Tree diversity monitoring method and system for large-area forest
By combining UAV lidar and ground-based measurement data, a single-tree information set is constructed, which solves the problems of low efficiency and insufficient accuracy in monitoring tree diversity in large-area forests, and realizes efficient and accurate tree diversity monitoring and ecosystem management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-05
- Publication Date
- 2026-03-10
AI Technical Summary
Existing methods for monitoring forest tree diversity cannot achieve efficient and accurate monitoring in large areas of forest, especially in dangerous areas. Furthermore, existing technologies cannot accurately extract key linear size features of trees, have low tree species identification accuracy, and cannot obtain data related to tree growth stages and lifespan.
Large-scale, precise measurements are conducted using drone-based lidar, and benchmark calibration is performed using ground-based measured data. This allows for the construction of individual tree information sets. Through feature recognition and environmental monitoring, tree growth trends are deduced, and tree diversity reports are generated.
It enables efficient and accurate monitoring of tree diversity in large-scale forests, providing scientific evidence for forest resource management and ecological environment assessment, and generating reports that include the current status of diversity, tree species distribution patterns and future trends, supporting policy-making by forestry departments and research by scientific researchers.
Smart Images
Figure CN121634128A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of forest resource monitoring, in particular to a tree diversity monitoring method and system for large-area forests. BACKGROUND
[0002] Forest tree diversity is a core indicator of ecosystem stability and function, especially in areas such as virgin forests and high mountain forests that are difficult for humans to enter or where there are risks of wildlife and geological disasters. Diversity monitoring is a key requirement for biodiversity protection, forest resource management, and ecological environment assessment. However, existing forest tree diversity monitoring methods have many defects, including: First, relying on staff to investigate each tree in the forest, although accurate species information and growth parameters can be obtained, the efficiency is extremely low, it is difficult to cover large-area forests, and there are safety hazards in dangerous areas, making it difficult to achieve regular monitoring; Second, using satellite remote sensing or unmanned aerial vehicle visible light remote sensing for large-scale monitoring, but due to vegetation obstruction and resolution limitations, it is difficult to accurately extract key linear size features such as tree diameter and crown width, the tree species recognition accuracy is low, and it is difficult to directly obtain tree growth stage and life-related data; Therefore, there is an urgent need for a method and system that can efficiently and accurately monitor the tree diversity, growth status, and trends of large-area dangerous forests. SUMMARY
[0004] The present application provides a tree diversity monitoring method and system for large-area forests, which realizes integrated monitoring of tree species, growth status, diversity, and life in dangerous and difficult-to-enter areas through large-scale accurate measurement by unmanned aerial vehicle laser radar and benchmark calibration of ground measurement data, improves monitoring efficiency and accuracy, and provides technical support for forest resource management.
[0005] The present application provides a tree diversity monitoring method for large-area forests, comprising: Step 1: On-site measurement of target forests to obtain a plurality of sets of measurement data, identifying the tree information contained in each set of measurement data, and constructing a single tree information set of the target forest; Step 2: Feature recognition of each single tree information set to determine the types of trees contained in the target forest and the distribution range of each tree type in the target forest; Step 3: Environmental monitoring of each distribution range to derive the growth trend of each tree species in the target forest and determine the vegetation coverage of the target forest at different prediction time periods; Step 4: Analyze the increase and decrease of tree species for each predicted time period based on the vegetation coverage, generate a tree diversity report and display it.
[0006] In one feasible embodiment, step 1 includes: Step 11: Generate UAV flight trajectory based on the forest area characteristics of the target forest, control the UAV to perform radar scan on the target forest according to the UAV flight trajectory, obtain several sets of measured data of the target forest, and arrange the measured data according to the scan coverage area corresponding to each set of measured data. Step 12: Perform tree identification on the data arrangement results to identify several trees in the target forest, and obtain the tree appearance features corresponding to each identified tree. Map each tree appearance feature to the corresponding measured data to obtain the unidentified data contained in each measured data. Step 13: Preprocess each of the unidentified data to determine the data coverage area corresponding to each of the unidentified data, construct the identified range of the target forest based on the tree appearance characteristics corresponding to each identified tree, and determine the range correlation features between each of the data coverage areas and the identified range. Step 14: Draw several possible presentation contours of the unidentified data according to the data presentation characteristics of each unidentified data in the corresponding data coverage area, filter the matching possible presentation contours corresponding to each unidentified data based on the corresponding range-related features, update the identified range to obtain several tree information, and construct a single tree information set.
[0007] In one feasible approach, the process of filtering matching suspected presentation contours corresponding to each of the unidentified data based on the corresponding range-related features, updating the identified range to obtain several tree information sets, and constructing a single tree information set includes: Based on the range-related features, determine a number of contour connection positions between the data coverage area and the corresponding identified area, as well as the connection relationship corresponding to each contour connection position; In each of the suspected contours, locate the contour presentation position corresponding to the contour connection position, and identify the degree of matching between each contour presentation position and the identified range based on the connection relationship. The suspected presentation contour with the highest sum of matching degree is regarded as the suspected presentation contour corresponding to the unidentified data; Based on the connection relationship, each of the matched suspected presentation contours is used to expand the identified range to obtain the full identification range of the target forest; Divide all the identification information contained within the entire identification range into several tree information; Information on trees of the same type is aggregated to obtain a set of individual tree information for the target forest.
[0008] In one feasible embodiment, step 2 includes: Step 21: Identify several tree features contained in each of the single tree information sets, and draw the tree appearance corresponding to each single tree information set. Perform iterative spatial registration between the tree features and the tree appearance to generate a preliminary tree distribution map of the target forest. Step 22: Use AI to identify several tree types contained in the preliminary tree distribution map, and construct tree samples corresponding to each tree type. Use each tree sample to traverse the preliminary tree distribution map to obtain a preliminary distribution map of each tree type in the target forest. Step 23: Overlay and map all the preliminary distribution maps to obtain the identified tree map of the target forest. Simultaneously, construct the overall tree concept map of the target forest based on the measured data. When the identified tree map and the overall tree concept map... Figure One At that time, the distribution range of the corresponding tree type in the target forest is determined according to the preliminary distribution map; Step 24: Conversely, each tree sample is used to traverse the image differences between the identified tree map and the overall tree concept map. The preliminary tree distribution map is updated based on the traversal results to obtain the updated distribution map corresponding to each tree type. The distribution range of the tree type in the target forest is determined by combining the corresponding preliminary distribution map.
