Forest resource management method and system based on multi-source remote sensing data fusion

Abstract

The invention provides a multi-source remote sensing data fusion forest resource management method and system, and the method comprises the steps: extracting forest dynamic change features from cross-modal multi-source remote sensing historical data, generating a time sequence correlation curve between cross-modal data, and determining a cross-modal time sequence deviation value between different remote sensing modals through the time sequence correlation curve; calculating the characteristic coincidence degree of different remote sensing modes under the normalized spatial scale for the real-time multi-source remote sensing data, and determining the spatial scale adaptation degree of the real-time multi-source remote sensing data according to all the characteristic coincidence degrees; the fusion feature confidence of the real-time multi-source remote sensing data is calculated through the cross-modal time sequence deviation value and the spatial scale adaptation degree; and generating regionalized management parameters of forest resources of the target forest region based on the fusion feature confidence, and performing resource level-to-level management on the target forest region according to the regionalized management parameters. By adopting the scheme, precise forest resource level-to-level management can be realized based on time consistency and space consistency of multi-source remote sensing data fusion.

Description

Technical Field

[0001] This application relates to the field of forest resource management technology, and more specifically, to a method and system for forest resource management based on multi-source remote sensing data fusion. Background Technology

[0002] Forest resource management refers to the planning, organization, coordination, and supervision of the entire life cycle of forest resources—including cultivation, protection, utilization, and monitoring—through a series of scientific and systematic methods and measures, in order to achieve sustainable development of forest resources. Its core objective is to balance ecological, economic, and social benefits while ensuring the ecological functions of forests, ensuring that forest resources meet the needs of the present generation without compromising the ability of future generations to meet their own needs. Effective forest resource management can improve resource utilization efficiency and reduce the risk of ecological damage, making it crucial for achieving sustainable forestry development.

[0003] Forest resource management based on multi-source remote sensing data fusion refers to the process of integrating remote sensing data of different types, platforms, and time phases during forest resource management. Data fusion technology is used to overcome the limitations of single data sources, improving the accuracy, timeliness, and reliability of forest resource information extraction. This results in a more precise, efficient, and dynamic management model for forest resource monitoring, assessment, and decision support. However, current technologies often neglect temporal and spatial resolution differences between different modalities, leading to data inconsistencies or feature deviations within the same region. This results in insufficient handling of temporal and spatial consistency in multi-source remote sensing data fusion, causing unstable assessments of forest change trends and difficulty in unifying management indicators. Consequently, it fails to meet the current needs for refined and dynamic forest management. Therefore, how to achieve precise hierarchical management of forest resources based on the temporal and spatial consistency of multi-source remote sensing data fusion has become a challenge for the industry. Summary of the Invention

[0004] This application provides a forest resource management method and system based on multi-source remote sensing data fusion, which can achieve precise hierarchical management of forest resources based on the temporal and spatial consistency of multi-source remote sensing data fusion.

[0005] In a first aspect, this application provides a forest resource management method based on multi-source remote sensing data fusion, comprising the following steps: Acquire multi-source remote sensing historical data across modalities in the target forest area; Forest dynamic change features are extracted from the multi-source remote sensing historical data. Based on the temporal gradient of the forest dynamic change features, a temporal correlation curve between cross-modal data in the multi-source remote sensing historical data is generated. Then, the temporal correlation curve is combined with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal temporal deviation value between different remote sensing modes. The real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and then the feature overlap degree of different remote sensing modes under the normalized spatial scale is calculated. The spatial scale fit degree of the real-time multi-source remote sensing data is determined based on all feature overlap degrees. The confidence level of the fusion feature of the real-time multi-source remote sensing data is calculated by using the cross-modal temporal deviation value and the spatial scale fit. Based on the confidence level of the fusion features, regionalized management parameters for forest resources in the target forest area are generated, and resource classification management of the target forest area is carried out according to the regionalized management parameters.

[0006] In some embodiments, extracting forest dynamic change features from the multi-source remote sensing historical data specifically includes: Time axis alignment is performed on the different remote sensing modes in the multi-source remote sensing historical data; Spectral features, radar features, and three-dimensional structural features of the target forest area were extracted from multi-source remote sensing historical data aligned with the time axis. The dynamic changes of the forest are determined by the spectral features, the radar features, and the three-dimensional structural features.

[0007] In some embodiments, generating a temporal correlation curve between cross-modal data in the multi-source remote sensing historical data based on the temporal gradient of the forest dynamic change characteristics specifically includes: Determine the temporal gradient of the dynamic change characteristics of the forest; The cross-modal similarity between different remote sensing modes is determined by the temporal gradient. Based on the cross-modal similarity between the different remote sensing modalities, a time-series correlation curve is generated between the cross-modal data in the multi-source remote sensing historical data.

[0008] In some embodiments, determining the cross-modal temporal deviation value between different remote sensing modes by combining the time-series correlation curve with real-time multi-source remote sensing data collected from the target forest area specifically includes: Acquire real-time multi-source remote sensing data of the target forest area; The real-time multi-source remote sensing data is aligned with the time-series correlation curve to obtain the aligned correlation curve. The cross-modal temporal deviation values ​​between different remote sensing modes are calculated using the alignment correlation curve.

[0009] In some embodiments, performing spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculating the feature overlap degree of different remote sensing modalities at the normalized spatial scale, specifically includes: The original spatial resolution information of the real-time multi-source remote sensing data is obtained, and then the target resolution is determined based on the original spatial resolution information. The different remote sensing modes in the real-time multi-source remote sensing data are resampled to the target resolution to obtain different remote sensing modes at the normalized spatial scale. Determine the feature overlap of each remote sensing mode at the normalized spatial scale.

[0010] In some embodiments, determining the spatial scale fit of the real-time multi-source remote sensing data based on the overlap of all features specifically includes: The overlap ratio of resource management in the target forest area is determined by the overlap of all features; The spatial scale contribution of each remote sensing mode is determined based on the importance of different remote sensing modes in forest resource monitoring in the real-time multi-source remote sensing data. The spatial scale fit of the real-time multi-source remote sensing data is determined by the overlap ratio and the contribution of all spatial scales.

