vegetation cover change detection and classification tasks.

[0101] S22: Build a converter based on the Transformer architecture, design a spatiotemporal feature integration module, use the self-attention mechanism, and adopt a multi-task learning strategy for change detection and semantic segmentation to refine and strengthen the multi-scale spatial features of the encoder output to identify changes in vegetation cover over time.

[0102] The key to the design of the converter is to refine and strengthen the multi-scale spatial features obtained from the encoder based on the Transformer architecture, with the goal of accurately identifying changes in vegetation cover over time. This architecture uses the Transformer mechanism to achieve efficient integration of features at different spatial scales, ensuring that key vegetation cover change information can be identified. The built-in self-attention mechanism enhances the model's ability to interpret complex relationships between features, effectively reveals the spatial change pattern of vegetation cover, and is an indispensable component in the task of vegetation cover change detection.

[0103] In the implementation process, the converter accurately processes the multi-scale spatial feature maps output by the encoder by designing a spatiotemporal feature integration module to identify the temporal changes in vegetation cover. The spatiotemporal feature integration module uses the self-attention mechanism to fuse and analyze the features of the same geographical location at different time points, enhancing the perception of temporal changes, especially in identifying significant change areas, while retaining key spatial details and extracting key spatiotemporal change features. By implementing a multi-task learning strategy for change detection and semantic segmentation, the converter not only independently processes each detection target, but also optimizes the detection performance through the information exchange mechanism between tasks.

[0104] S23: Construct a decoder based on multi-task processing to perform change detection and semantic segmentation on vegetation cover, embed a self-attention mechanism, conduct in-depth analysis of spatiotemporal features, optimize the feature fusion process, accurately analyze feature maps based on rich details, fuse the multi-scale spatiotemporal features output by the encoder and the composite features of vegetation cover changes over time output by the converter, and accurately detect vegetation cover changes.

[0105] The decoder design is based on the principle of multi-tasking, aiming to accurately perform vegetation cover change detection and semantic segmentation. The design comprehensively applies the multi-layer spatial features and spatiotemporal features generated by the encoder and converter to enhance the accurate detection of vegetation cover changes. Through the embedded self-attention mechanism, the analysis of spatiotemporal features is deepened, and the feature fusion process is optimized to ensure accurate analysis based on detailed feature maps. At the same time, the decoder adopts a specific multi-task learning strategy to independently handle detection and classification tasks by customizing the task-related output layers, and establishes an information exchange mechanism to promote mutual learning between tasks.

[0106] In implementation, the decoder will fuse the composite features from the encoder and converter, and accurately analyze through the self-attention mechanism to highlight key changes and details. Applying the multi-task learning framework, the decoder is equipped with specific output layers to independently optimize each task for vegetation change detection and semantic segmentation. An information exchange mechanism is set up to promote mutual learning between tasks and improve the overall performance of the model. The performance of the decoder is continuously monitored and evaluated. Based on performance feedback, the parameters of the self-attention mechanism, the weight distribution of the multi-task learning strategy, and the details of feature fusion and information exchange are adjusted to maximize the efficiency and accuracy of the decoder.

[0107] S3: Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a sample dataset is constructed, and a multi-task fully convolutional neural network model is trained to obtain a forest vegetation type change detection model.

[0108] When constructing the sample data set, the image data is enhanced and the sample data set is constructed based on the time window sliding technology.

[0109] The sample data construction first generates a sample set reflecting the change spots based on the existing remote sensing images and spot data. The calibrated samples include the vegetation cover type before and after and the change type samples.

[0110] Using two periods of high spatial resolution remote sensing images from the same sensor in adjacent quarters, vegetation cover remote sensing image change calibration is carried out by province or region based on remote sensing image features and vegetation cover changes. According to the vegetation cover type reflected by the attributes of fixed sample plots or forest map patches, the remote sensing image features of the location are used for comparative analysis, and representative plots with relatively consistent remote sensing image features are selected. The vegetation cover type before and after, remote sensing data source, remote sensing data acquisition time, spatial resolution and other attribute information are calibrated according to a 512m×512m square vector frame.