[0009] In one implementable manner, step 24 includes: Step 241: When the identified tree map and the overall tree concept map are inconsistent, draw the outline of the unidentified area of the target forest according to the different information of the image, and use each tree sample to traverse the outline of the unidentified area to obtain the updated identification information corresponding to each unidentified area outline. Step 242: Update the preliminary tree distribution map using the updated identification information, and record several update information of the preliminary tree distribution map. Match each update information with several corresponding identified tree maps, and update each identified tree map according to the matching results to obtain the corresponding updated distribution map. Step 243: Based on the identified tree map and updated distribution map corresponding to each tree type, redraw the tree distribution of the tree type in the target forest to obtain the distribution range of each tree type in the target forest.
[0010] In one feasible embodiment, step 3 includes: Step 31: Deploy several monitoring devices in each distribution range using drones, construct monitoring environment information corresponding to the distribution range based on the monitoring range corresponding to each monitoring device, and simultaneously acquire climate information of the environment where the target forest is located to construct climate-environment characteristics corresponding to each distribution range. Step 32: Construct a simulated environment using the climate-environment characteristics, simulate the growth trend corresponding to the tree type in the simulated environment, and identify the positive growth time period and negative growth time period corresponding to the tree type in the growth trend; Step 33: Determine the current vegetation coverage of the target forest based on the measured data, deduce the growth data in the corresponding positive growth period and the decay data in the corresponding negative growth period based on the growth trend of each tree type, and use the growth data and decay data to perform dynamic analysis on the current vegetation coverage to obtain the vegetation coverage of the target forest in several predicted time periods.
[0011] In one feasible embodiment, step 4 includes: Step 41: Based on the vegetation coverage corresponding to different prediction time periods and the growth trend of each tree species in the target forest, determine the tree increase area and tree decrease area corresponding to each prediction time period, and determine the increase reason for each tree increase area and the decrease reason for each tree decrease area respectively. Step 42: Based on each of the reasons for increase, reverse deduce several first regional environmental characteristics of the target forest corresponding to different prediction time periods, and based on each of the reasons for decrease, reverse deduce several corresponding second environmental characteristics of the target forest corresponding to different prediction time periods. Step 43: Construct the environmental feature changes of the target forest based on the first environmental feature and the second environmental feature corresponding to different prediction time periods, and obtain the trees that can grow and do not match the environmental feature changes. Step 44: Analyze the growth probability of each of the growable trees based on the measured data, and analyze the growth inhibition probability of each of the existing trees due to environmental changes, determine the increase and decrease of tree species in the target forest during different predicted time periods, generate a tree diversity report and display it.
[0012] One feasible approach also includes: When the reduction ratio of the tree species is higher than the prescribed ratio, it is determined that the target forest has been damaged, and the corresponding distribution range is monitored in a key area and a real-time report is generated for display.
[0013] This invention provides a tree diversity monitoring system for large-area forests, comprising: The information collection module is used to conduct on-site measurements of the target forest to obtain several sets of measured data, identify the information of several trees contained in each set of measured data, and construct a single tree information set of the target forest. The distribution analysis module is used to perform feature recognition on each of the single tree information sets, determine the several tree types contained in the target forest, and determine the distribution range of each tree type in the target forest. The trend analysis module is used to perform environmental monitoring on each of the distribution ranges, deduce the growth trend of each tree species in the target forest, and determine the vegetation coverage of the target forest at different predicted time periods. The report generation module is used to analyze the increase and decrease of tree species corresponding to each predicted time period based on the vegetation coverage, generate a tree diversity report, and display it.
[0014] In one implementable manner, the information acquisition module includes: The scanning execution unit is used to generate a UAV flight trajectory based on the forest area characteristics of the target forest, control the UAV to perform radar scanning on the target forest according to the UAV flight trajectory, obtain several sets of measured data of the target forest, and arrange the measured data according to the scanning coverage area corresponding to each set of measured data. The tree identification unit is used to identify several trees in the target forest by performing tree identification on the data arrangement results, and to obtain the tree appearance features corresponding to each identified tree, and to map each tree appearance feature to the corresponding measured data to obtain the unidentified data contained in each measured data. The range analysis unit is used to preprocess each of the unidentified data to determine the data coverage range corresponding to each of the unidentified data, construct the identified range of the target forest based on the tree appearance characteristics corresponding to each identified tree, and determine the range correlation features between each of the data coverage ranges and the identified ranges. The information confirmation unit is used to draw several possible presentation contours of each unidentified data according to the data presentation characteristics of each unidentified data in the corresponding data coverage area, filter the matching possible presentation contours corresponding to each unidentified data based on the corresponding range-related features, update the identified range to obtain several tree information, and construct a single tree information set. The beneficial effects of the above technical solution are as follows: To accurately adapt to the monitoring needs of large-area forests, firstly, comprehensive on-site measurements are conducted on the target forest to obtain measured data including tree diameter at breast height (DBH), tree height, crown width, and tree species morphology. Then, the independent information of each tree is accurately identified to construct a single-tree information set. Next, feature recognition is performed on the single-tree information set to accurately determine the tree type, and the distribution range of each tree type is clarified through spatial positioning data. This technical method can shorten the data collection time to several weeks, avoiding misidentification of tree species due to insufficient experience in manual identification. Furthermore, targeted environmental monitoring is conducted based on the distribution range of each tree type, and growth trends are deduced by combining tree growth patterns, thereby predicting... By dynamically predicting vegetation cover and tree species change trends at different time periods, a scientific basis for proactive management of forest ecosystems can be provided. Finally, based on the vegetation cover at different predicted time periods, the potential for increasing or decreasing tree species can be accurately analyzed, generating a tree diversity report that includes the current diversity status, tree species distribution patterns, future change trends, and protection and intervention recommendations for managers. This approach can provide core data support for forestry departments to formulate protection policies and ecological restoration plans, and also provide basic data for researchers to conduct forest ecological research, effectively improving the scientific rigor and practicality of forest diversity monitoring, and providing reliable technical support for forest resource protection and ecological balance maintenance.
