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

Abstract

The application provides a forest resource management method and system based on multi-source remote sensing data fusion. Forest dynamic change features are extracted from multi-source remote sensing historical data across modalities, and a time sequence correlation curve between the cross-modal data is generated. The cross-modal time sequence deviation value between different remote sensing modalities is determined by the time sequence correlation curve. The feature coincidence degree of different remote sensing modalities in the normalized spatial scale is calculated for real-time multi-source remote sensing data, and the spatial scale adaptation degree of the real-time multi-source remote sensing data is determined according to all the feature 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. The regionalization management parameters of the forest resources of the target forest region are generated based on the fusion feature confidence, and the resource hierarchical management of the target forest region is performed according to the regionalization management parameters. With the scheme, the precise forest resource hierarchical management can be realized based on the time consistency and spatial consistency of the multi-source remote sensing data fusion.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of forest resource management, and more particularly, to a forest resource management method and system based on multi-source remote sensing data fusion. BACKGROUND

[0002] Forest resource management refers to a management activity that plans, organizes, coordinates and supervises the whole life cycle of forest resources, such as cultivation, protection, utilization and monitoring, through a series of scientific and systematic means and measures, so as to realize the sustainable development of forest resources. The core goal is to balance ecological, economic and social benefits on the premise of guaranteeing the ecological function of forests, to ensure that forest resources can meet the needs of the current generation and do not harm the ability of future generations to meet their needs; effective forest resource management can improve resource utilization efficiency and reduce the risk of ecological damage, and is the key to realizing the sustainable development of forestry.

[0003] Forest resource management based on multi-source remote sensing data fusion refers to a management mode that realizes more accurate, efficient and dynamic forest resource monitoring, evaluation and decision support by integrating different types, different platforms and different time phases of remote sensing data, using data fusion technology to compensate for the limitations of a single data source, and improving the precision, timeliness and reliability of forest resource information extraction; however, in the prior art, the time difference and spatial resolution difference between different modalities are often ignored in the multi-source remote sensing data fusion process, resulting in inconsistent data or feature deviation in the same area, so that the multi-source remote sensing data fusion often faces the problem of insufficient time and spatial consistency processing, thereby causing unstable forest change trend evaluation and difficulty in unifying management indicators, which is difficult to meet the current fine and dynamic forest management needs. Therefore, how to realize accurate forest resource hierarchical management based on the time consistency and spatial consistency of multi-source remote sensing data fusion has become a difficult problem in the industry. SUMMARY

[0004] The present application provides a forest resource management method and system based on multi-source remote sensing data fusion, which can realize accurate forest resource hierarchical management based on the time consistency and spatial consistency of multi-source remote sensing data fusion.

[0005] In a first aspect, the present application provides a forest resource management method based on multi-source remote sensing data fusion, comprising the following steps:

[0006] Obtaining multi-source remote sensing historical data across modalities in a target forest area;

[0007] extracting forest dynamic change features from the multi-source remote sensing historical data, generating a time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data according to a time sequence gradient of the forest dynamic change features, and then determining a cross-modal time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in a target forest region;

[0008] performing spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculating a feature coincidence degree of different remote sensing modalities in a normalized spatial scale, and determining a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all feature coincidence degrees;

[0009] calculating a fusion feature confidence of the real-time multi-source remote sensing data through the cross-modal time sequence deviation value and the spatial scale adaptation degree;

[0010] generating a regionalized management parameter of forest resources in a target forest region based on the fusion feature confidence, and performing resource hierarchical management on the target forest region according to the regionalized management parameter.

[0011] In some embodiments, extracting forest dynamic change features from the multi-source remote sensing historical data specifically includes:

[0012] aligning time axes of different remote sensing modalities in the multi-source remote sensing historical data;

[0013] extracting spectral features, radar features, and three-dimensional structure features of a target forest region based on the multi-source remote sensing historical data after time axis alignment;

[0014] determining forest dynamic change features through the spectral features, the radar features, and the three-dimensional structure features.

[0015] In some embodiments, generating a time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data according to a time sequence gradient of the forest dynamic change features specifically includes:

[0016] determining a time sequence gradient of the forest dynamic change features;

[0017] determining a cross-modal similarity between different remote sensing modalities through the time sequence gradient;

[0018] generating a time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data based on the cross-modal similarity between different remote sensing modalities.

[0019] In some embodiments, determining a cross-modal time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in a target forest region specifically includes:

[0020] acquire real-time multi-source remote sensing data of a target forest area;

[0021] align the real-time multi-source remote sensing data with the time-series correlation curve to obtain an aligned correlation curve;

[0022] calculate a cross-modal time-series deviation value between different remote sensing modalities through the aligned correlation curve.

[0023] In some embodiments, the real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and then the feature coincidence degree of different remote sensing modalities under the normalized spatial scale is calculated, which specifically includes:

[0024] acquire original spatial resolution information of the real-time multi-source remote sensing data, and then determine a target resolution based on the original spatial resolution information;

[0025] resample different remote sensing modalities in the real-time multi-source remote sensing data to the target resolution to obtain different remote sensing modalities under the normalized spatial scale;

[0026] determine the feature coincidence degree of each remote sensing modality under the normalized spatial scale.

[0027] In some embodiments, the spatial scale adaptation degree of the real-time multi-source remote sensing data is determined according to all the feature coincidence degrees, which specifically includes:

[0028] determine a coincidence proportion of resource management of the target forest area through all the feature coincidence degrees;

[0029] determine the spatial scale contribution degree of each remote sensing modality based on the importance of different remote sensing modalities in forest resource monitoring in the real-time multi-source remote sensing data;

[0030] determine the spatial scale adaptation degree of the real-time multi-source remote sensing data from the coincidence proportion and all the spatial scale contribution degrees.