[0111] Based on the established remote sensing images and data labels of each vegetation cover type, combined with the remote sensing images and forest and grass resource data of the previous and next two periods, and referring to the attribute information of forest and grass resource spots, the characteristic change areas of the remote sensing images of the previous and next two periods are compared and determined, and the change type and the corresponding remote sensing data source, remote sensing data acquisition time, spatial resolution and other attribute information are calibrated according to the 512m×512m square vector frame. Specific change types include but are not limited to construction project use, logging, reclamation, disaster events, and ecological protection and restoration measures.

[0112] This step aims to establish a large-scale, diverse and spatially distributed sample database to support iterative training and parameter optimization of deep learning models. The goal is to increase the training accuracy of the model to more than 85%. Models that do not meet this accuracy standard will be inspected in detail, supplemented with labels for interpretation signs and change types, and training parameters will be adjusted and iterative training will be repeated to achieve higher accuracy.

[0113] The data annotation strategy focuses on the identification of changes in remote sensing images at different time points for each pixel and the accurate classification of vegetation types. By comparing image data at two or more time points, we can determine the change status of a specific pixel and the vegetation cover type before and after the change. This method aims to simplify the data annotation process while ensuring that the model can accurately learn to identify vegetation change patterns in multi-temporal data, providing effective data for subsequent accurate change detection and semantic segmentation tasks.

[0114] The training of the multi-task fully convolutional neural network model includes:

[0115] A fusion loss function is used to integrate the cross entropy loss of semantic segmentation and the binary cross entropy loss of change detection. An adjustable weight coefficient is introduced to dynamically adjust the weight according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance.

[0116] Use the Adam optimizer with adaptive learning rate to improve training efficiency and model convergence speed;

[0117] Integrate dynamic loss weight adjustment and learning rate scheduling strategies to help the model maintain a balance in task learning between training stages and optimize the overall performance of the model;

[0118] Use early stopping mechanism and model regularization measures to prevent overfitting, enhance the generalization ability of the model, and ensure the stability of the training process;

[0119] Use multiple indicators, including but not limited to precision, recall, F1 score, and overall accuracy, to comprehensively evaluate model performance;

[0120] Based on the comprehensive evaluation results, the network model is fine-tuned.

[0121] Model training includes multiple steps such as loss function selection, optimizer selection and learning rate management.

[0122] In order to effectively handle the dual tasks of semantic segmentation and change detection, this method adopts a fusion loss function. This function combines the cross entropy loss of semantic segmentation with the binary cross entropy loss of change detection, and introduces adjustable weight coefficients. These coefficients allow the model to dynamically adjust the weights according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance.

[0123] When training the model, the Adam optimizer with adaptive learning rate is selected to accelerate model convergence and improve efficiency. By integrating dynamic loss weight adjustment and learning rate scheduling strategy, the model is assisted to maintain a balance of task learning between each training stage and optimize the overall model performance. To prevent overfitting, early stopping mechanism and model regularization measures such as Dropout and batch normalization are adopted to enhance the generalization ability of the model and ensure the stability of the training process.

[0124] In order to accurately evaluate the performance of the model, multiple indicators are used, including precision, recall rate, F1 score, and overall accuracy, etc. Based on the comprehensive evaluation results, the network model is fine-tuned and improved.

[0125] S4: Input the multi-temporal and high spatial resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results.

[0126] This method focuses on two types of change detection in forest areas. First, according to the vegetation coverage type classification system of the "National Technical Regulations for Comprehensive Monitoring of Forest and Grass Ecology", combined with the specific vegetation coverage types of the project technology integration demonstration base, the change patches within the vegetation coverage are extracted; second, according to the remote sensing interpretation and verification change types in the "National Technical Regulations for Comprehensive Monitoring of Forest and Grass Ecology", the change types of the change patches obtained in the previous step are extracted, including construction project use, logging, reclamation, disasters, and ecological protection and restoration measures.

[0127] The output of the detection model of this method consists of two parts: one is the semantic segmentation result for each pixel, that is, identifying which vegetation type each pixel belongs to; the other is the change detection result, which identifies whether each pixel has changed between the time points under investigation and the nature of the change. These two parts of the output together constitute the model's comprehensive analysis of multi-temporal remote sensing data, which can provide rich information for subsequent application scenarios.