[0015] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.
[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram illustrating the workflow of a tree diversity monitoring method for large-area forests according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the composition of a tree diversity monitoring system for large-area forests according to an embodiment of the present invention. Detailed Implementation
[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.
[0021] Example 1: This example provides a method for monitoring tree diversity in large-area forests, such as... Figure 1 As shown, it includes: Step 1: Conduct on-site measurements of the target forest to obtain several sets of measured data, identify the information of several trees contained in each set of measured data, and construct a single tree information set for the target forest; Step 2: Perform feature recognition on each of the single tree information sets to determine the several tree types contained in the target forest, and determine the distribution range of each tree type in the target forest; Step 3: Conduct environmental monitoring for each of the aforementioned distribution ranges, deduce the growth trend of each of the aforementioned tree species in the target forest, and determine the vegetation coverage of the target forest at different predicted time periods; Step 4: Analyze the increase and decrease of tree species for each predicted time period based on the vegetation coverage, generate a tree diversity report and display it.
[0022] In this example, the measured data represents the data obtained through laser scanning, which includes information such as tree diameter at breast height, tree height, crown width, and tree species morphology; In this example, the single tree information set contains all the appearance information of a tree; In this example, tree type refers to the various types of trees contained in the target forest; In this example, the distribution range refers to the range corresponding to the distribution of the same tree type in the target forest; In this example, the prediction period is one year. In this example, there are two scenarios for adding tree species: the first is that a species that did not exist before appears, and the second is that the number of existing species increases. In this example, reducing the number of tree species means that the number of existing species has been reduced.
[0023] The working principle and beneficial effects of the above technical solution are as follows: To accurately adapt to the monitoring needs of large-area forests, firstly, comprehensive on-site measurements are conducted on the target forest to obtain measured data including tree diameter at breast height (DBH), tree height, crown width, and tree species morphology. Then, the independent information of each tree is accurately identified to construct a single-tree information set. Next, feature recognition is performed on the single-tree information set to accurately determine the tree type, and the distribution range of each tree type is clarified through spatial positioning data. This technical method can shorten the data collection time to several weeks, avoiding misidentification of tree species due to insufficient experience in manual identification. Furthermore, targeted environmental monitoring is conducted based on the distribution range of each tree type, and growth trends are deduced by combining tree growth patterns, thereby predicting... By dynamically predicting vegetation cover and tree species change trends at different time periods, a scientific basis for proactive management of forest ecosystems can be provided. Finally, based on the vegetation cover at different predicted time periods, the potential for increasing or decreasing tree species can be accurately analyzed, generating a tree diversity report that includes the current diversity status, tree species distribution patterns, future change trends, and protection and intervention recommendations for managers. This approach can provide core data support for forestry departments to formulate protection policies and ecological restoration plans, and also provide basic data for researchers to conduct forest ecological research, effectively improving the scientific rigor and practicality of forest diversity monitoring, and providing reliable technical support for forest resource protection and ecological balance maintenance.
[0024] Example 2: Based on Example 1, the method for monitoring tree diversity in large-area forests is characterized in that step 1 includes: Step 11: Generate UAV flight trajectory based on the forest area characteristics of the target forest, control the UAV to perform radar scan on the target forest according to the UAV flight trajectory, obtain several sets of measured data of the target forest, and arrange the measured data according to the scan coverage area corresponding to each set of measured data. Step 12: Perform tree identification on the data arrangement results to identify several trees in the target forest, and obtain the tree appearance features corresponding to each identified tree. Map each tree appearance feature to the corresponding measured data to obtain the unidentified data contained in each measured data. Step 13: Preprocess each of the unidentified data to determine the data coverage area corresponding to each of the unidentified data, construct the identified range of the target forest based on the tree appearance characteristics corresponding to each identified tree, and determine the range correlation features between each of the data coverage areas and the identified range. Step 14: Draw several possible presentation contours of the unidentified data according to the data presentation characteristics of each unidentified data in the corresponding data coverage area, filter the matching possible presentation contours corresponding to each unidentified data based on the corresponding range-related features, update the identified range to obtain several tree information, and construct a single tree information set.
[0025] In this example, forest footprint characteristics refer to the forest's area, topography, and elevation. In this example, the drone flight path represents the flight trajectory of the drone when it is collecting data; In this example, identifying trees means recognizing the trees in the data arrangement results; In this example, tree appearance features refer to the characteristics presented by the tree's appearance; In this example, unidentified data refers to data in the measured data that was not identified; In this example, the data coverage area represents the range corresponding to the unidentified data; In this example, preprocessing includes: denoising, enhancement, and coordinate calibration; In this example, the range-related feature represents the positional relationship between the data coverage area and the identified area; In this example, the apparent outline represents an outline constructed based on unidentified data; In this example, matching suspected rendering contours refers to suspected rendering contours that match unidentified data.