[0011] In some embodiments, calculating the fusion feature confidence level of the real-time multi-source remote sensing data using the cross-modal temporal deviation value and the spatial scale fit specifically includes: Determine the temporal and spatial fusion weight coefficients during the fusion of multi-source remote sensing data in the target forest area; The confidence level of the fusion features of the real-time multi-source remote sensing data is determined based on the fusion weight coefficient, the cross-modal temporal deviation value, and the spatial scale fit.

[0012] In some embodiments, generating regionalized management parameters for forest resources in a target forest area based on the fusion feature confidence level specifically includes: Divide the target forest area into multiple management units; Obtain forest metrics for each management unit; The forest indicators for each management unit are weighted and adjusted based on the confidence level of the fusion features. Regionalized management parameters for forest resources in the target forest area are generated based on the revised forest indicators.

[0013] In some embodiments, resource-level management of a target forest area based on the regionalized management parameters specifically includes: Set multiple threshold levels for forest resource management in target forest areas; Based on the regionalized management parameters and the set level thresholds, different management units in the target forest area are classified into levels to obtain a management intensity grading map of the target forest area; Based on the aforementioned management intensity grading map, resource management measures for the target forest area are formulated, which are then used to guide the graded management of resources in the target forest area.

[0014] Secondly, this application provides a forest resource management system based on multi-source remote sensing data fusion, comprising: The acquisition module is used to acquire multi-source remote sensing historical data across modalities in the target forest area; The processing module is used to extract forest dynamic change features from the multi-source remote sensing historical data, generate a time-series correlation curve between cross-modal data in the multi-source remote sensing historical data based on the time-series gradient of the forest dynamic change features, and then combine the time-series correlation curve with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal time-series deviation value between different remote sensing modes. The processing module is also used to perform spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculate the feature overlap degree of different remote sensing modes under the normalized spatial scale, and determine the spatial scale fit degree of the real-time multi-source remote sensing data based on all feature overlap degrees. The processing module is also used to calculate the fusion feature confidence of the real-time multi-source remote sensing data through the cross-modal temporal deviation value and the spatial scale fit. The execution module is used to generate regionalized management parameters for forest resources in the target forest area based on the confidence level of the fusion features, and to perform resource hierarchical management of the target forest area according to the regionalized management parameters.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In this application, multi-source remote sensing historical data across modalities is acquired in a target forest area; forest dynamic change features are extracted from the multi-source remote sensing historical data; a temporal correlation curve between the cross-modal data in the multi-source remote sensing historical data is generated based on the temporal gradient of the forest dynamic change features; the temporal correlation curve is then combined with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal temporal deviation value between different remote sensing modalities; the real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and the feature overlap degree of different remote sensing modalities at the normalized spatial scale is calculated; the spatial scale fit of the real-time multi-source remote sensing data is determined based on all feature overlap degrees; the fusion feature confidence degree of the real-time multi-source remote sensing data is calculated through the cross-modal temporal deviation value and the spatial scale fit degree; regionalized management parameters of forest resources in the target forest area are generated based on the fusion feature confidence degree, and resource hierarchical management of the target forest area is carried out based on the regionalized management parameters.

[0016] Therefore, in this application, firstly, by combining the time-series correlation curve with real-time multi-source remote sensing data collected from the target forest area to determine the cross-modal time-series deviation value between different remote sensing modalities, the temporal deviation of different remote sensing modal data can be effectively quantified, so that the multi-source remote sensing fusion data can be uniformly corrected in the time dimension, thereby accurately depicting the changing trend of forest resources in the target forest area; secondly, by determining the spatial scale fit of the real-time multi-source remote sensing data based on the overlap of all features, the problem of spatial information inconsistency caused by resolution differences in multi-source remote sensing data can be solved, making the spatial characteristics of forest resources more accurate and coordinated, ensuring spatial consistency, avoiding the spread of local errors, and ensuring the stability of forest resource management indicators at different spatial levels, laying a spatial scale foundation for precise forest management; then, the fusion feature confidence of the real-time multi-source remote sensing data is calculated through the cross-modal time-series deviation value and the spatial scale fit, and the fusion... Feature confidence simultaneously quantifies the degree of coordination between time and space, enhancing the robustness of the fusion process and dynamically reflecting the reliability of fused data across different forest regions and time periods. This avoids forest resource management deviations caused by imbalances in a single indicator. Finally, based on the fused feature confidence, regionalized management parameters for forest resources in the target forest region are generated. These regionalized management parameters, as a product of temporal and spatial consistency fusion, comprehensively reflect the intensity of forest resource changes and management needs. Furthermore, resource-level management of the target forest region can be carried out based on these regionalized management parameters, enabling a shift from extensive to intensive forest management. This provides layered and hierarchical dynamic control capabilities, meeting the differentiated and precise needs of actual forest resource management and promoting the modernization and intelligence of forest resource management. In summary, this scheme can achieve precise hierarchical management of forest resources based on the temporal and spatial consistency of multi-source remote sensing data fusion. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an exemplary flowchart of a forest resource management method based on multi-source remote sensing data fusion, as shown in some embodiments of this application. Figure 2 This is an exemplary flowchart illustrating the extraction of dynamic change features of forests according to some embodiments of this application; Figure 3 This is an exemplary flowchart illustrating the determination of cross-modal timing deviation values ​​according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a forest resource management system based on multi-source remote sensing data fusion according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device for implementing a forest resource management method that integrates multi-source remote sensing data, according to some embodiments of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] refer to Figure 1 The figure is an exemplary flowchart of a forest resource management method based on multi-source remote sensing data fusion, according to some embodiments of this application. This method mainly includes the following steps: In step 101, multi-source remote sensing historical data across modalities are acquired in the target forest area.

