A Smart Assessment Method for Tree Support Stability Based on Multi-Sensor Fusion

By integrating information from multiple sensors, a dynamic evaluation link for the internal support relationships of a tree swarm is constructed, generating a mutual thrust evaluation map. This addresses the risk of implicit structural imbalance in tree swarm stability assessment and improves the stability and monitoring accuracy of the tree swarm support structure.

CN123117773APending Publication Date: 2026-11-17SHENZHEN INCERE ENVIRONMENTAL ART ENG
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611245530.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-17
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately identify changes in the direction of force transmission at binding nodes, changes in root zone support contribution, and the evolution of the mutual push relationship caused by canopy eccentric loading in the overall stability assessment of tree groves. This leads to a discrepancy between the overall stability of the tree grove and the actual stress conditions, making it impossible to detect hidden structural imbalance risks in a timely manner and reducing the accuracy of long-term stability assessments of spliced ​​tree support systems.

Method used

By establishing a growth difference-driven mutual push identification mechanism for multi-tree planting, integrating the force transmission information of binding nodes, tree posture information, and canopy spatial distribution information, a dynamic evaluation link for the internal support relationship of the tree group is constructed, generating a mutual push force evaluation map, and realizing the hierarchical labeling and dynamic callback analysis of the mechanical coupling relationship between multiple trees.

Benefits of technology

It improves the accuracy of stability assessment of the supporting structure of multi-species trees, realizes comprehensive judgment and early warning of tree group structure, and enhances the safety and long-term stability monitoring accuracy of complex landscape tree groups.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117773A_ABST
    Figure CN123117773A_ABST
Patent Text Reader

Abstract

This invention discloses an intelligent assessment method for tree support stability based on multi-sensor fusion, belonging to the field of support assessment technology. It addresses the problem of discrepancies between the overall stability of a tree grove and its actual stress conditions, leading to the inability to promptly detect hidden structural imbalance risks and reducing the accuracy of long-term stability assessments for multi-tree support systems. By collecting data on the radial thickening of tree trunks during the monitoring period, a growth difference spectrum among trees is constructed and cross-propulsion risk states are identified. Combined with humidity and tension sensor data from shared binding nodes, a tension and humidity correction baseline is established to obtain the net conducted tension at the nodes. Furthermore, by integrating trunk inclination angle, root zone soil pressure distribution, and canopy offset characteristics, the mechanical coupling relationships among multiple trees are graded, labeled, and dynamically analyzed to generate a cross-propulsion force assessment map. This enables comprehensive judgment and early warning of the stability of multi-tree support structures, improving the safety and long-term stability monitoring accuracy of complex landscape tree groves.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of tree support assessment technology, and more specifically, to an intelligent assessment method for tree support stability based on multi-sensor fusion. Background Technology

[0002] In the process of urban landscaping, landscape enhancement, and ecological construction of large public spaces, in order to quickly create a landscape effect with rich layers and visual integrity, it is common to use the method of planting multiple trees together to construct a combined tree group. In the early stage of planting, such combined tree groups usually need to maintain overall stability through measures such as binding and support, and joint fixation, so that multiple trees can form a collaborative support structure within a certain period of time to resist the risk of toppling caused by wind load, soil disturbance, and unbalanced growth.

[0003] The existing technology has the following shortcomings: Currently, most existing technologies rely on independent assessments based on the tilt state of individual trees, the condition of supporting components, or the local soil environment. This makes it difficult to accurately identify changes in the direction of force transmission at binding nodes, changes in root zone support contribution, and the evolution of the mutual push relationship caused by canopy eccentricity during the continuous accumulation of growth differences between trees as they grow. This leads to a discrepancy between the overall stability of the tree group and the actual stress situation, making it impossible to detect hidden structural imbalance risks in a timely manner and reducing the accuracy of long-term stability assessment of the spliced ​​tree support system. Therefore, this paper proposes an intelligent assessment method for tree support stability based on multi-sensor fusion.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an intelligent assessment method for tree support stability based on multi-sensor fusion. By establishing a growth difference-driven mutual push identification mechanism for intercropped trees, the method integrates and correlates the force transmission information of binding nodes, tree posture information, root zone pressure information, and canopy spatial distribution information to construct a dynamic assessment link for the support relationship within the tree group. Furthermore, it generates a mutual push force assessment map based on a multi-dimensional feature callback mechanism to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent assessment method for tree support stability based on multi-sensor fusion, comprising the following steps: Step S1: Set a monitoring period, collect the radial thickening of each tree in the clump of trees to be tested within the monitoring period, generate a growth difference spectrum between trees based on the radial thickening, and determine whether to enter the mutual push identification mechanism through the growth difference spectrum between trees. Step S2: When entering the mutual push identification mechanism, detect the ambient humidity of the shared binding node and generate a tension humidity correction baseline, retrieve the binding tension of the shared binding node, perform humidity removal on the binding tension according to the tension humidity correction baseline, and calculate the net conduction tension of the node. Step S3: Identify the trunk elevation angle of each tree in the tree clump to be tested, correct the net conduction tension of the nodes according to the trunk elevation angle, collect the soil pressure distribution of the common root zone in the tree clump to be tested, and classify and mark each tree in the tree clump to be tested in combination with the corrected net conduction tension of the nodes. Step S4: Extract the canopy projection offset of each tree in the clump of trees to be tested, generate canopy offset features based on the canopy projection offset, determine whether to perform a callback to the classification label based on the canopy offset features, and generate a mutual thrust assessment map using the callback results.