[0026] The working principle and beneficial effects of the above technical solution are as follows: To achieve efficient collection and accurate database construction of information on trees in large-area forests, a customized UAV flight trajectory is first generated based on the land area characteristics of the target forest. Then, the UAV is controlled to acquire measured data through radar scanning, and the measured data is arranged in an orderly manner according to the scanning coverage area, forming a structured dataset from massive amounts of measured data. This avoids inefficiency caused by data clutter in subsequent analysis, providing comprehensive and orderly basic data for accurate tree identification. Next, tree identification is performed on the orderly arranged measured data to identify the trees and extract their appearance features. These appearance features are then mapped back into the measured data to accurately locate unidentified data. Finally, the unidentified data is preprocessed to clarify its data coverage area. The system first identifies the distribution range of trees and then analyzes the range-related characteristics of the two to provide spatial constraints for the subsequent discovery of unidentified trees. Finally, for the unidentified data, several possible outlines are drawn based on their data presentation characteristics. These outlines are then selected and matched based on the range-related characteristics of the previous steps. The identified range is then updated, and all tree information is integrated to construct a single tree information set. In this way, full-area coverage can be achieved through customized drone trajectories. By identifying rare broad-leaved trees hidden in the valley through layered identification and eliminating false rock identifications through outline matching, the final complete information set provides a solid data guarantee for subsequent tree species classification, distribution range analysis, and growth trend prediction, thereby improving the scientificity and reliability of the entire monitoring method from the source.
[0027] Example 3: Based on Example 2, the method for monitoring tree diversity in large-area forests includes the following steps: Based on the corresponding range-related features, filtering the matching suspected outlines corresponding to each unidentified data point, updating the identified range to obtain several tree information points, and constructing a single tree information set. Based on the range-related features, determine a number of contour connection positions between the data coverage area and the corresponding identified area, as well as the connection relationship corresponding to each contour connection position; In each of the suspected contours, locate the contour presentation position corresponding to the contour connection position, and identify the degree of matching between each contour presentation position and the identified range based on the connection relationship. The suspected presentation contour with the highest sum of matching degree is regarded as the suspected presentation contour corresponding to the unidentified data; Based on the connection relationship, each of the matched suspected presentation contours is used to expand the identified range to obtain the full identification range of the target forest; Divide all the identification information contained within the entire identification range into several tree information; Information on trees of the same type is aggregated to obtain a set of individual tree information for the target forest.
[0028] In this example, the contour connection position indicates the position where the connection operation can be performed when connecting the data coverage area and the identified area; In this example, the connection relationship represents the method used to connect the data coverage area and the identified area, as well as the appearance after the connection is completed; In this example, the degree of matching indicates the degree of fusion when the outline's position is connected to the already identified area; In this example, the total identified range represents the range that has already been identified in the target forest.
[0029] The working principle and beneficial effects of the above technical solution are as follows: It utilizes range-related features to determine the contour connection positions and quantification relationships, providing a reference for matching unidentified contours and avoiding misjudgment of non-tree targets with no reasonable connection, thus laying the foundation for accurate matching. Furthermore, it filters the optimal contour through multi-dimensional matching degree calculation and sum-based sorting, covering factors such as spatial distance, terrain environment, and vegetation distribution, significantly reducing the misjudgment rate. This technology aligns with the spatial correlation characteristics of large-area forests where trees are distributed in patches and saplings grow around mature trees. It can also expand the identified range based on the matched contours, achieving seamless connection and coverage of the monitoring range without blind spots. Ensuring the ecological continuity between newly added areas and existing areas provides a complete spatial basis for subsequent tree species distribution analysis and growth trend deduction. Furthermore, by classifying individual tree information and aggregating similar information, a structured single-tree information set is constructed. This ensures the traceability of information for each tree and allows data to be clearly presented according to tree species classification. It supports the efficient implementation of subsequent steps such as tree type determination and diversity report generation. Ultimately, it solves the problems of high misjudgment, scattered scope, and disorganized information in traditional tree information extraction. It establishes an accurate, complete, and highly usable information foundation for large-scale forest tree diversity monitoring, significantly improving the scientific nature and practicality of the overall monitoring.
[0030] Example 4: Based on Example 1, the method for monitoring tree diversity in large-area forests, step 2 includes: Step 21: Identify several tree features contained in each of the single tree information sets, and draw the tree appearance corresponding to each single tree information set. Perform iterative spatial registration between the tree features and the tree appearance to generate a preliminary tree distribution map of the target forest. Step 22: Use AI to identify several tree types contained in the preliminary tree distribution map, and construct tree samples corresponding to each tree type. Use each tree sample to traverse the preliminary tree distribution map to obtain a preliminary distribution map of each tree type in the target forest. Step 23: Overlay and map all the preliminary distribution maps to obtain the identified tree map of the target forest. Simultaneously, construct the overall tree concept map of the target forest based on the measured data. When the identified tree map and the overall tree concept map... Figure One At that time, the distribution range of the corresponding tree type in the target forest is determined according to the preliminary distribution map; Step 24: Conversely, each tree sample is used to traverse the image differences between the identified tree map and the overall tree concept map. The preliminary tree distribution map is updated based on the traversal results to obtain the updated distribution map corresponding to each tree type. The distribution range of the tree type in the target forest is determined by combining the corresponding preliminary distribution map.
[0031] In this example, iterative spatial registration represents the process of aligning tree features with another image or tree features through multiple iterations of optimization. In this example, the preliminary tree distribution map represents an image showing the distribution of trees in the target forest; In this example, the preliminary distribution map shows the distribution of trees of the same tree type in the target forest. The preliminary tree distribution map contains both the individual characteristics of the trees and clear spatial coordinates, providing highly accurate data for subsequent AI identification and avoiding AI relying solely on features while ignoring spatial correlations. In this example, the identified tree map represents an image of the trees that have been identified in the target forest; In this example, the tree sample refers to a sample generated using AI to represent the appearance characteristics of a tree type.