[0021] In specific implementation, acquiring multi-modal, multi-source remote sensing historical data of the target forest area can be achieved in the following ways: Multi-modal, multi-source remote sensing historical data of the target forest area can be obtained from official channels (such as the National Geographic Information Public Service Platform, the European Space Agency, etc.), including: optical remote sensing images, synthetic aperture radar (SAR) images, and lidar (Light Detection and Ranging) point cloud data of the target forest area, along with their corresponding acquisition times. Optical remote sensing images utilize multispectral sensors to acquire reflectance information of different bands of the land surface, reflecting the spectral characteristics of vegetation; SAR radar images acquire land surface structure and water content information by actively emitting microwaves and receiving echo signals using synthetic aperture radar, unaffected by cloud and rain interference; LiDAR point cloud data acquires the three-dimensional structure and canopy height information of the forest through laser scanning. Furthermore, the multi-source remote sensing historical data covers at least three complete forest growth cycles in the target forest area. Other methods can also be used in other embodiments, which are not limited here.

[0022] It should be noted that the multi-modal, multi-source remote sensing historical data in this application refers to a collection of historical data of spatial observation data of the target forest area acquired from different sensors at different times. It is used to analyze the long-term dynamic change trend of the target forest area and includes modal remote sensing data such as optical remote sensing data, SAR radar data and LiDAR point cloud data in the target forest area.

[0023] In step 102, forest dynamic change features are extracted from the multi-source remote sensing historical data. Based on the temporal gradient of the forest dynamic change features, a temporal correlation curve between cross-modal data in the multi-source remote sensing historical data is generated. Then, the temporal correlation curve is combined with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal temporal deviation value between different remote sensing modes.

[0024] In some embodiments, reference Figure 2 As shown, this figure is an exemplary flowchart for extracting forest dynamic change features in some embodiments of this application. In this embodiment, the extraction of forest dynamic change features from the multi-source remote sensing historical data can be achieved by the following steps: First, in step 1021, the different remote sensing modes in the multi-source remote sensing historical data are aligned along the time axis; Secondly, in step 1022, the spectral features, radar features, and three-dimensional structural features of the target forest area are extracted based on the multi-source remote sensing historical data aligned with the time axis. Finally, in step 1023, the dynamic change characteristics of the forest are determined by the spectral features, the radar features, and the three-dimensional structural features.

[0025] In specific implementation, time axis alignment of different remote sensing modalities in the multi-source remote sensing historical data can be achieved in the following way: Time axis alignment is performed on different remote sensing modalities (where different remote sensing modalities refer to remote sensing data types with different imaging principles, such as optical, microwave, and lidar) in the multi-source remote sensing historical data. For example, the data acquisition time of each modality's remote sensing data can be organized, and linear interpolation or nearest neighbor methods can be used to map asynchronously acquired data to a unified time reference (such as monthly or quarterly time nodes) to ensure that the data of different modalities are aligned. Based on the temporal comparability, the extraction of spectral features, radar features, and three-dimensional structural features of the target forest area from time-aligned multi-source remote sensing historical data can be achieved in the following manner: In the time-aligned multi-source remote sensing historical data, vegetation index calculation methods (such as normalized difference vegetation index and red edge position index) are used for optical remote sensing data to extract spectral features reflecting forest leaf area index and photosynthetic activity, thereby reflecting vegetation photosynthetic activity, chlorophyll content, and canopy status; polarization decomposition and backscattering coefficient calculation methods are used for SAR radar data. The method obtains radar features related to changes in forest structure density and water content to reflect the structural density, moisture status, and roughness of forest vegetation. Canopy height models are constructed and vertical profiles are analyzed using LiDAR point cloud data to extract three-dimensional structural features such as the rate of change of forest canopy height and canopy structural complexity, reflecting the vertical structural morphology and canopy distribution of the forest. This provides descriptive information of the target forest area in three dimensions: spectral, microwave, and structural, offering multi-faceted support for comprehensive analysis of forest dynamics. The determination of forest dynamic change characteristics through the spectral features, radar features, and three-dimensional structural features can be achieved by using time series analysis methods (such as Theil-Sen trend analysis and Mann-Kendall mutation test) to calculate the slope and abrupt change time points of the modal features (spectral features, radar features, and three-dimensional structural features) within the observation period of the target forest area. The set of results corresponding to each modal feature is then used as the forest dynamic change characteristics. Other methods can also be used in other embodiments, which are not limited here.

[0026] It should be noted that, in this application, time axis alignment refers to adjusting multi-source remote sensing data acquired at different times to the same time reference frame to eliminate the impact of time differences on cross-modal feature analysis; spectral features in this application refer to the reflection / radiation characteristics of ground objects to electromagnetic waves of different wavelengths, which reflect the vegetation growth status of the target forest area; radar features in this application refer to the reflection / scattering characteristics of ground objects to microwave signals, which are affected by the vegetation structure and humidity in the target forest area; three-dimensional structural features in this application refer to the spatial distribution characteristics of forests in the vertical and horizontal directions, which are used to describe the vertical structural morphology and canopy distribution of the target forest area; forest dynamic change features in this application refer to a multi-dimensional feature set that can quantify the state changes of different modal features (spectral, radar, three-dimensional structural features, etc.) of the target forest area in the time dimension, which provides a change signal with clear physical and ecological significance for subsequent time-series correlation analysis of cross-modal data, so as to achieve quantifiable comparison of forest resource change trends.

[0027] In some embodiments, generating temporal correlation curves between cross-modal data in the multi-source remote sensing historical data based on the temporal gradient of the forest dynamic change characteristics can be achieved using the following steps: Determine the temporal gradient of the dynamic change characteristics of the forest; The cross-modal similarity between different remote sensing modes is determined by the temporal gradient. Based on the cross-modal similarity between the different remote sensing modalities, a time-series correlation curve is generated between the cross-modal data in the multi-source remote sensing historical data.