[0128] S5: Based on the detection results of forest vegetation type changes, a spatiotemporal distribution map of forest vegetation type changes is constructed.

[0129] After obtaining the forest vegetation type change detection results using the forest vegetation type change detection model, these data are integrated to construct a spatiotemporal distribution map showing the distribution of forest vegetation type changes and their time process, providing important information for environmental monitoring and decision-making analysis.

[0130] The beneficial effects of this embodiment are: this method can also conduct a comprehensive evaluation of the network model based on multiple indicators, and fine-tune and improve the network model according to the evaluation results; according to the detection results of the changes in forest vegetation types, the data is integrated to construct a spatiotemporal distribution map of the changes in forest vegetation types, which provides important information for environmental monitoring and decision-making analysis.

[0131] Embodiment 3

[0132] A forest vegetation type change detection system, based on the forest vegetation type change detection method as described in Embodiment 1 or 2, comprises:

[0133] Data acquisition module, used to obtain multi-temporal high spatial resolution remote sensing image data of forest areas;

[0134] The model building module is used to build a multi-task full convolutional neural network model based on the convolutional neural network model and multi-task learning according to the multi-temporal high-spatial resolution remote sensing image data of the forest area;

[0135] The training module is used to construct a sample data set based on multi-temporal and high-spatial-resolution remote sensing image data of forest areas, train a multi-task fully convolutional neural network model, and obtain a forest vegetation type change detection model;

[0136] The detection module is used to input the multi-temporal and high-spatial-resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results.

[0137] Preferably, the system further comprises:

[0138] The preprocessing module is used to preprocess the multi-temporal and high spatial resolution remote sensing image data of the forest area;

[0139] The distribution map construction module is used to construct a spatiotemporal distribution map of forest vegetation type changes based on the forest vegetation type change detection results.

[0140] Embodiment 4

[0141] A computer program, which, when executed, implements the steps of the forest vegetation type change detection method as described in embodiment one or two.

[0142] Embodiment 5

[0143] A readable storage medium stores a computer program as described in Example 4, and when the computer program is executed, the steps of the forest vegetation type change detection method as described in Example 1 or 2 are implemented.

[0144] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with the art within the technical scope disclosed in the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for detecting changes in vegetation types in forest areas, characterized in that: The following steps are involved: Acquire multi-temporal and high spatial resolution remote sensing image data of forest areas; According to the multi-temporal high-spatial resolution remote sensing image data of the forest area, a multi-task full convolutional neural network model was constructed based on the convolutional neural network model and multi-task learning; Based on the multi-temporal and high-spatial-resolution remote sensing image data of the forest area, a sample data set was constructed, and a multi-task fully convolutional neural network model was trained to obtain a forest vegetation type change detection model. Input the multi-temporal high spatial resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection results; The multi-task full convolutional neural network model constructed based on the convolutional neural network model and multi-task learning includes: A hierarchical feature extraction network is constructed as an encoder to deeply analyze the complex spatial information contained in high spatial resolution remote sensing image data, integrate the spatiotemporal attention mechanism, integrate the change information of the time dimension, and extract comprehensive multi-scale spatiotemporal features from multi-phase high spatial resolution remote sensing image data; A converter is built based on the Transformer architecture, and a spatiotemporal feature integration module is designed. The self-attention mechanism is used to adopt a multi-task learning strategy for change detection and semantic segmentation to refine and strengthen the multi-scale spatial features output by the encoder to identify changes in vegetation cover over time. The decoder is constructed based on multi-task processing to detect changes in vegetation cover and perform semantic segmentation. The self-attention mechanism is embedded to conduct in-depth analysis of spatiotemporal features, optimize the feature fusion process, accurately analyze feature maps based on rich details, and fuse the multi-scale spatiotemporal features output by the encoder and the composite features of vegetation cover changes over time output by the converter to accurately detect vegetation cover changes. The training of the multi-task fully convolutional neural network model includes: A fusion loss function is used to integrate the cross entropy loss of semantic segmentation and the binary cross entropy loss of change detection. An adjustable weight coefficient is introduced to dynamically adjust the weight according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance. Use the Adam optimizer with adaptive learning rate to improve training efficiency and model convergence speed; Integrate dynamic loss weight adjustment and learning rate scheduling strategies to help the model maintain a balance in task learning between training stages and optimize the overall performance of the model; Use early stopping mechanism and model regularization measures to prevent overfitting, enhance the generalization ability of the model, and ensure the stability of the training process; Use multiple indicators, including precision, recall, F1 score, and overall accuracy, to comprehensively evaluate model performance; Based on the comprehensive evaluation results, the network model is fine-tuned.