[0031] In some embodiments, the fusion feature confidence of the real-time multi-source remote sensing data is calculated from the cross-modal time-series deviation value and the spatial scale adaptation degree, which specifically includes:

[0032] determine the fusion weight coefficient of time and space in the multi-source remote sensing data fusion process of the target forest area;

[0033] determine the fusion feature confidence of the real-time multi-source remote sensing data based on the fusion weight coefficient, the cross-modal time-series deviation value, and the spatial scale adaptation degree.

[0034] In some embodiments, the regional management parameter of forest resources of the target forest area is generated based on the fusion feature confidence, which specifically includes:

[0035] dividing a plurality of management units in the target forest region;

[0036] obtaining forest indexes of each management unit;

[0037] weighting and correcting the forest indexes of each management unit according to the fusion feature confidence;

[0038] generating regionalized management parameters of forest resources of the target forest region based on the corrected forest indexes.

[0039] In some embodiments, the resource hierarchical management of the target forest region according to the regionalized management parameters specifically comprises:

[0040] setting a plurality of grade thresholds for forest resource management of the target forest region;

[0041] dividing different management units in the target forest region into different grades according to the regionalized management parameters and the set grade thresholds, to obtain a management intensity grading map of the target forest region;

[0042] formulating resource management measures of the target forest region based on the management intensity grading map, to guide the resource hierarchical management of the target forest region.

[0043] In a second aspect, the present application provides a forest resource management system based on multi-source remote sensing data fusion, comprising:

[0044] an acquisition module, configured to acquire multi-source remote sensing historical data of different modalities in a target forest region;

[0045] a processing module, configured to extract forest dynamic change features from the multi-source remote sensing historical data, generate a time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data according to time sequence gradients of the forest dynamic change features, and then determine cross-modal time sequence deviation values between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in the target forest region;

[0046] The processing module is further configured to perform spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculate feature coincidence degrees of different remote sensing modalities in the normalized spatial scale, and determine a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all the feature coincidence degrees.

[0047] The processing module is further configured to calculate a fusion feature confidence of the real-time multi-source remote sensing data by using the cross-modal time sequence deviation values and the spatial scale adaptation degree.

[0048] an execution module, configured to generate regionalized management parameters of forest resources of the target forest region based on the fusion feature confidence, and perform resource hierarchical management of the target forest region according to the regionalized management parameters.

[0049] The technical scheme provided by the embodiments disclosed in the application has the following beneficial effects:

[0050] In the application, multi-source remote sensing historical data across modalities in a target forest region is acquired; forest dynamic change features are extracted from the multi-source remote sensing historical data, a time sequence correlation curve between cross-modality data in the multi-source remote sensing historical data is generated according to a time sequence gradient of the forest dynamic change features, then the time sequence correlation curve is combined with real-time multi-source remote sensing data collected in the target forest region to determine a cross-modality time sequence deviation value between different remote sensing modalities; the real-time multi-source remote sensing data is subjected to spatial scale normalization processing, then a feature coincidence degree of different remote sensing modalities under a normalized spatial scale is calculated, and a spatial scale adaptation degree of the real-time multi-source remote sensing data is determined according to all the feature coincidence degrees; a fusion feature confidence of the real-time multi-source remote sensing data is calculated through the cross-modality time sequence deviation value and the spatial scale adaptation degree; regionalized management parameters of forest resources in the target forest region are generated based on the fusion feature confidence, and resource hierarchical management is performed on the target forest region according to the regionalized management parameters.

[0051] It can be seen that, in the present application, firstly, the cross-modal time sequence deviation value between different remote sensing modalities is determined by combining the time sequence correlation curve with the real-time multi-source remote sensing data collected in the target forest region, which can effectively quantify the time deviation of different remote sensing modalities, so as to unify and correct the multi-source remote sensing fusion data in the time dimension, thereby accurately depicting the forest resource change trend in the target forest region; secondly, the spatial scale adaptation degree of the real-time multi-source remote sensing data is determined according to all the feature coincidence degrees, which can solve the problem of inconsistent spatial information caused by resolution difference of multi-source remote sensing data, make the spatial features of forest resources more accurate and coordinated, ensure the spatial consistency, avoid local error diffusion, and ensure 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 sequence deviation value and the spatial scale adaptation degree, which quantifies the coordination degree of time and space at the same time, not only enhances the robustness of the fusion process, but also dynamically reflects the credibility of the fusion data in different forest regions and time periods, thereby avoiding the deviation of forest resource management caused by the imbalance of a single indicator; finally, the regionalized management parameters of forest resources in the target forest region are generated based on the fusion feature confidence, which as a product of fusion time and spatial consistency can comprehensively reflect the change intensity and management demand of forest resources, and then the target forest region is managed according to the regionalized management parameters, which can realize the transformation of forest management from extensive to fine, provide hierarchical and graded dynamic management and control capability, meet the differentiated and precise needs in actual forest resource management, and promote the modernization and intelligentization of forest resource management; in summary, the scheme can realize precise forest resource hierarchical management based on the time consistency and spatial consistency of multi-source remote sensing data fusion. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0053] Figure 1 is an exemplary flowchart of a multi-source remote sensing data fusion forest resource management method according to some embodiments of the present application;

[0054] Figure 2 is an exemplary flowchart of extracting forest dynamic change features according to some embodiments of the present application;

[0055] Figure 3is an exemplary flowchart of determining a cross-modal timing bias value according to some embodiments of the present application;

[0056] Figure 4 is a structural schematic diagram of a forest resource management system for multi-source remote sensing data fusion according to some embodiments of the present application;

[0057] Figure 5 is a structural schematic diagram of a computer device for implementing a forest resource management method for multi-source remote sensing data fusion according to some embodiments of the present application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0059] Reference Figure 1 The figure is an exemplary flowchart of a forest resource management method for multi-source remote sensing data fusion according to some embodiments of the present application, which mainly includes the following steps:

[0060] In step 101, cross-modal multi-source remote sensing historical data in a target forest region is acquired.