[0028] In specific implementation, determining the temporal gradient of the forest dynamic change characteristics can be achieved in the following way: the rate of change of the forest dynamic change characteristics (including spectral characteristics, radar characteristics, and three-dimensional structural characteristics) can be calculated on a time series, that is, the difference of the forest dynamic change characteristics at adjacent time nodes can be obtained, or the gradient vector can be calculated based on the central difference formula, and the obtained gradient vector can be used as the temporal gradient of the forest dynamic change characteristics. In addition, in order to reduce the interference of high-frequency noise, the forest dynamic change characteristics can be smoothed first using Savitzky-Golay smoothing filtering or moving average method. Quantifying the rate and direction of change of forest features over time in the target forest area establishes a comparable numerical basis for subsequent cross-modal comparisons. Determining the cross-modal similarity between different remote sensing modalities through the temporal gradient can be achieved by normalizing the gradient values ​​of each modal feature in the temporal gradient of the forest dynamics to the same numerical range. This eliminates the influence of differences in feature dimensions and value ranges, allowing direct comparison of the changing trends of different modal remote sensing data within a unified numerical space, thus improving cross-modal similarity. To improve the accuracy of the correlation, multivariate dynamic time warping algorithms or canonical correlation analysis can be used to model the multimodal correlation of the temporal gradients of different remote sensing modes in the normalized multi-source remote sensing historical data. This allows the distribution of temporal variation similarity among different remote sensing modes to be used as the cross-modal similarity, thereby uncovering the commonalities and differences in the temporal variation patterns of different remote sensing modes. Generating temporal correlation curves between cross-modal data in the multi-source remote sensing historical data based on the cross-modal similarity can be achieved in the following way: the cross-modal correlation curves between different remote sensing modes can be used as a basis for further analysis. The similarity over time is plotted as a similarity curve. This similarity curve can also be fitted using B-spline interpolation or Gaussian regression to reduce the influence of local outliers and make the curve smoother and more continuous. The vertical axis of the similarity curve represents the correlation strength between different remote sensing modes, and the horizontal axis represents time. Finally, this similarity curve is used as the temporal correlation curve between cross-modal data in the multi-source remote sensing historical data. The purpose of this step is to visualize and quantify the complex cross-modal variation relationship so as to calculate the cross-modal deviation value in subsequent calculations. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0029] It should be noted that the temporal gradient in this application refers to the rate of change of forest dynamics characteristics between adjacent observation points over time, which is used to reveal the speed and direction of forest change; the cross-modal similarity in this application represents the degree of similarity in the change patterns of different remote sensing modes in multi-source remote sensing data; the cross-modal data in this application refers to data modes from different remote sensing platforms or sensors; the temporal correlation curve in this application refers to a curve describing the strength of correlation between different remote sensing modes in the temporal change patterns, which is used to reveal the correlation and difference patterns of change between different remote sensing modes at the visualization and quantitative levels.

[0030] In some embodiments, reference Figure 3 As shown in the figure, this is an exemplary flowchart for determining cross-modal time series deviation values ​​in some embodiments of this application. The determination of cross-modal time series deviation values ​​between different remote sensing modes by combining the time series correlation curve with real-time multi-source remote sensing data collected from the target forest area can be achieved by the following steps: Acquire real-time multi-source remote sensing data of the target forest area; The real-time multi-source remote sensing data is aligned with the time-series correlation curve to obtain the aligned correlation curve. The cross-modal temporal deviation values ​​between different remote sensing modes are calculated using the alignment correlation curve.

[0031] In specific implementation, acquiring real-time multi-source remote sensing data of the target forest area can be achieved in the following way: Real-time multi-source remote sensing data, such as optical remote sensing data, SAR radar data, and LiDAR point cloud data, are simultaneously acquired through satellite receiving stations or UAV platforms. Preprocessing, including radiometric calibration and geometric correction, is then performed to ensure consistency in radiometric and spatial characteristics between the acquired real-time multi-source remote sensing data and historical multi-source remote sensing data. The real-time multi-source remote sensing data is then aligned with the time-series correlation curve to obtain the aligned correlation curve. This can be achieved in the following way: First, real-time spectral features, radar features, and three-dimensional structural features can be obtained using the same feature extraction method as for historical multi-source remote sensing data. Then, the cross-modal similarity between different remote sensing modes in the real-time multi-source remote sensing data is calculated. The obtained cross-modal similarity is then mapped and registered to the corresponding time node of the time-series correlation curve according to the corresponding acquisition time. The most recent time interpolation method or spline interpolation method can be used to map the real-time features to the same time scale as the time-series correlation curve. In addition, if the acquisition time corresponding to the real-time multi-source remote sensing data falls between two historical sampling points, the correlation strength between different remote sensing modes corresponding to that historical time point is calculated by linear interpolation to ensure that the real-time features and historical trends are aligned in time, and the aligned curve is used as the alignment correlation curve. The cross-modal temporal deviation value between different remote sensing modes can be calculated through the alignment correlation curve in the following way: at each time point of the alignment correlation curve, the absolute difference between the correlation strength between different remote sensing modes corresponding to the real-time multi-source remote sensing data and the correlation strength between different remote sensing modes on the temporal correlation curve corresponding to the multi-source remote sensing historical data is calculated. This absolute difference reflects the degree of deviation between the real-time forest state and the historical cross-modal trend. Then, all the obtained absolute differences are accumulated and averaged at all time points, and the obtained value is normalized (i.e., mapped to the 0-1 interval) and used as the cross-modal temporal deviation value between different remote sensing modes. Other methods can also be used to determine this in other embodiments, which are not limited here.

[0032] It should be noted that the real-time multi-source remote sensing data in this application refers to multiple remote sensing modal data of the target forest area collected at the current moment, which is used to reflect the immediate state of the target forest area; the feature temporal alignment in this application refers to the process of mapping the real-time features of the target forest area to the same time scale as the historical features, which is used to ensure the temporal consistency between historical and real-time multi-source remote sensing data; the alignment correlation curve in this application represents the curve obtained after aligning the real-time cross-modal trend to the same time scale as the historical cross-modal trend, which is used to reflect the degree of coordinated change of different remote sensing modalities of the target forest area at the current moment; the cross-modal temporal deviation value in this application refers to the difference between the historical cross-modal trend and the real-time cross-modal trend of the target forest area, which is used to quantify the degree to which the current state of the target forest area deviates from the historical change pattern. In addition, the cross-modal temporal deviation value can measure the difference in feature changes of different remote sensing data types (optical, SAR, LiDAR, etc.) on the same time series, thereby reflecting the consistency or difference in their responses when monitoring the target forest area.