[0032] The working principle and beneficial effects of the above technical solution are as follows: First, tree features are extracted from a single tree information set. Iterative registration ensures accurate matching between the abstract features and spatial appearance of each tree, thereby generating a preliminary tree distribution map. Then, an AI model is used to identify the tree types in the preliminary distribution map, and a dual-dimensional tree sample (features and appearance) is constructed for each type. The preliminary distribution map is then generated by iterating through the samples, ensuring high consistency in features and appearance among trees of the same type, avoiding distribution range deviations caused by inconsistent classification standards. Furthermore, all preliminary distribution maps are overlaid and mapped. Simultaneously, an overall tree concept map is constructed based on measured data, and the identified trees are compared... The tree map and the overall tree concept map are used to verify the completeness of type identification. If they match, the distribution range of each type is directly determined. Otherwise, tree samples are used to traverse different information in the image. Finally, the updated distribution map and the preliminary distribution map are combined to determine the precise distribution range of each type. In this way, feature and appearance registration ensures that the spatial location of each tree is bound to the type features. AI quickly classifies and generates a preliminary distribution map. The concept map is then used to verify and supplement the identification of hidden areas of maple saplings. Finally, the precise distribution range of the three tree species is determined, providing key data for subsequent analysis of the growth trends of different tree species in mountains and valleys. This improves the scientificity and practicality of tree diversity monitoring from the core link.
[0033] Example 5: Based on Example 4, the method for monitoring tree diversity in large-area forests, step 24 includes: Step 241: When the identified tree map and the overall tree concept map are inconsistent, draw the outline of the unidentified area of the target forest according to the different information of the image, and use each tree sample to traverse the outline of the unidentified area to obtain the updated identification information corresponding to each unidentified area outline. Step 242: Update the preliminary tree distribution map using the updated identification information, and record several update information of the preliminary tree distribution map. Match each update information with several corresponding identified tree maps, and update each identified tree map according to the matching results to obtain the corresponding updated distribution map. Step 243: Based on the identified tree map and updated distribution map corresponding to each tree type, redraw the tree distribution of the tree type in the target forest to obtain the distribution range of each tree type in the target forest.
[0034] The working principle and beneficial effects of the above technical solution are as follows: To further achieve accurate completion and reliable optimization of tree type distribution, firstly, based on the differences between the identified tree images and the overall tree concept map, the outlines of unidentified areas are accurately drawn using different information from the images. Then, the outlines of these areas are traversed one by one using the constructed tree samples. The outlines of unidentified areas are accurately locked through image differences, avoiding repeated checks on identified areas and significantly improving the efficiency of supplementary identification. Then, the initial tree distribution map is corrected using updated identification information. At the same time, the updated information is recorded in detail, and each update information is combined with the corresponding identified tree image to generate an updated distribution map. This avoids data contradictions where the distribution map labels new types but the identified images are not updated. Finally, for each tree type, the original identified tree image and the updated distribution map are spatially integrated to determine the precise distribution range of each tree type. By integrating the original and updated data in this way, it is ensured that the distribution range includes both the initially identified core area and the supplementary identification edge and hidden areas, ensuring the continuity of the distribution range and providing a foundation for subsequent work.
[0035] Example 6: Based on Example 1, the method for monitoring tree diversity in large-area forests, step 3 includes: Step 31: Deploy several monitoring devices in each distribution range using drones, construct monitoring environment information corresponding to the distribution range based on the monitoring range corresponding to each monitoring device, and simultaneously acquire climate information of the environment where the target forest is located to construct climate-environment characteristics corresponding to each distribution range. Step 32: Construct a simulated environment using the climate-environment characteristics, simulate the growth trend corresponding to the tree type in the simulated environment, and identify the positive growth time period and negative growth time period corresponding to the tree type in the growth trend; Step 33: Determine the current vegetation coverage of the target forest based on the measured data, deduce the growth data in the corresponding positive growth period and the decay data in the corresponding negative growth period based on the growth trend of each tree type, and use the growth data and decay data to perform dynamic analysis on the current vegetation coverage to obtain the vegetation coverage of the target forest in several predicted time periods.
[0036] In this example, climate-environment characteristics represent the combined characteristics of a distribution range, enhanced by its climate and environmental features. In this example, the monitoring device refers to a device used for temperature monitoring, humidity monitoring, and light monitoring; In this example, a positive growth period indicates that the tree shows a growth trend during that period, such as spring and summer, while a negative growth period indicates that the tree shows a decline trend during that period, such as autumn and winter. In this example, the current vegetation cover rate represents the current vegetation cover rate of the target forest; In this example, dynamic analysis refers to the process of adjusting the current vegetation cover value based on growth and decay data.
[0037] The working principle and beneficial effects of the above technical solution are as follows: First, monitoring devices are deployed in a targeted manner to monitor the distribution range of tree types, avoiding the shortcomings of traditional manual random deployment which makes it difficult to cover remote areas and causes a disconnect between tree species distribution and monitoring range. Simultaneously, local monitoring environmental information and global climate data of the target forest are integrated to construct climate-environment characteristics. Then, a simulated environment that closely matches real growth conditions is built to simulate the growth trend of corresponding tree types. From the trend, positive growth time periods and negative growth time periods are extracted to accurately capture the sensitive growth thresholds of rare tree species, providing about one month of active intervention time for tree protection and effectively reducing vegetation damage rate. Finally, based on measured data, the current vegetation coverage rate is determined. Combined with the growth trend, the growth data of different tree types during the positive period and the decay data during the negative period are derived. Finally, the current coverage rate is dynamically adjusted through data coupling to obtain the vegetation coverage rate for multiple predicted time periods, providing key technical support for the dynamic monitoring of tree diversity in large-area forests, ecological risk prevention and control, and scientific management.