[0061] In specific implementation, the cross-modal multi-source remote sensing historical data in the target forest region can be acquired in the following manner, i.e., the cross-modal multi-source remote sensing historical data in the target forest region can be acquired from official channels (such as the national geographic information public service platform, the European Space Agency, etc.), for example, optical remote sensing images, synthetic aperture (SAR) radar images and laser radar (LiDAR) point cloud data and corresponding acquisition times of the target forest region, wherein the optical remote sensing images are used to reflect the spectral characteristics of vegetation by acquiring the reflection information of different wavebands on the ground surface using a multi-spectral sensor; the SAR radar images are used to acquire the structure and water information on the ground surface by actively emitting microwaves and receiving return signals, and are not affected by clouds and rain; the LiDAR point cloud data are used to acquire the three-dimensional structure and canopy height information of the forest by laser scanning; in addition, the multi-source remote sensing historical data covers at least three complete tree growth cycles of the target forest region; in other embodiments, other methods can also be used for determination, which is not limited here.

[0062] 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.

[0063] 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.

[0064] 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:

[0065] First, in step 1021, the different remote sensing modes in the multi-source remote sensing historical data are aligned along the time axis;

[0066] 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.

[0067] 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.

[0068] In a specific implementation, the time axis alignment of different remote sensing modalities in the multi-source remote sensing historical data can be achieved in the following manner: the time axis alignment of different remote sensing modalities (wherein different remote sensing modalities refer to remote sensing data types of different imaging principles such as optical, microwave, and laser radar) in the multi-source remote sensing historical data, for example, the data collection time of each modality remote sensing data can be sorted, and linear interpolation or nearest neighbor method can be used to map the non-synchronous collected data to a unified time reference (such as monthly, seasonal time nodes), to ensure that different modal remote sensing data can be compared in the time dimension; the extraction of spectral features, radar features, and three-dimensional structure features of the target forest region based on the time axis aligned multi-source remote sensing historical data can be achieved in the following manner: in the time axis aligned multi-source remote sensing historical data, the optical remote sensing data is calculated by using a vegetation index calculation method (such as normalized difference vegetation index, red edge position index), to extract spectral features reflecting forest leaf area index and photosynthetic activity, to reflect vegetation photosynthetic activity, chlorophyll content, and crown layer state; the SAR radar data is calculated by using a polarization decomposition and backscattering coefficient calculation method, to obtain radar features related to forest structure density and water content change, to reflect the structure density, water condition, and roughness of forest vegetation; the LiDAR point cloud data is analyzed by using crown height model construction and vertical profile analysis, to extract three-dimensional structure features such as forest crown height change rate and crown structure complexity, to reflect the vertical structure form and crown distribution of the forest; thereby obtaining the description information of the target forest region in the spectral, microwave, and structure dimensions, to provide multi-angle basis for comprehensive analysis of forest dynamics; the determination of forest dynamic change features by using the spectral features, the radar features, and the three-dimensional structure features can be achieved in the following manner: the change slope and mutation time point of the modality features such as the spectral features, the radar features, and the three-dimensional structure features in the observation period of the target forest region are calculated by using a time series analysis method (such as Theil-Sen trend analysis, Mann-Kendall mutation test), and then the set composed of the results corresponding to each modality feature calculated is taken as the forest dynamic change features; in other embodiments, other methods can also be used for implementation, which are not limited here.

[0069] It should be noted that the time axis alignment in the present application refers to adjusting multi-source remote sensing data acquired at different times to the same time reference system to eliminate the influence of time difference on cross-modal feature analysis; the spectral feature in the present application refers to the reflection / radiation characteristics of ground objects to electromagnetic waves of different wavelengths, which reflects the vegetation growth state of the target forest region; the radar feature in the present application refers to the reflection / scattering characteristics of ground objects to microwave signals, which is affected by the vegetation structure and humidity in the target forest region; the three-dimensional structure feature in the present application refers to the spatial distribution characteristics of the forest in the vertical and horizontal directions, which is used to describe the vertical structure form and canopy distribution of the target forest region; the forest dynamic change feature in the present application refers to a multi-dimensional feature set that can quantify the state change of different modal features (spectral, radar, three-dimensional structure, etc.) of the target forest region in the time dimension, which provides a change signal with clear physical and ecological significance for the subsequent time sequence correlation analysis of cross-modal data, so as to realize the quantifiable comparison of the change trend of forest resources.

[0070] In some embodiments, generating the time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data according to the time sequence gradient of the forest dynamic change feature can be achieved by the following steps:

[0071] determining the time sequence gradient of the forest dynamic change feature;

[0072] determining the cross-modal similarity between different remote sensing modalities through the time sequence gradient;

[0073] generating the time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data based on the cross-modal similarity between different remote sensing modalities.