[0033] In step 103, the real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and then the feature overlap degree of different remote sensing modes under the normalized spatial scale is calculated. The spatial scale fit degree of the real-time multi-source remote sensing data is determined based on all feature overlap degrees.

[0034] In some embodiments, the spatial scale normalization processing of the real-time multi-source remote sensing data, and the subsequent calculation of the feature overlap degree of different remote sensing modes at the normalized spatial scale, can be achieved by the following steps: The original spatial resolution information of the real-time multi-source remote sensing data is obtained, and then the target resolution is determined based on the original spatial resolution information. The different remote sensing modes in the real-time multi-source remote sensing data are resampled to the target resolution to obtain different remote sensing modes at the normalized spatial scale. Determine the feature overlap of each remote sensing mode at the normalized spatial scale.

[0035] In specific implementation, obtaining the original spatial resolution information of the real-time multi-source remote sensing data and then determining the target resolution based on the original spatial resolution information can be achieved in the following way: First, read the metadata of each data file in the real-time multi-source remote sensing data to obtain the original spatial resolution information. Then, combine the original spatial resolution information with the forest resource management unit scale (such as the minimum management patch side length of the small plot) and the anti-aliasing principle to select a target resolution that balances information effectiveness and operability. The selection strategy is, for example, to take a safe downsampling scale that is not lower than the highest resolution of each remote sensing modality, or to set it according to the integer factor of the management unit side length so that the grid and the small plot boundary are as divisible as possible. The target resolution that is unified, physically reasonable and matches the resource management unit is used as the common scale for cross-modal fusion, so as to suppress noise / aliasing caused by excessive detail and avoid the loss of key management features by coarse scale. Resampling different remote sensing modalities in the real-time multi-source remote sensing data to the target resolution to obtain different remote sensing modalities at the normalized spatial scale can be achieved in the following way: The different remote sensing modalities in the real-time multi-source remote sensing data can be resampled to the target resolution using methods such as wavelet transform, multi-resolution pyramid, or bilinear / cubic convolution interpolation. For LiDAR point cloud data, grid data is first generated through gridded statistics (such as maximum height and average height), and then resampled according to the target resolution to obtain different remote sensing modalities at a normalized spatial scale, thereby completing the spatial scale normalization processing of the real-time multi-source remote sensing data. The feature overlap of each remote sensing modality at the normalized spatial scale can be determined as follows: normalized multimodal features (such as spectrum, SAR backscatter, and LiDAR height) can be extracted at the normalized spatial scale, and the matching ratio of each remote sensing modality in the feature space can be calculated by pixel-by-pixel comparison or regional statistics based on vector boundaries. The obtained matching ratio is then used as the feature overlap of each remote sensing modality at the normalized spatial scale to quantify the degree of information consistency of different remote sensing modalities at the same spatial scale. Other methods can also be used in other embodiments, which are not limited here.

[0036] It should be noted that the original spatial resolution information in this application refers to the set of ground sampling accuracy parameters naturally formed by different remote sensing modalities during the data acquisition stage; the multi-scale spatial normalization processing in this application refers to the process of converting multi-modal data with different resolutions in real-time multi-source remote sensing data into the same spatial scale, which is used to eliminate the information mismatch problem caused by differences in spatial accuracy; the target resolution in this application refers to the uniform pixel size selected for anti-aliasing constraints of different remote sensing modalities, which is the common scale for cross-modal resampling and comparison; the normalized spatial scale in this application refers to the spatial expression of each remote sensing modality under the same pixel size and the same raster template after geometric uniformity and resampling, which is used to eliminate the spatial heterogeneity bias caused by differences in resolution and sampling grid; the feature overlap in this application represents the degree of information consistency of different remote sensing modalities in real-time multi-source remote sensing data at the same spatial scale, which is used to reflect the degree of consistency of the expression of the same phenomenon (such as canopy closure, tree height, humidity) of different remote sensing modalities in the same management unit at the normalized spatial scale, so as to understand the consistency and reliability of cross-modal information in the target forest.

[0037] In some embodiments, determining the spatial scale fit of the real-time multi-source remote sensing data based on the overlap of all features can be achieved using the following steps: The overlap ratio of resource management in the target forest area is determined by the overlap of all features; The spatial scale contribution of each remote sensing mode is determined based on the importance of different remote sensing modes in forest resource monitoring in the real-time multi-source remote sensing data. The spatial scale fit of the real-time multi-source remote sensing data is determined by the overlap ratio and the contribution of all spatial scales.

[0038] In specific implementation, determining the overlap ratio of resource management in the target forest area based on the overlap of all features can be achieved as follows: the average value of all feature overlap can be used as the overlap ratio of resource management in the target forest area. Determining the spatial scale contribution of each remote sensing mode based on its importance in forest resource monitoring from the real-time multi-source remote sensing data can be achieved as follows: weights can be assigned to each remote sensing mode according to its importance in forest resource monitoring from the real-time multi-source remote sensing data. This assignment process can employ expert scoring or analytic hierarchy process (AHP). For example, LiDAR is more sensitive to forest stand structure, and its weights can be set accordingly. The weight assigned to each remote sensing modality is 0.4, and the corresponding weight is used as the spatial scale contribution of each remote sensing modality. The spatial scale fit of the real-time multi-source remote sensing data can be determined by the overlap ratio and all spatial scale contributions. This can be achieved by fusing the overlap ratio with the spatial scale contribution of each remote sensing modality using a linear weighting method, and normalizing the fusion result to make its value between 0 and 1. The final result is then used as the spatial scale fit of the real-time multi-source remote sensing data. Other methods can also be used in other embodiments, which are not limited here.

[0039] It should be noted that the overlap ratio in this application represents the degree of information consistency of different remote sensing modalities in the resource management of the target forest area; the spatial scale contribution in this application represents the comprehensive contribution of remote sensing modalities to the forest resource monitoring target under the condition of spatial scale normalization; and the spatial scale fit in this application represents the degree of adaptation of real-time multi-source remote sensing data of the target forest area to information fusion under a unified spatial scale.

[0040] In step 104, the confidence level of the fusion features of the real-time multi-source remote sensing data is calculated using the cross-modal temporal deviation value and the spatial scale fit.