[0038] Example 7: Based on Example 1, the method for monitoring tree diversity in large-area forests, step 4 includes: Step 41: Based on the vegetation coverage corresponding to different prediction time periods and the growth trend of each tree species in the target forest, determine the tree increase area and tree decrease area corresponding to each prediction time period, and determine the increase reason for each tree increase area and the decrease reason for each tree decrease area respectively. Step 42: Based on each of the reasons for increase, reverse deduce several first regional environmental characteristics of the target forest corresponding to different prediction time periods, and based on each of the reasons for decrease, reverse deduce several corresponding second environmental characteristics of the target forest corresponding to different prediction time periods. Step 43: Construct the environmental feature changes of the target forest based on the first environmental feature and the second environmental feature corresponding to different prediction time periods, and obtain the trees that can grow and do not match the environmental feature changes. Step 44: Analyze the growth probability of each of the growable trees based on the measured data, and analyze the growth inhibition probability of each of the existing trees due to environmental changes, determine the increase and decrease of tree species in the target forest during different predicted time periods, generate a tree diversity report and display it.
[0039] In this example, the first regional environmental characteristics represent the environmental characteristics of the area with increased trees, and the second regional environmental characteristics represent the environmental characteristics of the area with decreased trees. In this example, trees that can grow are those that are not affected by changes in environmental characteristics; In this example, the growth inhibition probability represents the probability that environmental changes will inhibit tree growth.
[0040] The working principle and beneficial effects of the above technical solution are as follows: In order to achieve in-depth analysis and scientific prediction of the relationship between tree diversity changes and the environment, the system first combines vegetation cover and tree growth trends at different prediction time periods to accurately divide areas of increased and decreased tree populations. Then, it traces the specific reasons for these two types of areas to ensure that subsequent analyses do not deviate from the core influencing factors, providing precise regional improvement basis for diversity protection. Next, it infers the environmental characteristics of the first area based on the reasons for increased populations and infers the environmental characteristics of the second area based on the reasons for decreased populations. It further integrates the two types of characteristics to construct the environmental characteristic changes of the target forest at different prediction time periods. Finally, it calculates the growth probability of cultivable trees and the inhibition probability of existing trees based on measured data, ultimately determining the types of trees to be added or removed and generating a report. In this way, it solves the problem of traditional predictions relying on experience to recommend tree species without data support, providing managers with scientific and effective data.
[0041] Example 8: Based on Example 7, the method for monitoring tree diversity in large-area forests further includes: When the reduction ratio of the tree species is higher than the prescribed ratio, it is determined that the target forest has been damaged, and the corresponding distribution range is monitored in a key area and a real-time report is generated for display.
[0042] In this example, the specified ratio is 1%.
[0043] The working principle and beneficial effects of the above technical solution are as follows: it enables focused monitoring of tree species whose numbers are rapidly decreasing, helping managers to make correct management decisions.
[0044] Example 9: This example provides a tree diversity monitoring system for large-area forests, such as... Figure 2 As shown, it includes: The information collection module is used to conduct on-site measurements of the target forest to obtain several sets of measured data, identify the information of several trees contained in each set of measured data, and construct a single tree information set of the target forest. The distribution analysis module is used to perform feature recognition on each of the single tree information sets, determine the several tree types contained in the target forest, and determine the distribution range of each tree type in the target forest. The trend analysis module is used to perform environmental monitoring on each of the distribution ranges, deduce the growth trend of each tree species in the target forest, and determine the vegetation coverage of the target forest at different predicted time periods. The report generation module is used to analyze the increase and decrease of tree species corresponding to each predicted time period based on the vegetation coverage, generate a tree diversity report, and display it.
[0045] In this example, the measured data represents the data obtained through laser scanning, which includes information such as tree diameter at breast height, tree height, crown width, and tree species morphology; In this example, the single tree information set contains all the appearance information of a tree; In this example, tree type refers to the various types of trees contained in the target forest; In this example, the distribution range refers to the range corresponding to the distribution of the same tree type in the target forest; In this example, the prediction period is one year. In this example, there are two scenarios for adding tree species: the first is that a species that did not exist before appears, and the second is that the number of existing species increases. In this example, reducing the number of tree species means that the number of existing species has been reduced.
[0046] The working principle and beneficial effects of the above technical solution are as follows: To accurately adapt to the monitoring needs of large-area forests, firstly, comprehensive on-site measurements are conducted on the target forest to obtain measured data including tree diameter at breast height (DBH), tree height, crown width, and tree species morphology. Then, the independent information of each tree is accurately identified to construct a single-tree information set. Next, feature recognition is performed on the single-tree information set to accurately determine the tree type, and the distribution range of each tree type is clarified through spatial positioning data. This technical method can shorten the data collection time to several weeks, avoiding misidentification of tree species due to insufficient experience in manual identification. Furthermore, targeted environmental monitoring is conducted based on the distribution range of each tree type, and growth trends are deduced by combining tree growth patterns, thereby predicting... By dynamically predicting vegetation cover and tree species change trends at different time periods, a scientific basis for proactive management of forest ecosystems can be provided. Finally, based on the vegetation cover at different predicted time periods, the potential for increasing or decreasing tree species can be accurately analyzed, generating a tree diversity report that includes the current diversity status, tree species distribution patterns, future change trends, and protection and intervention recommendations for managers. This approach can provide core data support for forestry departments to formulate protection policies and ecological restoration plans, and also provide basic data for researchers to conduct forest ecological research, effectively improving the scientific rigor and practicality of forest diversity monitoring, and providing reliable technical support for forest resource protection and ecological balance maintenance.
[0047] Example 10: Based on Example 9, the tree diversity monitoring system for large-area forests, wherein the information acquisition module includes: The scanning execution unit is used to generate a UAV flight trajectory based on the forest area characteristics of the target forest, control the UAV to perform radar scanning on the target forest according to the UAV flight trajectory, obtain several sets of measured data of the target forest, and arrange the measured data according to the scanning coverage area corresponding to each set of measured data. The tree identification unit is used to identify several trees in the target forest by performing tree identification on the data arrangement results, and to obtain the tree appearance features corresponding to each identified tree, and to map each tree appearance feature to the corresponding measured data to obtain the unidentified data contained in each measured data. The range analysis unit is used to preprocess each of the unidentified data to determine the data coverage range corresponding to each of the unidentified data, construct the identified range of the target forest based on the tree appearance characteristics corresponding to each identified tree, and determine the range correlation features between each of the data coverage ranges and the identified ranges. The information confirmation unit is used to draw several possible presentation contours of the unidentified data according to the data presentation characteristics of each unidentified data in the corresponding data coverage area, filter the matching possible presentation contours corresponding to each unidentified data based on the corresponding range-related features, update the identified range to obtain several tree information, and construct a single tree information set.