[0074] In a specific implementation, the time series gradient of the forest dynamic change feature can be determined in the following manner: the rate of change of the forest dynamic change feature (including spectral features, radar features, and three-dimensional structure features) over time can be calculated, that is, the difference of the forest dynamic change feature at adjacent time nodes or the gradient vector based on the central difference formula can be calculated, and the obtained gradient vector is taken as the time series gradient of the forest dynamic change feature. In addition, in order to reduce the interference of high-frequency noise, the forest dynamic change feature can be smoothed by using a Savitzky-Golay smoothing filter or a moving average method, so as to quantify the change speed and direction of the forest feature in the target forest area over time, and establish a comparable numerical basis for subsequent cross-modal comparison. The cross-modal similarity between different remote sensing modalities determined by the time series gradient can be determined in the following manner: the gradient values of the gradient of each modality feature in the time series gradient of the forest dynamic change feature can be converted to the same numerical range by normalization processing (such as Z-score standardization or Min-Max normalization), so as to eliminate the influence caused by different feature dimensions and value range differences, make the change trend of different modal remote sensing data directly comparable in a unified numerical space, improve the accuracy of cross-modal correlation, and then use a multivariate dynamic time warping algorithm or canonical correlation analysis to model the multi-modal correlation of the time series gradient of different remote sensing modalities in the multi-source remote sensing historical data after normalization, so as to take the time change similarity distribution between different remote sensing modalities as the cross-modal similarity between different remote sensing modalities, so as to mine the commonality and difference of different remote sensing modalities in the time change rule. The time series correlation curve between cross-modal data in the multi-source remote sensing historical data based on the cross-modal similarity between different remote sensing modalities can be determined in the following manner: the cross-modal similarity between different remote sensing modalities over time can be plotted as a similarity curve, and the similarity curve can also be fitted by B-spline interpolation or Gaussian regression to weaken the influence of local outliers, so that the curve is smoother and more continuous, wherein the ordinate of the similarity curve represents the correlation strength between different remote sensing modalities, and the abscissa is time. Finally, the similarity curve is taken as the time series 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 change relationship, so as to calculate the cross-modal deviation value subsequently. In other embodiments, other methods can also be used to determine, which are not limited here.

[0075] It should be noted that the time sequence gradient in the present application refers to the change rate of the forest dynamic change characteristics between adjacent observation points in time, which is used to reveal the speed and direction of forest change; the cross-modal similarity in the present application represents the similarity of different remote sensing modalities in the change rule in multi-source remote sensing data; the cross-modal data in the present application refers to the data modalities from different remote sensing platforms or sensors; the time sequence correlation curve in the present application refers to a curve describing the correlation between different remote sensing modalities in the time change rule, which is used to reveal the change correlation and difference mode between different remote sensing modalities in the visualization and quantification level.

[0076] In some embodiments, reference Figure 3 As shown in the figure, which is an exemplary flow chart for determining the cross-modal time sequence deviation value in some embodiments of the present application, the following steps can be used to determine the cross-modal time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with the real-time multi-source remote sensing data collected in the target forest area:

[0077] Obtain real-time multi-source remote sensing data of the target forest area;

[0078] Align the features of the real-time multi-source remote sensing data and the time sequence correlation curve, and obtain an aligned correlation curve;

[0079] Calculate the cross-modal time sequence deviation value between different remote sensing modalities through the aligned correlation curve.

[0080] In a specific implementation, the real-time multi-source remote sensing data of the target forest area can be acquired in the following manner: through a satellite receiving station or a UAV platform, real-time multi-source remote sensing data such as optical remote sensing data, SAR radar data and LiDAR point cloud data are synchronously collected, and preprocessing such as radiometric calibration and geometric correction is completed to ensure the consistency of the collected real-time multi-source remote sensing data and multi-source remote sensing historical data in terms of radiation and spatial characteristics; the real-time multi-source remote sensing data and the time series correlation curve are aligned in feature time to obtain an aligned correlation curve, which can be achieved in the following manner: first, the same feature extraction method as that of the multi-source remote sensing historical data is used to obtain real-time spectral features, radar features and three-dimensional structure features, and the cross-modal similarity between different remote sensing modalities in the real-time multi-source remote sensing data is calculated, and the obtained cross-modal similarity is mapped and registered to the corresponding time nodes of the time series correlation curve according to the corresponding collection time, wherein the nearest time interpolation method or the 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 collection time of the real-time multi-source remote sensing data is between two historical sampling points, the correlation strength between different remote sensing modalities corresponding to the time point is calculated by linear interpolation to ensure that the real-time features and the historical trend are aligned in time, so as to obtain the curve as the aligned correlation curve; the cross-modal time series deviation value between different remote sensing modalities can be calculated in the following manner through the aligned correlation curve: at each time point of the aligned correlation curve, the absolute difference between the correlation strength between different remote sensing modalities corresponding to the real-time multi-source remote sensing data and the correlation strength between different remote sensing modalities on the time series correlation curve corresponding to the multi-source remote sensing historical data is calculated, which reflects the deviation degree between the real-time forest state and the historical cross-modal trend, and then all the obtained absolute difference values at all time points are accumulated and averaged, and finally the obtained value is normalized (i.e., mapped to the interval of 0-1) as the cross-modal time series deviation value between different remote sensing modalities; in other embodiments, other methods can also be used to determine, which are not limited here.

[0081] It should be noted that the real-time multi-source remote sensing data in the present application refers to the multi-source remote sensing data of the target forest region collected at the current time, which is used to reflect the instant state of the target forest region; the feature time alignment in the present application refers to the process of mapping the real-time features of the target forest region to the same time scale as the historical features, which functions to ensure the consistency of the historical and real-time multi-source remote sensing data in time; the alignment correlation curve in the present application refers to 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 cooperative change of different remote sensing modalities of the target forest region at the current time; the cross-modal time sequence deviation value in the present application refers to the difference between the historical cross-modal trend and the real-time cross-modal trend of the target forest region, which is used to quantify the degree of deviation of the current state of the target forest region from the historical change rule, in addition, the cross-modal time sequence deviation value can measure the feature change difference of different remote sensing data types (optical, SAR, LiDAR, etc.) in the same time sequence, so as to reflect the response consistency or difference size of them when monitoring the target forest region.

[0082] In step 103, the real-time multi-source remote sensing data is subjected to spatial scale normalization processing, and then the feature coincidence degree of different remote sensing modalities under the normalized spatial scale is calculated, and the spatial scale adaptation degree of the real-time multi-source remote sensing data is determined according to all the feature coincidence degrees.