[0041] In some embodiments, the confidence level of the fusion feature of the real-time multi-source remote sensing data can be calculated using the cross-modal temporal deviation value and the spatial scale fit as follows: Determine the temporal and spatial fusion weight coefficients during the fusion of multi-source remote sensing data in the target forest area; The confidence level of the fusion features of the real-time multi-source remote sensing data is determined based on the fusion weight coefficient, the cross-modal temporal deviation value, and the spatial scale fit.

[0042] In practice, determining the temporal and spatial fusion weight coefficients during the fusion of multi-source remote sensing data in the target forest area can be achieved in the following ways: First, by analyzing historical remote sensing data of the target forest area, the influence of the time dimension on the fusion results during different seasons and phenological periods can be statistically analyzed (e.g., vegetation cover changes rapidly during the rainy season, requiring a higher time weight). The basic weight of the time dimension can then be calculated using the coefficient of variation method. Second, based on forest type zoning (e.g., coniferous forest, broadleaf forest), spatial autocorrelation analysis can be used to quantify the spatial characteristic differences between different areas of the target forest area, assigning differentiated weights to the spatial dimension. Finally, the ratio of time and spatial weights during the multi-source remote sensing data fusion process can be adjusted through cross-validation (e.g., dividing historical data into training and validation sets) so that the sum of the two weights is 1 (e.g., time weight is 0.3 and spatial weight is 0.7). This ensures that the fusion weight coefficient reflects both the impact of temporal dynamics on data consistency and the effect of spatial heterogeneity on fusion accuracy. The final adjusted result is used as the temporal and spatial fusion weight coefficient in the multi-source remote sensing data fusion process in the target forest area, wherein the fusion weight coefficient consists of temporal weight and spatial weight. The confidence level of the fusion feature of the real-time multi-source remote sensing data based on the fusion weight coefficient, the cross-modal temporal deviation value, and the spatial scale adaptation can be determined in the following way: the confidence level of the fusion feature of the real-time multi-source remote sensing data can be calculated by constructing a fusion feature confidence level formula, for example: fusion feature confidence level = (1 - cross-modal temporal deviation value) × temporal weight + spatial scale adaptation × spatial weight, so as to comprehensively consider the two core dimensions of temporal consistency and spatial adaptability, quantify the reliability of the fusion feature of multi-source remote sensing data, and provide a reliable feature quality assessment basis for subsequent forest resource monitoring applications in the target forest area (such as biomass estimation and pest detection); other methods can also be used in other embodiments, which are not limited here.

[0043] It should be noted that the fusion weight coefficient in this application represents the relative weight of the temporal consistency term and the spatial consistency term in the confidence fusion process during the fusion of multi-source remote sensing data in the target forest area; the fusion feature confidence in this application represents the reliability of multi-source data in terms of temporal and spatial consistency, which is a quality indicator for measuring the usability of the multi-source remote sensing data fusion results in the target forest area for forest resource monitoring applications.

[0044] In step 105, regionalized management parameters for forest resources in the target forest area are generated based on the confidence level of the fusion features, and resource classification management is carried out in the target forest area according to the regionalized management parameters.

[0045] In some embodiments, generating regionalized management parameters for forest resources in a target forest area based on the confidence level of the fused features can be achieved through the following steps: Divide the target forest area into multiple management units; Obtain forest metrics for each management unit; The forest indicators for each management unit are weighted and adjusted based on the confidence level of the fusion features. Regionalized management parameters for forest resources in the target forest area are generated based on the revised forest indicators.

[0046] In specific implementation, dividing the target forest area into multiple management units can be achieved in the following ways: The smallest management unit for forest resources can be divided within the target forest area, for example, using existing stand sub-compartments or regular grids generated based on digital elevation models as the basic unit, thus dividing the target forest area into multiple management units. Obtaining forest indicators for each management unit can be achieved in the following way: Forest indicators for each management unit can be estimated from real-time multi-source remote sensing data of the target forest area. These forest indicators include, for example, stand density, average tree height, canopy closure, and stock volume. Weighting and correcting the forest indicators of each management unit based on the confidence level of the fusion features can be achieved in the following way: Within each management unit, the forest indicators (such as stand density, average tree height, canopy closure, and stock volume) can be weighted based on the confidence level of the fusion features. This weighting process, for example, involves directly relying on the multi-source remote sensing fusion results for weighted correction of the remote sensing data of high-confidence management units, while combining historical data with the remote sensing data of low-confidence management units. Historical experience or sample survey data are weighted and corrected to leverage the advantages of high-confidence areas while avoiding data distortion in low-confidence areas. The generation of regionalized management parameters for forest resources in the target forest area based on the corrected forest indicators can be achieved in the following way: First, a multi-indicator comprehensive evaluation method, such as the analytic hierarchy process (AHP), entropy method, or expert weighting method, can be used to synthesize the corrected forest indicators for each management unit, forming a management score reflecting the forest structure, utilization pressure, and resource status within each management unit. Then, a range standardization method or a percentage conversion method can be used to map the management score of each management unit to a fixed range (e.g., 0-100). This management score reflects the comprehensive level of each management unit in terms of utilization intensity, space occupation, and structural stability. Finally, the set of all management units and their corresponding management scores is used as the regionalized management parameters for forest resources in the target forest area to understand the required management intervention intensity for different management units in the target forest area. Other methods can also be used in other embodiments, which are not limited here.

[0047] It should be noted that, in this application, the management unit refers to the smallest spatial division unit in the forest resource management of the target forest area, which serves as the spatial carrier for remote sensing data processing and management decision-making; the forest index in this application refers to quantitative parameters that can reflect the forest resource status and the impact of management activities within the management unit; the weighted correction in this application refers to the process of adjusting the weight allocation of the original forest index of the management unit by fusing feature confidence, which is used to filter out interference from low-quality data, improve the accuracy of the forest index, and provide a more reliable basis for the subsequent generation of regionalized management parameters; the regionalized management parameters in this application represent the intensity of management intervention required for different management units within the target forest area, which can intuitively present the intensity differences of forest resource management in different management units within the target forest area, providing a decision-making basis for formulating scientific forest resource management strategies and optimizing resource allocation.