[0007] In a preferred embodiment, in step S1, the radial thickening of the main trunk is the diameter increase of each main trunk at a set calibration section during the monitoring period. The radial thickening of the main trunk of multiple plants was statistically analyzed to obtain the mean and standard deviation of the radial thickening of the main trunk. The standard deviation of the radial thickening of the main stem is normalized by using the mean and standard deviation of the radial thickening of the main stem to obtain the growth difference value between plants.

[0008] In a preferred embodiment, in step S1, the inter-plant growth difference spectrum includes the direction and intensity of the difference between each plant. Wherein, the difference intensity is the absolute value of the difference in growth between plants, and the difference direction is the sign direction of the difference in growth between plants. The maximum value of the difference intensity of each plant in the inter-plant growth difference spectrum is compared with the cross-inference trigger threshold: When the maximum value of the difference intensity in the inter-plant growth difference spectrum is greater than or equal to the mutual inference trigger threshold, the mutual inference recognition mechanism is entered. Conversely, it will not enter the mutual push identification mechanism.

[0009] In a preferred embodiment, in step S2, the shared binding node is the mechanical coupling point formed by binding multiple main trunks of trees in the grove of trees to be tested with the same hemp rope at the intersection. The environmental humidity and binding tension of the shared binding node are detected. The tension humidity correction baseline is calculated using linear regression based on the ambient humidity of the shared binding node. Humidity rejection is performed on the retrieved binding tension, and the net conduction tension of the node is obtained by subtracting the binding tension from the humidity correction baseline.

[0010] In a preferred embodiment, in step S3, the trunk elevation tilt angle refers to the deviation angle of the upper end of each trunk from its initial vertical position relative to the initial installation position, including the direction vector; For each shared binding node, retrieve the trunk high-position tilt angle vector of the two adjacent plants, and calculate the tilt angle and deviation angle of the two plants based on the trunk high-position tilt angle vector; Compare the principal tilt angles and deflection angles of two adjacent plants. The plant with the smaller principal tilt angle and deflection angle is marked as the initiator; the plant with the larger principal tilt angle and deflection angle is marked as the receiver. The net conduction tension at the node is corrected by adjusting the direction vector from the actuator to the receiver to obtain the corrected net conduction tension at the node.

[0011] In a preferred embodiment, in step S3, the pressure values ​​of the four symmetrical directions of the common root domain are collected and the earth pressure distribution of the common root domain is constructed. Based on the earth pressure distribution in the common root domain, the difference in earth pressure components in the symmetrical direction is calculated separately. The difference in earth pressure components in the two axes is combined into an unbalanced vector of earth pressure distribution in the common root domain. The amplitude of the unbalanced vector is used as the unbalanced amplitude. The rules for classifying and labeling each tree in the grove of trees to be tested are as follows: the tree marked as the initiator, whose corrected net conduction tension amplitude is greater than the average net conduction tension amplitude, and whose imbalance vector points to the location of the initiator, is marked as the dominant tree in mutual push; the tree marked as the initiator, whose corrected net conduction tension amplitude is greater than the average net conduction tension amplitude, and whose imbalance vector points from the location of the initiator to the location of the corresponding tree, is marked as the initiated tree in mutual push. The remaining plants were marked as neutral plants.