[0083] In some embodiments, the feature coincidence degree of different remote sensing modalities under the normalized spatial scale can be realized by the following steps:

[0084] 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;

[0085] The different remote sensing modalities in the real-time multi-source remote sensing data are resampled to the target resolution to obtain different remote sensing modalities under the normalized spatial scale;

[0086] The feature coincidence degree of each remote sensing modality under the normalized spatial scale is determined.

[0087] In a specific implementation, the original spatial resolution information of the real-time multi-source remote sensing data is acquired, and then the target resolution is determined based on the original spatial resolution information. The determination can be implemented in the following manner: the metadata of each data file in the real-time multi-source remote sensing data is read to acquire the original spatial resolution information, and then the original spatial resolution information is combined with the management unit scale of forest resources (for example, the minimum management patch length of a small class) and the anti-aliasing principle to select a target resolution that takes into account the information effectiveness and operability. For example, a safe downsampling scale that is not lower than the highest resolution of each remote sensing modality is selected, or an integer factor of the management unit length is set to divide the grid and small class boundary as much as possible, so as to unify the target resolution, which is physically reasonable and matched with the resource management unit, as the common scale for cross-modal fusion, so as to suppress the noise / aliasing caused by excessive details and avoid the loss of key management features due to rough scale. The different remote sensing modalities in the real-time multi-source remote sensing data are resampled to the target resolution to obtain different remote sensing modalities under the normalized spatial scale. The resampling can be implemented in the following manner: the different remote sensing modalities in the real-time multi-source remote sensing data are resampled to the target resolution by using methods such as wavelet transform, multi-resolution pyramid, or bilinear / three times convolution interpolation. For LiDAR point cloud data, grid data is first generated by grid statistics (such as maximum height and average height), and then resampled according to the target resolution, so as to obtain different remote sensing modalities under the normalized spatial scale, thereby completing the spatial scale normalization processing of the real-time multi-source remote sensing data. The feature coincidence degree of each remote sensing modality under the normalized spatial scale can be determined in the following manner: the normalized multi-modal features (such as spectrum, SAR backscatter, and LiDAR height) are extracted under the normalized spatial scale, and the matching proportion of each remote sensing modality in the feature space is calculated by pixel-by-pixel comparison or regional statistics based on vector boundaries, and then the obtained matching proportion is used as the feature coincidence degree of each remote sensing modality under the normalized spatial scale, so as to quantify the degree of information consistency of different remote sensing modalities under the same spatial scale. In other embodiments, other methods can also be used for determination, which is not limited here.

[0088] It should be noted that the original spatial resolution information in the present application refers to a set of ground sampling accuracy parameters naturally formed by different remote sensing modalities in the data collection stage; the multi-scale spatial normalization processing in the present application refers to the process of converting multi-modal data of 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 the difference in spatial accuracy; the target resolution in the present application refers to the unified pixel size selected by anti-aliasing constraint for different remote sensing modalities, which is the common scale for cross-modal resampling and comparison; the normalized spatial scale in the present application refers to the spatial expression of each remote sensing modality under the same pixel size and the same grid template after geometric consistency and resampling, which is used to eliminate the spatial heterogeneity bias caused by the difference in resolution and sampling grid; the feature coincidence degree in the present application represents the consistency of information of different remote sensing modalities in real-time multi-source remote sensing data under the same spatial scale, which is used to reflect the consistency of expression of different remote sensing modalities to the same phenomenon (such as canopy closure, tree height, humidity) in the same management unit under the normalized spatial scale, so as to understand the consistency and reliability of cross-modal information in the target forest.

[0089] In some embodiments, determining the spatial scale adaptation degree of the real-time multi-source remote sensing data according to all feature coincidence degrees can be achieved by the following steps:

[0090] Determining the coincidence proportion of the resource management of the target forest region through all feature coincidence degrees;

[0091] Determining the spatial scale contribution degree of each remote sensing modality based on the importance of different remote sensing modalities in forest resource monitoring in the real-time multi-source remote sensing data;

[0092] Determining the spatial scale adaptation degree of the real-time multi-source remote sensing data from the coincidence proportion and all spatial scale contribution degrees.

[0093] In a specific implementation, the following methods can be used to determine the coincidence proportion of the target forest area resource management through all feature coincidence degrees: the average value of all feature coincidence degrees can be used as the coincidence proportion of the target forest area resource management; the following methods can be used to determine the spatial scale contribution degree of each remote sensing mode based on the importance of different remote sensing modes in forest resource monitoring in the real-time multi-source remote sensing data: weights can be assigned to each remote sensing mode according to the importance of different modes in forest resource monitoring in the real-time multi-source remote sensing data, and the assignment process can use expert scoring or analytic hierarchy process, for example, LiDAR is more sensitive to stand structure, and the weight can be set to 0.4, and finally the weight assigned to each remote sensing mode is used as the spatial scale contribution degree of each remote sensing mode; the following methods can be used to determine the spatial scale adaptation degree of the real-time multi-source remote sensing data from the coincidence proportion and all spatial scale contribution degrees: the coincidence proportion and the spatial scale contribution degree of each remote sensing mode can be fused, the fusion process can use linear weighting, and the obtained fusion result is normalized to make the value between 0 and 1, and then the finally obtained result is used as the spatial scale adaptation degree of the real-time multi-source remote sensing data; in other embodiments, other methods can also be used for implementation, which are not limited here.

[0094] It should be noted that the coincidence proportion in the present application represents the consistency degree of different remote sensing mode features in the target forest area resource management; the spatial scale contribution degree in the present application represents the comprehensive contribution of the remote sensing mode to the forest resource monitoring target under the condition of spatial scale normalization; and the spatial scale adaptation degree in the present application represents the adaptation degree of the real-time multi-source remote sensing data of the target forest area under the condition of unified spatial scale information fusion.