[0048] In some embodiments, resource-level management of a target forest area based on the regionalized management parameters can be achieved through the following steps: Set multiple threshold levels for forest resource management in target forest areas; Based on the regionalized management parameters and the set level thresholds, different management units in the target forest area are classified into levels to obtain a management intensity grading map of the target forest area; Based on the aforementioned management intensity grading map, resource management measures for the target forest area are formulated, which are then used to guide the graded management of resources in the target forest area.

[0049] In practical implementation, setting multiple threshold levels for forest resource management in the target forest area can be achieved in the following way: Based on the value range of regionalized management parameters, multiple threshold levels for forest resource management in the target forest area can be set, for example, dividing it into three management levels: low, medium, and high. The threshold levels can be determined by combining national forest management standards, expert experience, and historical survey data. For example, when the value range of the regionalized management parameters is 0~100, the threshold levels can be set as follows: 0~30 defined as low intensity, 31~70 defined as medium intensity, and 71~100 defined as high intensity. Dividing different management units in the target forest area into levels based on the regionalized management parameters and the set threshold levels to obtain a management intensity grading map of the target forest area can be achieved in the following way: Different management units in the target forest area are classified according to regionalized management parameters... The management intensity grading map is generated by classifying management parameters and set level thresholds to cover the target forest area. Based on this map, resource management measures are formulated to guide the graded management of resources in the target forest area. This can be achieved by: developing differentiated resource management measures for different management levels; for example, low-level units only require routine patrols and nature conservation; medium-level units require appropriate thinning, harvesting, and replanting; and high-level units require key interventions, such as clear-cutting, integrated pest management, or soil and water conservation projects. The management intensity grading map and resource management measures are then stored in a forest resource management database for spatial display and traceable recording to guide the graded management of resources in the target forest area. Other methods can also be used in other embodiments, which are not limited here.

[0050] It should be noted that the grade threshold in this application refers to the numerical boundary for dividing different management intensity levels in the target forest area, which provides a quantitative boundary for the hierarchical management of forest units; the management intensity grading map in this application refers to a spatial distribution map of management intensity levels in the target forest area; the resource grading management in this application refers to a management method that divides different management units in the target forest area into grades based on regionalized management parameters and matches differentiated management measures to different grades. It is used to transform the complex state of forest resources into an operable grading management scheme to achieve scientific, precise and dynamic forest resource management.

[0051] Furthermore, in another aspect of this application, in some embodiments, this application provides a forest resource management system based on multi-source remote sensing data fusion, with reference to... Figure 4 The figure is a schematic diagram of the structure of a forest resource management system based on multi-source remote sensing data fusion according to some embodiments of this application. The forest resource management system 400 based on multi-source remote sensing data fusion includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described below: The acquisition module 401 in this application is mainly used to acquire multi-source remote sensing historical data across modalities in the target forest area; Processing module 402 in this application is mainly used to extract forest dynamic change features from the multi-source remote sensing historical data, generate a time-series correlation curve between cross-modal data in the multi-source remote sensing historical data based on the time-series gradient of the forest dynamic change features, and then combine the time-series correlation curve with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal time-series deviation value between different remote sensing modes. The processing module 402 described in this application is further used to perform spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculate the feature overlap degree of different remote sensing modes under the normalized spatial scale, and determine the spatial scale fit degree of the real-time multi-source remote sensing data based on all feature overlap degrees. The processing module 402 described in this application is also used to calculate the fusion feature confidence of the real-time multi-source remote sensing data through the cross-modal temporal deviation value and the spatial scale fit; The execution module 403 in this application is mainly used to generate regionalized management parameters of forest resources in the target forest area based on the confidence level of the fusion features, and to perform resource classification management of the target forest area according to the regionalized management parameters.

[0052] The foregoing has detailed examples of a forest resource management method and system based on multi-source remote sensing data fusion provided in the embodiments of this application. It is understood that the corresponding apparatus, in order to achieve the above functions, includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0053] In some embodiments, this application also provides a computer device, the computer device including a memory and a processor, the memory for storing a computer program, and the processor for calling and running the computer program from the memory, so that the computer device performs the above-described forest resource management method of multi-source remote sensing data fusion.

[0054] In some embodiments, reference Figure 5The dashed lines in the figure indicate that the unit or module is optional. This figure is a schematic diagram of the computer device used to implement the forest resource management method based on multi-source remote sensing data fusion of this application. The forest resource management method based on multi-source remote sensing data fusion in the above embodiments can be implemented through... Figure 5 The computer device 500 shown is used to implement this, and the computer device 500 includes at least one processor 501, a memory 502 and at least one communication unit 505. The computer device 500 may be a terminal device, a server or a chip.

[0055] The processor 501 can be a general-purpose processor or a special-purpose processor. For example, the processor 501 can be a central processing unit (CPU). The CPU can be used to control the computer device 500, execute software programs, and process data from the software programs. The computer device 500 may also include a communication unit 505 for inputting (receiving) and outputting (transmitting) signals.

[0056] For example, computer device 500 may be a chip, communication unit 505 may be the input and / or output circuit of the chip, or communication unit 505 may be the communication interface of the chip, and the chip may be a component of terminal device, network device or other device.

[0057] For example, computer device 500 may be a terminal device or a server, and communication unit 505 may be a transceiver of the terminal device or the server, or communication unit 505 may be a transceiver circuit of the terminal device or the server.

[0058] The computer device 500 may include one or more memories 502 storing a program 504. The program 504 can be executed by a processor 501 to generate instructions 503, causing the processor 501 to perform the methods described in the above method embodiments according to the instructions 503. Optionally, the memory 502 may also store data (such as a target audit model). Optionally, the processor 501 may also read data stored in the memory 502, which may be stored at the same storage address as the program 504, or the data may be stored at a different storage address than the program 504.

[0059] The processor 501 and memory 502 can be configured separately or integrated together, for example, integrated on the system-on-chip (SOC) of the terminal device.

[0060] It should be understood that each step of the above method embodiment can be completed by hardware logic circuits or software instructions in the processor 501. The processor 501 can be a CPU, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, such as discrete gate, transistor logic devices, or discrete hardware components.