[0048] In this example, forest footprint characteristics refer to the forest's area, topography, and elevation. In this example, the drone flight path represents the flight trajectory of the drone when it is collecting data; In this example, identifying trees means recognizing the trees in the data arrangement results; In this example, tree appearance features refer to the characteristics presented by the tree's appearance; In this example, unidentified data refers to data in the measured data that was not identified; In this example, the data coverage area represents the range corresponding to the unidentified data; In this example, preprocessing includes: denoising, enhancement, and coordinate calibration; In this example, the range-related feature represents the positional relationship between the data coverage area and the identified area; In this example, the apparent outline represents an outline constructed based on unidentified data; In this example, matching suspected rendering contours refers to suspected rendering contours that match unidentified data.
[0049] The working principle and beneficial effects of the above technical solution are as follows: To achieve efficient collection and accurate database construction of information on trees in large-area forests, a customized UAV flight trajectory is first generated based on the land area characteristics of the target forest. Then, the UAV is controlled to acquire measured data through radar scanning, and the measured data is arranged in an orderly manner according to the scanning coverage area, forming a structured dataset from massive amounts of measured data. This avoids inefficiency caused by data clutter in subsequent analysis, providing comprehensive and orderly basic data for accurate tree identification. Next, tree identification is performed on the orderly arranged measured data to identify the trees and extract their appearance features. These appearance features are then mapped back into the measured data to accurately locate unidentified data. Finally, the unidentified data is preprocessed to clarify its data coverage area. The system first identifies the distribution range of trees and then analyzes the range-related characteristics of the two to provide spatial constraints for the subsequent discovery of unidentified trees. Finally, for the unidentified data, several possible outlines are drawn based on their data presentation characteristics. These outlines are then selected and matched based on the range-related characteristics of the previous steps. The identified range is then updated, and all tree information is integrated to construct a single tree information set. In this way, full-area coverage can be achieved through customized drone trajectories. By identifying rare broad-leaved trees hidden in the valley through layered identification and eliminating false rock identifications through outline matching, the final complete information set provides a solid data guarantee for subsequent tree species classification, distribution range analysis, and growth trend prediction, thereby improving the scientificity and reliability of the entire monitoring method from the source.
[0050] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.
Claims
1. A method for tree diversity monitoring for large areas of forest, characterized in that, The method comprises the following steps: Step 1: obtaining a plurality of groups of measured data by on-site measurement of a target forest, identifying a plurality of tree information contained in each of the groups of measured data, and constructing a single tree information set of the target forest; Step 2: identifying the characteristics of each of the single tree information sets, determining a plurality of tree types contained in the target forest, and determining the distribution range of each of the tree types in the target forest; Step 3: monitoring the environment of each of the distribution ranges, deriving the growth trend of each of the tree types in the target forest, and determining the vegetation coverage of the target forest corresponding to different prediction time periods; Step 4: analyzing the increase and decrease of tree species corresponding to each of the prediction time periods according to the vegetation coverage, generating a tree diversity report, and displaying the tree diversity report.
2. A method for monitoring tree diversity in large areas of forest according to claim 1, characterized in that, The step 1 comprises the following steps: Step 11: generating a UAV flight track according to the forest area characteristics of the target forest, controlling a UAV to perform radar scanning on the target forest according to the UAV flight track, obtaining a plurality of groups of measured data of the target forest, and arranging the measured data according to the scanning coverage range corresponding to each of the groups of measured data; Step 12: identifying a plurality of identified trees in the target forest by tree identification on the data arrangement result, respectively acquiring tree appearance characteristics corresponding to each of the identified trees, respectively mapping each of the tree appearance characteristics in the corresponding measured data, and obtaining un-identified data contained in each of the measured data; Step 13: respectively pre-processing each of the un-identified data to determine a data coverage range corresponding to each of the un-identified data, constructing an identified range of the target forest according to the tree appearance characteristics corresponding to each of the identified trees, and determining a range-related feature between each of the data coverage ranges and the identified range; Step 14: drawing a plurality of suspected presentation contours of the un-identified data according to the data presentation features of each of the un-identified data in the corresponding data coverage range, filtering a matching suspected presentation contour corresponding to each of the un-identified data based on the corresponding range-related feature, updating the identified range to obtain a plurality of tree information, and constructing a single tree information set.
3. A method for monitoring tree diversity in large areas of forest according to claim 2, characterized in that, The process of filtering a matching suspected presentation contour corresponding to each of the un-identified data based on the corresponding range-related feature, updating the identified range to obtain a plurality of tree information, and constructing a single tree information set comprises the following steps: determining a plurality of contour connection positions between the data coverage range and the corresponding identified range according to the range-related feature, and a connection relationship corresponding to each of the contour connection positions; locating a contour presentation position corresponding to each of the contour connection positions in each of the suspected presentation contours, and identifying a matching degree between each of the contour presentation positions and the identified range according to the connection relationship; regarding the suspected presentation contour with the highest matching degree sum as the matching suspected presentation contour corresponding to the un-identified data. According to the connection relationship, the identified range is expanded by each matching suspected presentation profile to obtain the total identified range of the target forest; All the identified information contained in the total identified range is divided into several tree information; The tree information of the same type is summarized to obtain a single tree information set of the target forest.