[0095] In step 104, the fusion feature confidence of the real-time multi-source remote sensing data is calculated through the cross-modal temporal deviation value and the spatial scale adaptation degree.

[0096] In some embodiments, the following steps can be 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 adaptation degree:

[0097] Determine the fusion weight coefficient of time and space in the multi-source remote sensing data fusion process in the target forest area;

[0098] Determine the fusion feature confidence 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 degree.

[0099] In a specific implementation, the determination of the fusion weight coefficient of time and space in the multi-source remote sensing data fusion process in the target forest region can be achieved in the following manner: the influence of the time dimension on the fusion result in different seasons and phenological periods (e.g., the vegetation coverage changes rapidly in the rainy season, and the time weight needs to be increased) can be calculated by analyzing the historical remote sensing data of the target forest region, and the basic weight of the time dimension can be calculated using the coefficient of variation method; secondly, based on the forest type partition (e.g., coniferous forest, broad-leaved forest), the spatial characteristics difference of different regions of the target forest region is quantified using spatial autocorrelation analysis, and a differentiated weight is assigned to the spatial dimension; finally, the proportion of the time and space weights in the multi-source remote sensing data fusion process is adjusted through cross-validation (e.g., dividing the historical data into a training set and a validation set), so that the sum of the two weights is 1 (e.g., the time weight is 0.3 and the space weight is 0.7), to ensure that the fusion weight coefficient reflects both the influence of time dynamics on data consistency and the role of spatial heterogeneity on fusion accuracy. The result obtained after adjustment is used as the fusion weight coefficient of time and space in the multi-source remote sensing data fusion process in the target forest region, wherein the fusion weight coefficient is composed of a time weight and a space weight; the determination of the fusion feature confidence of the real-time multi-source remote sensing data based on the fusion weight coefficient, the cross-modal time sequence bias value, and the spatial scale adaptation degree can be achieved in the following manner: the fusion feature confidence of the real-time multi-source remote sensing data can be calculated using the fusion feature confidence formula constructed, for example: fusion feature confidence = (1-cross-modal time sequence bias value) x time weight + spatial scale adaptation degree x space weight, to comprehensively consider the two core dimensions of time consistency and spatial adaptability, quantify the reliability of the multi-source remote sensing data fusion feature, and provide a reliable feature quality evaluation basis for subsequent forest resource monitoring applications (e.g., biomass estimation, pest detection) in the target forest region. In other embodiments, other methods can also be used for determination, which are not limited here.

[0100] It should be noted that the fusion weight coefficient in the present application represents the relative weight of the time consistency term and the space consistency term in the confidence fusion in the multi-source remote sensing data fusion process in the target forest region; the fusion feature confidence in the present application represents the reliability of the multi-source data in terms of time and space consistency, which is a quality index for measuring the applicability of the multi-source remote sensing data fusion result in the target forest region to forest resource monitoring applications.

[0101] In step 105, the regionalized management parameters of the forest resources of the target forest region are generated based on the fusion feature confidence, and the target forest region is managed in a resource classification manner according to the regionalized management parameters.

[0102] In some embodiments, the generation of the regionalized management parameters of the forest resources of the target forest region based on the fusion feature confidence can be achieved in the following steps:

[0103] dividing a plurality of management units in the target forest region;

[0104] obtaining forest indexes of each management unit;

[0105] correcting the forest indexes of each management unit according to the fusion feature confidence;

[0106] generating regionalized management parameters of forest resources in the target forest region based on the corrected forest indexes.

[0107] In specific implementation, the plurality of management units in the target forest region can be divided in the following manner: the minimum management unit of forest resources in the target forest region can be divided, for example, the existing stand sub-compartment or the regular grid generated based on digital elevation model cutting is taken as the basic unit to divide a plurality of management units in the target forest region; the forest indexes of each management unit can be obtained in the following manner: the forest indexes of each management unit can be estimated from real-time multi-source remote sensing data of the target forest region, for example, stand density, average tree height, canopy density, and stock volume; the forest indexes of each management unit can be corrected according to the fusion feature confidence in the following manner: in each management unit, the forest indexes (such as stand density, average tree height, canopy density, and stock volume) can be subjected to credibility weighting processing based on the fusion feature confidence, for example, the remote sensing data of the high-confidence management unit is directly subjected to weighted correction based on the multi-source remote sensing fusion result, while the remote sensing data of the low-confidence management unit is subjected to weighted correction in combination with historical experience or sample plot investigation data, so as to take advantage of the high-confidence area while avoiding data distortion of the low-confidence area; the regionalized management parameters of forest resources in the target forest region can be generated based on the corrected forest indexes in the following manner: first, a multi-index comprehensive evaluation method, for example, the analytic hierarchy process, the entropy method, or the expert weight assignment method, is used to comprehensively evaluate the corrected forest indexes of each management unit to form a management score reflecting the forest structure, utilization pressure, and resource status in each management unit, and then the range standardization method or the percentage conversion method is used to map the management score of each management unit to a fixed interval (such as 0-100), which is used to reflect the comprehensive level of each management unit in utilization intensity, spatial occupation, and structure stability, and finally the set composed of all the management units and the corresponding management scores is taken as the regionalized management parameters of forest resources in the target forest region to understand the management intervention intensity required by different management units in the target forest region; in other embodiments, other methods can also be used for determination, which is not limited here.

[0108] It should be noted that the management unit in the present application refers to the smallest spatial division unit in the forest resource management of the target forest area, which is used as the spatial carrier for remote sensing data processing and management decision; the forest index in the present application refers to the quantitative parameter that can reflect the forest resource condition and the influence of management activities in the management unit; the weighted correction in the present application refers to the process of weight distribution adjustment of the original forest index of the management unit by fusing the feature confidence, which is used to filter the interference of low-quality data, improve the accuracy of the forest index, and provide a more reliable basis for the generation of subsequent regionalized management parameters; the regionalized management parameter in the present application represents the required management intervention intensity of different management units in the target forest area, which can directly present the intensity difference of forest resource management of different management units in the target forest area, and provide decision basis for formulating scientific forest resource management strategy and optimizing resource allocation.