[0061] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0062] For example, in some embodiments, this application also provides a computer-readable storage medium storing instructions or code that, when executed on a computer, cause the computer to implement the above-described forest resource management method based on multi-source remote sensing data fusion.

[0063] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0064] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A forest resource management method based on multi-source remote sensing data fusion, characterized in that, Includes the following steps: Acquire multi-source remote sensing historical data across modalities in the target forest area; Forest dynamic change features are extracted from the multi-source remote sensing historical data. Based on the temporal gradient of the forest dynamic change features, a temporal correlation curve between cross-modal data in the multi-source remote sensing historical data is generated. Then, the temporal correlation curve is combined with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal temporal deviation value between different remote sensing modes. The real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and then the feature overlap degree of different remote sensing modes under the normalized spatial scale is calculated. The spatial scale fit degree of the real-time multi-source remote sensing data is determined based on all feature overlap degrees. The confidence level of the fusion feature of the real-time multi-source remote sensing data is calculated by using the cross-modal temporal deviation value and the spatial scale fit. Based on the confidence level of the fusion features, regionalized management parameters for forest resources in the target forest area are generated, and resource classification management of the target forest area is carried out according to the regionalized management parameters.

2. The method as described in claim 1, characterized in that, Extracting forest dynamic change features from the multi-source remote sensing historical data specifically includes: Time axis alignment is performed on the different remote sensing modes in the multi-source remote sensing historical data; Spectral features, radar features, and three-dimensional structural features of the target forest area were extracted from multi-source remote sensing historical data aligned with the time axis. The dynamic changes of the forest are determined by the spectral features, the radar features, and the three-dimensional structural features.

3. The method as described in claim 1, characterized in that, Generating time-series correlation curves between cross-modal data in the multi-source remote sensing historical data based on the time-series gradient of the forest dynamics characteristics specifically includes: Determine the temporal gradient of the dynamic change characteristics of the forest; The cross-modal similarity between different remote sensing modes is determined by the temporal gradient. Based on the cross-modal similarity between the different remote sensing modalities, a time-series correlation curve is generated between the cross-modal data in the multi-source remote sensing historical data.

4. The method as described in claim 1, characterized in that, Determining the cross-modal temporal deviation values ​​between different remote sensing modes by combining the aforementioned time-series correlation curves with real-time multi-source remote sensing data collected from the target forest area specifically includes: Acquire real-time multi-source remote sensing data of the target forest area; The real-time multi-source remote sensing data is aligned with the time-series correlation curve to obtain the aligned correlation curve. The cross-modal temporal deviation values ​​between different remote sensing modes are calculated using the alignment correlation curve.

5. The method as described in claim 1, characterized in that, The spatial scale normalization process for the real-time multi-source remote sensing data, followed by the calculation of feature overlap at the normalized spatial scale for different remote sensing modalities, specifically includes: The original spatial resolution information of the real-time multi-source remote sensing data is obtained, and then the target resolution is determined based on the original spatial resolution information. The different remote sensing modes in the real-time multi-source remote sensing data are resampled to the target resolution to obtain different remote sensing modes at the normalized spatial scale. Determine the feature overlap of each remote sensing mode at the normalized spatial scale.

6. The method as described in claim 1, characterized in that, Determining the spatial scale fit of the real-time multi-source remote sensing data based on the overlap of all features specifically includes: The overlap ratio of resource management in the target forest area is determined by the overlap of all features; The spatial scale contribution of each remote sensing mode is determined based on the importance of different remote sensing modes in forest resource monitoring in the real-time multi-source remote sensing data. The spatial scale fit of the real-time multi-source remote sensing data is determined by the overlap ratio and the contribution of all spatial scales.

7. The method as described in claim 1, characterized in that, Calculating the confidence level of the fusion features of the real-time multi-source remote sensing data using the cross-modal temporal deviation value and the spatial scale fit specifically includes: Determine the temporal and spatial fusion weight coefficients during the fusion of multi-source remote sensing data in the target forest area; The confidence level of the fusion features of the real-time multi-source remote sensing data is determined based on the fusion weight coefficient, the cross-modal temporal deviation value, and the spatial scale fit.

8. The method as described in claim 1, characterized in that, The specific methods for generating regionalized management parameters for forest resources in the target forest area based on the confidence level of the fused features include: Divide the target forest area into multiple management units; Obtain forest metrics for each management unit; The forest indicators for each management unit are weighted and adjusted based on the confidence level of the fusion features. Regionalized management parameters for forest resources in the target forest area are generated based on the revised forest indicators.

9. The method as described in claim 1, characterized in that, The specific implementation of resource classification management for target forest areas based on the aforementioned regionalized management parameters includes: Set multiple threshold levels for forest resource management in target forest areas; Based on the regionalized management parameters and the set level thresholds, different management units in the target forest area are classified into levels to obtain a management intensity grading map of the target forest area; Based on the aforementioned management intensity grading map, resource management measures for the target forest area are formulated, which are then used to guide the graded management of resources in the target forest area.

10. A forest resource management system that integrates multi-source remote sensing data, characterized in that, include: The acquisition module is used to acquire multi-source remote sensing historical data across modalities in the target forest area; The processing module is used to extract forest dynamic change features from the multi-source remote sensing historical data, generate a time-series correlation curve between cross-modal data in the multi-source remote sensing historical data based on the time-series gradient of the forest dynamic change features, and then combine the time-series correlation curve with real-time multi-source remote sensing data collected in the target forest area to determine the cross-modal time-series deviation value between different remote sensing modes. The processing module is also used to perform spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculate the feature overlap degree of different remote sensing modes under the normalized spatial scale, and determine the spatial scale fit degree of the real-time multi-source remote sensing data based on all feature overlap degrees. The processing module is also used to calculate the fusion feature confidence of the real-time multi-source remote sensing data through the cross-modal temporal deviation value and the spatial scale fit. The execution module is used to generate regionalized management parameters for forest resources in the target forest area based on the confidence level of the fusion features, and to perform resource classification management of the target forest area according to the regionalized management parameters.

Images