4. A method for monitoring tree diversity in large areas of forest as claimed in claim 1, wherein, The step 2 comprises: Step 21: respectively identify several tree characteristics contained in each single tree information set, respectively draw the tree appearance corresponding to each single tree information, iteratively space register the tree characteristics and the tree characteristics, and generate a preliminary tree distribution map of the target forest; Step 22: identify several tree types contained in the preliminary tree distribution map by using AI, and construct a tree sample corresponding to each tree type; each tree sample is used to traverse the preliminary tree distribution map to obtain a preliminary distribution map of each tree type in the target forest; Step 23: all the preliminary distribution maps are overlapped and mapped to obtain an identified tree map of the target forest, and an overall tree concept map of the target forest is constructed according to the measured data; when the identified tree map and the overall tree concept map are consistent, the distribution range of the corresponding tree type in the target forest is determined according to the preliminary distribution map; Step 24: otherwise, each tree sample is used to traverse the image different information between the identified tree map and the overall tree concept map, and the preliminary tree distribution map is updated according to the traversal result to obtain an updated distribution map corresponding to each tree type, and the distribution range of the tree type in the target forest is determined in combination with the corresponding preliminary distribution map.
5. A method for monitoring tree diversity in large areas of forest according to claim 4, characterized in that, The step 24 comprises: Step 241: when the identified tree map and the overall tree concept map are inconsistent, the un-identified area contour of the target forest is drawn according to the image different information, and each tree sample is used to traverse the un-identified area contour to obtain updated identification information corresponding to each un-identified area contour; Step 242: the preliminary tree distribution map is updated by using the updated identification information, and several updated information of the preliminary tree distribution map are recorded; each updated information is matched with corresponding several identified tree maps, and each identified tree map is updated according to the matching result to obtain a corresponding updated distribution map; Step 243: the tree distribution of each tree type in the target forest is redrawn according to the identified tree map and the updated distribution map corresponding to each tree type to obtain the distribution range of each tree type in the target forest.
6. A method for monitoring tree diversity in large areas of forest as claimed in claim 1, wherein, The step 3 comprises: Step 31: several monitoring devices are placed in each distribution range by using a drone, a monitoring environment information of the corresponding distribution range is constructed based on the monitoring range corresponding to each monitoring device, climate information of the environment where the target forest is located is obtained, and a climate-environment feature corresponding to each distribution range is constructed. Step 32: constructing a simulation environment by using the climate-environmental features, simulating a growth trend corresponding to the tree type in the simulation environment, and identifying a positive growth period and a reverse growth period corresponding to the tree type in the growth trend; Step 33: determining a current vegetation coverage of the target forest according to the measured data, deriving growth data in the positive growth period and decay data in the reverse growth period according to the growth trend of each tree type, dynamically analyzing the current vegetation coverage by using the growth data and the decay data, and obtaining vegetation coverage corresponding to the target forest in several prediction periods.
7. A method for monitoring tree diversity in large areas of forest as claimed in claim 1, wherein, The step 4 comprises: Step 41: determining a tree increase area and a tree decrease area corresponding to each prediction period according to the vegetation coverage corresponding to different prediction periods and the growth trend of each tree type in the target forest, and respectively determining an increase reason corresponding to each tree increase area and a decrease reason corresponding to each tree decrease area; Step 42: reversely deriving several first area environmental features corresponding to the target forest in different prediction periods according to each increase reason, and reversely deriving several second environmental features corresponding to the target forest in different prediction periods according to each decrease reason; Step 43: constructing an environmental feature change of the target forest according to the first environmental features and the second environmental features corresponding to different prediction periods, obtaining growable trees matching the environmental feature change, and obtaining existing growing trees not matching the environmental change; Step 44: analyzing a growth probability of each growable tree according to the measured data, and analyzing a growth inhibition probability of each existing growing tree by the environmental change, determining an increase tree type and a decrease tree type corresponding to the target forest in different prediction periods, generating a tree diversity report and displaying the tree diversity report.
8. A method for monitoring tree diversity in large areas of forest according to claim 7, characterized in that, Further comprising: When a decrease proportion corresponding to the decrease tree type is higher than a specified proportion, it is determined that the target forest is damaged, a distribution range corresponding to the decrease tree type is monitored, and a real-time report is generated and displayed.
9. A tree diversity monitoring system for large areas of forest, characterized by Comprise: An information collection module is configured to obtain several groups of measured data by performing on-site measurement on a target forest, identify several tree information included in each group of measured data, and construct a single tree information set of the target forest; A distribution analysis module is configured to identify features of each single tree information set, determine several tree types included in the target forest, and determine a distribution range of each tree type in the target forest; A trend analysis module is configured to monitor an environment of each distribution range, derive a growth trend of each tree type in the target forest, and determine a vegetation coverage corresponding to the target forest in different prediction periods; A report generation module is configured to analyze an increase tree type and a decrease tree type corresponding to each prediction period according to the vegetation coverage, generate a tree diversity report, and display the tree diversity report.
10. A tree diversity monitoring system for large areas of forest as claimed in claim 9, wherein, The information collection module comprises: The scanning execution unit is configured to generate a UAV flight track according to the forest land feature of the target forest, control the UAV to perform radar scanning on the target forest according to the UAV flight track, obtain a plurality of groups of measured data of the target forest, and arrange the measured data according to the scanning coverage range corresponding to each group of measured data; The tree identification unit is configured to perform tree identification on the data arrangement result, determine a plurality of identified trees in the target forest, and obtain the tree appearance feature corresponding to each identified tree, respectively, map each tree appearance feature in the corresponding measured data, and obtain the unidentified data contained in each measured data; The range analysis unit is configured to respectively pre-process each unidentified data, determine the data coverage range corresponding to each unidentified data, construct the identified range of the target forest according to the tree appearance feature corresponding to each identified tree, and determine the range-related feature between each data coverage range and the identified range; The information confirmation unit is configured to draw a plurality of suspected presentation contours of the unidentified data according to the data presentation feature of each unidentified data in the corresponding data coverage range, filter the matching suspected presentation contour corresponding to each unidentified data based on the corresponding range-related feature, update the identified range to obtain a plurality of tree information, and construct a single-tree information set.