[0109] In some embodiments, the resource hierarchical management of the target forest area according to the regionalized management parameter can be realized by the following steps:

[0110] Setting a plurality of grade thresholds for the forest resource management of the target forest area;

[0111] According to the regionalized management parameter and the set grade threshold, the different management units in the target forest area are divided into grades to obtain a management intensity grading map of the target forest area;

[0112] Based on the management intensity grading map, the resource management measures of the target forest area are formulated, which are used to guide the resource hierarchical management of the target forest area.

[0113] In a specific implementation, the setting of the multiple level thresholds for the forest resource management of the target forest region can be implemented in the following manner: multiple level thresholds for the forest resource management of the target forest region can be set according to the value range of the regionalization management parameter, for example, divided into three management levels of low, medium and high, wherein the level thresholds can be determined in combination with national forest management standards, expert experience and historical survey data, for example: when the value range of the regionalization management parameter is 0-100, the level threshold setting can be: 0-30 is defined as low intensity, 31-70 is defined as medium intensity, and 71-100 is defined as high intensity; the level division of different management units in the target forest region according to the regionalization management parameter and the set level threshold to obtain the management intensity classification map of the target forest region can be implemented in the following manner: different management units in the target forest region are divided into levels according to the regionalization management parameter and the set level threshold to generate a management intensity classification map covering the target forest region; the development of resource management measures for the target forest region based on the management intensity classification map and the use of the management intensity classification map to guide the resource classification management of the target forest region can be implemented in the following manner: differentiated resource management measures can be developed for different levels of management units, for example: low-level units only need to be regularly patrolled and naturally conserved; medium-level units need to be arranged for moderate tending and felling, and supplementary planting and updating; high-level units need to be intervened, for example, clear-cutting and updating, comprehensive pest control or soil and water conservation engineering, and then the management intensity classification map and the resource management measures are stored in the forest resource management database to realize spatialized display and traceable recording, so as to guide the resource classification management of the target forest region; in other embodiments, other methods can also be used for implementation, which are not limited here.

[0114] It should be noted that the level threshold in the present application refers to the numerical limit for dividing the different management intensity levels in the target forest region, which provides a quantitative boundary for the classification management of the forest unit; the management intensity classification map in the present application refers to the spatialized management intensity level distribution map of the target forest region; the resource classification management in the present application refers to the management method of dividing the different management units of the target forest region into levels based on the regionalization management parameter and matching differentiated management measures for different levels, which is used to convert the complex forest resource state into an operable classification management scheme to realize scientific, fine and dynamic forest resource management.

[0115] In addition, another aspect of the present application, in some embodiments, the present application provides a forest resource management system based on multi-source remote sensing data fusion, referring to Figure 4 The figure is a structural schematic diagram of a forest resource management system based on multi-source remote sensing data fusion according to some embodiments of the present application, which comprises an acquisition module 401, a processing module 402 and an execution module 403, which are described as follows:

[0116] The acquisition module 401 is mainly used for acquiring multi-source remote sensing historical data of different modalities in a target forest region in the present application.

[0117] The processing module 402 is mainly used for extracting forest dynamic change features from the multi-source remote sensing historical data, generating a time sequence correlation curve between cross-modal data in the multi-source remote sensing historical data according to a time sequence gradient of the forest dynamic change features, and then determining a cross-modal time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve and real-time multi-source remote sensing data collected in the target forest region.

[0118] The processing module 402 is also used for performing spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculating a feature coincidence degree of different remote sensing modalities in a normalized spatial scale, and determining a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all feature coincidence degrees.

[0119] The processing module 402 is also used for calculating a fusion feature confidence of the real-time multi-source remote sensing data through the cross-modal time sequence deviation value and the spatial scale adaptation degree.

[0120] The execution module 403 is mainly used for generating a regionalized management parameter of forest resources in a target forest region based on the fusion feature confidence, and performing resource hierarchical management on the target forest region according to the regionalized management parameter.

[0121] The above describes an example of a forest resource management method and system for multi-source remote sensing data fusion provided by the embodiments of the present application in detail. It can be understood that a corresponding device includes a hardware structure and / or a software module corresponding to each function to achieve the above functions. Those skilled in the art should easily realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is executed by hardware or computer software driven hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0122] In some embodiments, the present application also provides a computer device including a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned forest resource management method for multi-source remote sensing data fusion.

[0123] In some embodiments, reference Figure 5 The 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.

[0124] 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.

[0125] 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.

[0126] 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.

[0127] 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.

[0128] 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.

[0129] It should be understood that each step of the above method embodiments can be accomplished by logic circuits in the form of hardware or instructions in the form of software in the processor 501, which 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 device, discrete gate or transistor logic, discrete hardware components, or any combination thereof.

[0130] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0131] For example, in some embodiments, the present application also provides a computer readable storage medium, wherein instructions or codes are stored in the computer readable storage medium, and when the instructions or codes are run on a computer, the computer is caused to perform the above-mentioned forest resource management method of multi-source remote sensing data fusion.

[0132] Although preferred embodiments of the application have been described, those skilled in the art will appreciate that additional modifications and variations to the preferred embodiments can be made without departing from the spirit and scope of the application. Accordingly, it is intended that the appended claims be interpreted as including all such alterations and modifications as fall within the scope of the application.

[0133] Obviously, various modifications and changes can be made to the present application by those skilled in the art without departing from the spirit and scope of the present application. Thus, it is intended that the present application cover the modifications and changes as fall within the scope of the claims and their equivalents.