[0012] In a preferred embodiment, in step S4, the canopy projection offset refers to the horizontal offset vector of the center of the canopy projection on the horizontal ground relative to the initial projection position. The canopy projection offset includes two components: offset direction and offset magnitude. The canopy offset features are generated by combining the offset direction and offset amplitude of the canopy projection offset of each plant. The canopy offset features include the offset direction, offset amplitude and offset duality of the canopy of each plant. The duality of canopy offset characteristics is defined as follows: the angle between the offset direction of the canopy projection of the mutually pushed plant and the direction of the line connecting the location of the corresponding dominant plant is greater than 90 degrees, and the amplitude of the offset of the canopy projection of the mutually pushed plant is greater than the average amplitude of the offset of the canopy projection of all plants in the clump of trees to be tested.

[0013] In a preferred embodiment, in step S4, the canopy offset features are compared with the classification markers for consistency. Condition one is that the angle between the offset direction of the reciprocating plant and the direction of the line connecting the location of the corresponding dominant reciprocating plant is greater than 90 degrees; condition two is that the offset amplitude of the reciprocating plant is greater than the average of all offset amplitudes. If both conditions one and two are met, the callback will not be executed. When either condition one or condition two is not met, a callback is executed on the classification label based on the offset direction: calculate the azimuth angle between each plant and the opposite offset direction of the pushed plant, and re-label the plant with the smallest azimuth angle. If the re-comparison still fails to find a match, and the cumulative number of callbacks reaches the maximum number of callbacks, the callback will be terminated, and a review prompt will be output.

[0014] In a preferred embodiment, in step S4, a mutual push force evaluation map is generated based on the callback results: using the top view of the tree clump to be tested as the base map, the position, role (dominant tree, receptive tree, or neutral tree), mutual push direction vector, and mutual push intensity level of each tree are marked on the base map. The mutual push strength level is determined by the uneven distribution of earth pressure in the shared root zone and is divided into three levels: high, medium, and low.

[0015] The technical effects and advantages of this invention are as follows: This invention collects the radial thickening of tree trunks during the monitoring period, constructs a spectrum of inter-tree growth differences, identifies mutual push risk states, and establishes a tension and humidity correction baseline by combining humidity and tension sensor data of shared binding nodes to obtain the net conducted tension of the nodes. Furthermore, it integrates trunk inclination angle, root zone soil pressure distribution, and canopy offset characteristics to classify, label, and dynamically analyze the mechanical coupling relationship between multiple trees, generating a mutual push force assessment spectrum. This enables a comprehensive judgment and early warning of the stability of the support structure of multi-tree plantings, improving the safety and long-term stability monitoring accuracy of complex landscape tree group structures. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the implementation of an intelligent assessment method for tree support stability based on multi-sensor fusion, as described in this invention.

[0017] Figure 2 This is a schematic diagram illustrating the steps of an intelligent assessment method for tree support stability based on multi-sensor fusion according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention collects the radial thickening of tree trunks during the monitoring period, constructs a spectrum of inter-tree growth differences, identifies the risk status of mutual push, and establishes a tension and humidity correction baseline by combining humidity and tension sensor data of shared binding nodes to obtain the net conducted tension of the nodes. Furthermore, it integrates the trunk inclination angle, root zone soil pressure distribution, and canopy offset characteristics to classify, label, and dynamically analyze the mechanical coupling relationship between multiple trees, generating a mutual push force assessment spectrum, and realizing a comprehensive judgment and early warning of the stability of the support structure of multi-tree trees.

[0020] Example 1, as Figures 1 to 2 As shown, a smart assessment method for tree support stability based on multi-sensor fusion is described, and the specific operation process is as follows: Step S1: Set a monitoring period, collect the radial thickening of each tree in the clump of trees to be tested within the monitoring period, generate a growth difference spectrum between trees based on the radial thickening, and determine whether to enter the mutual push identification mechanism through the growth difference spectrum between trees. Step S2: When entering the mutual push identification mechanism, detect the ambient humidity of the shared binding node and generate a tension humidity correction baseline, retrieve the binding tension of the shared binding node, perform humidity removal on the binding tension according to the tension humidity correction baseline, and calculate the net conduction tension of the node. Step S3: Identify the trunk elevation angle of each tree in the tree clump to be tested, correct the net conduction tension of the nodes according to the trunk elevation angle, collect the soil pressure distribution of the common root zone in the tree clump to be tested, and classify and mark each tree in the tree clump to be tested in combination with the corrected net conduction tension of the nodes. Step S4: Extract the canopy projection offset of each tree in the clump of trees to be tested, generate canopy offset features based on the canopy projection offset, determine whether to perform a callback to the classification label based on the canopy offset features, and generate a mutual thrust assessment map using the callback results.