2. The method for detecting changes in forest vegetation types according to claim 1, characterized in that: The method for detecting changes in forest vegetation types also includes: preprocessing multi-temporal and high spatial resolution remote sensing image data of the forest area.

3. The method for detecting changes in forest vegetation types according to claim 2, characterized in that: The preprocessing of multi-temporal high-spatial-resolution remote sensing image data of forest areas includes: radiation calibration, atmospheric correction and image registration.

4. The method for detecting changes in forest vegetation types according to claim 1, characterized in that: When constructing the sample data set, the image data is enhanced and the sample data set is constructed based on the time window sliding technology.

5. The method for detecting changes in forest vegetation types according to claim 1, characterized in that: The forest vegetation type change detection method also includes the following steps: Based on the detection results of forest vegetation type changes, a spatiotemporal distribution map of forest vegetation type changes was constructed.

6. A forest vegetation type change detection system, characterized in that: include: Data acquisition module, used to obtain multi-temporal high spatial resolution remote sensing image data of forest areas; The model building module is used to build a multi-task full convolutional neural network model based on the convolutional neural network model and multi-task learning according to the multi-temporal high-spatial resolution remote sensing image data of the forest area; The training module is used to construct a sample data set based on multi-temporal and high-spatial-resolution remote sensing image data of forest areas, train a multi-task fully convolutional neural network model, and obtain a forest vegetation type change detection model; The detection module is used to input the multi-temporal high spatial resolution remote sensing image data of the forest area into the forest area vegetation type change detection model to obtain the forest area vegetation type change detection result; The multi-task full convolutional neural network model constructed based on the convolutional neural network model and multi-task learning includes: A hierarchical feature extraction network is constructed as an encoder to deeply analyze the complex spatial information contained in high spatial resolution remote sensing image data, integrate the spatiotemporal attention mechanism, integrate the change information of the time dimension, and extract comprehensive multi-scale spatiotemporal features from multi-phase high spatial resolution remote sensing image data; A converter is built based on the Transformer architecture, and a spatiotemporal feature integration module is designed. The self-attention mechanism is used to adopt a multi-task learning strategy for change detection and semantic segmentation to refine and strengthen the multi-scale spatial features output by the encoder to identify changes in vegetation cover over time. The decoder is constructed based on multi-task processing to detect changes in vegetation cover and perform semantic segmentation. The self-attention mechanism is embedded to conduct in-depth analysis of spatiotemporal features, optimize the feature fusion process, accurately analyze feature maps based on rich details, and fuse the multi-scale spatiotemporal features output by the encoder and the composite features of vegetation cover changes over time output by the converter to accurately detect vegetation cover changes. The training of the multi-task fully convolutional neural network model includes: A fusion loss function is used to integrate the cross entropy loss of semantic segmentation and the binary cross entropy loss of change detection. An adjustable weight coefficient is introduced to dynamically adjust the weight according to the characteristics and complexity of the task, thereby achieving a balance between tasks and optimizing model performance. Use the Adam optimizer with adaptive learning rate to improve training efficiency and model convergence speed; Integrate dynamic loss weight adjustment and learning rate scheduling strategies to help the model maintain a balance in task learning between training stages and optimize the overall performance of the model; Use early stopping mechanism and model regularization measures to prevent overfitting, enhance the generalization ability of the model, and ensure the stability of the training process; Use multiple indicators, including but not limited to precision, recall, F1 score, and overall accuracy, to comprehensively evaluate model performance; Based on the comprehensive evaluation results, the network model is fine-tuned.

7. A computer program product, characterized in that It comprises a computer program, which, when executed, implements the steps of the forest vegetation type change detection method as described in any one of claims 1 to 5.

Images