Claims

1. A forest resource management method of multi-source remote sensing data fusion, characterized in that, The method comprises the following steps: acquiring multi-source remote sensing historical data of different modalities in a target forest region; extracting forest dynamic change features from the multi-source remote sensing historical data, generating a time sequence correlation curve between cross-modality data in the multi-source remote sensing historical data according to a time sequence gradient of the forest dynamic change features, and then determining a cross-modality time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in the target forest region; performing spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculating feature coincidence degrees of different remote sensing modalities in the normalized spatial scale, and determining a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all the feature coincidence degrees; calculating a fusion feature confidence of the real-time multi-source remote sensing data through the cross-modality time sequence deviation value and the spatial scale adaptation degree; generating a regionalized management parameter of forest resources in the target forest region based on the fusion feature confidence, and performing resource hierarchical management on the target forest region according to the regionalized management parameter; wherein generating the time sequence correlation curve between cross-modality data in the multi-source remote sensing historical data according to the time sequence gradient of the forest dynamic change features comprises: determining the time sequence gradient of the forest dynamic change features; determining a cross-modality similarity between different remote sensing modalities through the time sequence gradient; generating the time sequence correlation curve between cross-modality data in the multi-source remote sensing historical data based on the cross-modality similarity between different remote sensing modalities.

2. The method of claim 1, wherein, Extracting forest dynamic change features from the multi-source remote sensing historical data comprises: aligning time axes of different remote sensing modalities in the multi-source remote sensing historical data; extracting spectral features, radar features and three-dimensional structure features of the target forest region based on the multi-source remote sensing historical data after time axis alignment; determining forest dynamic change features through the spectral features, the radar features and the three-dimensional structure features.

3. The method of claim 1, wherein, Determining a cross-modality time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in the target forest region comprises: acquiring real-time multi-source remote sensing data of the target forest region; performing feature time sequence alignment on the real-time multi-source remote sensing data and the time sequence correlation curve to obtain an aligned correlation curve; calculating a cross-modality time sequence deviation value between different remote sensing modalities through the aligned correlation curve.

4. The method of claim 1, wherein, Performing spatial scale normalization processing on the real-time multi-source remote sensing data, and then calculating feature coincidence degrees of different remote sensing modalities in the normalized spatial scale, and determining a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all the feature coincidence degrees, comprises: acquiring original spatial resolution information of the real-time multi-source remote sensing data, and then determining a target resolution based on the original spatial resolution information; resampling different remote sensing modalities in the real-time multi-source remote sensing data to the target resolution to obtain different remote sensing modalities in the normalized spatial scale; determining feature coincidence degrees of each remote sensing modality in the normalized spatial scale.

5. The method of claim 1, wherein, Determining a spatial scale adaptation degree of the real-time multi-source remote sensing data according to all the feature coincidence degrees comprises: determining a coincidence proportion of resource management of the target forest region through all the feature coincidence degrees; determining a spatial scale contribution degree of each remote sensing modality based on importance of different remote sensing modalities in forest resource monitoring in the real-time multi-source remote sensing data; determining a spatial scale fitness degree of the real-time multi-source remote sensing data from the coincidence ratio and all spatial scale contribution degrees.

6. The method of claim 1, wherein, calculating a fusion feature confidence degree of the real-time multi-source remote sensing data through the cross-modality time sequence deviation value and the spatial scale fitness degree specifically includes: determining a fusion weight coefficient of time and space in a multi-source remote sensing data fusion process in a target forest region; determining a fusion feature confidence degree of the real-time multi-source remote sensing data based on the fusion weight coefficient, the cross-modality time sequence deviation value and the spatial scale fitness degree.

7. The method of claim 1, wherein, generating a regionalized management parameter of forest resources in a target forest region based on the fusion feature confidence degree specifically includes: dividing a plurality of management units in the target forest region; obtaining forest indexes of each management unit; weighting and correcting the forest indexes of each management unit according to the fusion feature confidence degree; generating a regionalized management parameter of forest resources in the target forest region based on the corrected forest indexes.

8. The method of claim 1, wherein, performing resource hierarchical management on the target forest region according to the regionalized management parameter specifically includes: setting a plurality of grade thresholds for forest resource management of the target forest region; performing grade division on different management units in the target forest region according to the regionalized management parameter and the set grade thresholds to obtain a management intensity grading map of the target forest region; formulating resource management measures of the target forest region based on the management intensity grading map, and further used for guiding resource hierarchical management of the target forest region.

9. A forest resource management system using multi-source remote sensing data fusion, which adopts the method of any one of claims 1 to 8 for forest resource management, characterized in that, The forest resource management system includes: an acquisition module configured to acquire multi-source remote sensing historical data of cross modalities in a target forest region; a processing module configured to extract forest dynamic change features from the multi-source remote sensing historical data, generate a time sequence correlation curve between cross-modality data in the multi-source remote sensing historical data according to a time sequence gradient of the forest dynamic change features, and further determine a cross-modality time sequence deviation value between different remote sensing modalities by combining the time sequence correlation curve with real-time multi-source remote sensing data collected in the target forest region; the processing module is further configured to perform spatial scale normalization processing on the real-time multi-source remote sensing data, and further calculate a feature coincidence degree of different remote sensing modalities under a normalized spatial scale, and determine a spatial scale fitness degree of the real-time multi-source remote sensing data according to all feature coincidence degrees; the processing module is further configured to calculate a fusion feature confidence degree of the real-time multi-source remote sensing data through the cross-modality time sequence deviation value and the spatial scale fitness degree; an execution module configured to generate a regionalized management parameter of forest resources in a target forest region based on the fusion feature confidence degree, and perform resource hierarchical management on the target forest region according to the regionalized management parameter.

Images