[0021] The specific implementation is as follows: In step S1, a monitoring period is set, and the radial thickening of the main trunk of each tree in the grove of trees to be tested is collected within the monitoring period.

[0022] The tree clump to be tested refers to an overall landscape unit formed by several trees through the crossing of main trunks, close proximity of root balls, shared binding and compaction of backfill material between root balls.

[0023] The radial thickening of the trunk refers to the diameter increase of each trunk at a set calibration section during the monitoring period. In this example, it is collected by strain sensors arranged circumferentially along the trunk.

[0024] It should be added that the monitoring cycle is set according to the growth rate characteristics of the planted trees, for example, the quarterly scale is used for fast-growing trees; the strain sensors are deployed in the circumferential direction of the trunk to realize continuous reading of the trunk diameter change, or the diameter geometric change can be obtained by scanning the trunk cross section with an image scanning device. Those skilled in the art can choose according to the site conditions.

[0025] The radial thickening of the main stem of each plant was collected during each monitoring period. The radial thickening of the main stems of multiple plants was statistically analyzed to obtain the mean and standard deviation of the radial thickening of the main stems.

[0026] For each plant, calculate the deviation between its radial thickening of the main stem and the mean within the clump, and normalize it using the standard deviation. The calculation formula is as follows: ; in, Let be the inter-plant growth difference value of the i-th plant. The radial thickening of the main stem of the i-th plant is given. This is the average radial thickening of the main trunk. This is the standard deviation of the radial thickening of the main trunk.

[0027] The inter-plant growth difference spectrum is formed by the set of inter-plant growth difference values ​​of all plants. It records the direction and intensity of the difference for each plant. The direction of difference is the sign direction of the inter-plant growth difference value. A positive direction of difference indicates that the growth rate of the plant is higher than the average within the cluster, and a negative direction of difference indicates that the growth rate of the plant is lower than the average within the cluster. The intensity of difference is the absolute value of the inter-plant growth difference value.

[0028] The maximum value of the difference intensity of each plant in the inter-plant growth difference spectrum is compared with the cross-inference trigger threshold: When the maximum value of the difference intensity in the inter-plant growth difference spectrum is greater than or equal to the mutual push triggering threshold, it is determined that there is an inconsistent growth direction among the trees in the test clump and the mutual push identification mechanism is entered. When the maximum value of the difference intensity in the inter-plant growth difference spectrum is less than the cross-induction trigger threshold, it is determined that the growth of each plant in the test tree clump is basically the same, and the cross-induction identification mechanism will not be entered in this cycle, maintaining the regular monitoring rhythm.

[0029] The method for setting the cross-pushing trigger threshold is as follows: select several groups of mixed-species samples that have already experienced cross-pushing, statistically analyze the distribution range of the maximum difference intensity in the growth difference spectrum between plants, and take the median value of the distribution of the maximum difference intensity as the cross-pushing trigger threshold to balance the early identification sensitivity and false alarm rate.

[0030] In step S2, after entering the mutual push identification mechanism, the ambient humidity of the shared binding node is detected. The ambient humidity of the shared binding node is detected in real time by a miniature humidity probe at the shared binding node, reflecting the air humidity level around the node. It should be added that a shared binding node refers to a mechanical coupling point formed by binding the trunks of multiple trees in a grove of trees to be tested with the same hemp rope at the intersection. A single node simultaneously bears the force transmitted from multiple trees.

[0031] The binding tension is retrieved in real time by embedding a tension sensor in the hemp rope of the shared binding node, reflecting the total tension borne by the hemp rope at the node.

[0032] The tension sensor is embedded in the hemp rope by attaching a thin-film strain gauge to the hemp rope at the node and covering it with a waterproof wrapping layer. The deformation signal of the strain gauge is then converted into the tension data of the hemp rope.

[0033] A tension humidity correction baseline was constructed based on the environmental humidity of the shared binding nodes. A linear regression method was used to fit the spurious tension changes caused by the moisture absorption of the hemp rope. The calculation formula is as follows: ; Where B is the tension-humidity correction baseline. To share the ambient humidity of the binding nodes, , These are the baseline parameters.

[0034] The baseline parameter calibration method is as follows: Select a hemp rope sample with the same specifications as the one used on site and conduct a drying-humidification cycle test. Record the tension change data under different humidity levels and determine the value of the baseline parameter by least squares fitting.

[0035] The higher the ambient humidity of the shared binding node, the greater the false increase in tension caused by the moisture absorption and expansion of the hemp rope, and the larger the corresponding tension humidity correction baseline; the lower the ambient humidity, the smaller the tension humidity correction baseline.

[0036] Humidity rejection is performed on the retrieved binding tension, and the net conduction tension of the node is obtained by subtracting the binding tension from the humidity correction baseline.

[0037] The net conduction tension at the node reflects the lateral force transmitted between multiple plants through shared binding nodes.

[0038] It should be noted that the miniature humidity probe is a miniaturized humidity measurement device based on capacitive or resistive humidity-sensitive elements. It can be attached to the surface of the binding node rope or embedded inside the waterproof wrapping layer of the node. In this example, it is used to detect the air humidity around the shared binding node in real time. The tension sensor is a measurement device that converts tension into an electrical signal based on the deformation of elastic elements and the resistance strain effect.

[0039] In step S3, the high-level tilt angle of the main trunk of each tree in the grove of trees to be tested is identified. The high-level tilt angle of the main trunk refers to the deviation angle of the upper end of the main trunk from the initial vertical position, including the direction vector. It is identified in real time by a dual-axis tilt angle sensor installed at the high position of the main trunk of each tree.

[0040] It should be added that a dual-axis tilt sensor is a sensing device that simultaneously measures the tilt components of two orthogonal axes based on the principle of microelectromechanical systems accelerometers.

[0041] The high position of the trunk is taken at 1 / 3 to 1 / 2 of the tree height from the ground. This position can avoid the interference of soil deformation around the roots and is lower than the main range of the canopy wind load, so it can more accurately reflect the elastic bending displacement of the trunk caused by mutual push. The dual-axis tilt sensor records the east-west and north-south tilt components in a dual-axis manner.

[0042] Perform direction attribution correction on the net conduction tension at each shared binding node: For each shared binding node, retrieve the main trunk high-position tilt angle vector of the two adjacent plants, calculate the angle between the main trunk high-position tilt angle vector and the direction of the line connecting the two plants, and obtain the main tilt angle deviation angle of the corresponding plant. Compare the principal tilt angles and deflection angles of two adjacent plants. The plant with the smaller principal tilt angle and deflection angle is marked as the initiator; the plant with the larger principal tilt angle and deflection angle is marked as the receiver. The net conduction tension at the node is corrected by adjusting the direction vector from the actuator to the receiver to obtain the corrected net conduction tension at the node. The calculation formula is as follows: ; in, The corrected net conduction tension at the node. For the net conductive tension of the node, It represents the directional difference in the high-level inclination angle of two adjacent main trunks.

[0043] The greater the net conduction tension at the node after correction, the more it indicates that there is a definite lateral conduction at the node; its direction vector is the conduction path of the mutual thrust.

[0044] The shared root zone refers to the overall underground area formed by the soil balls of each tree and the backfill material between the soil balls in the clump of trees to be tested. Earth pressure sensors are set up in the backfill area between the soil balls in the shared root zone, and earth pressure sensors are placed in four symmetrical directions (east, south, west, and north) to collect earth pressure values ​​at each location. The set of earth pressure values ​​in the four directions is taken as the earth pressure distribution of the shared root zone.

[0045] It should be added that an earth pressure sensor is a sensing device that converts soil pressure into an electrical signal based on a diaphragm or vibrating wire pressure-sensitive element. In this example, it is buried in the backfill material between soil balls to collect the underground earth pressure values ​​at various locations, and the output results constitute the earth pressure distribution of the common root domain.

[0046] Based on the earth pressure distribution in the common root domain, the difference between the earth pressure components in the east-west axis direction and the difference between the earth pressure components in the north-south axis direction are calculated respectively. The difference between the east-west axis components is the difference between the earth pressure values ​​in the east and west directions, and the difference between the north-south axis components is the difference between the earth pressure values ​​in the south and north directions. The two axial component differences are combined into an unbalanced vector of the earth pressure distribution in the common root domain, and the amplitude of the unbalanced vector is taken as the unbalanced amplitude. The greater the degree of imbalance, the more the underground root system tends to push adjacent soil ball areas in a single direction, and the more concentrated the transmission of mutual thrust underground.

[0047] By combining the direction vector of the corrected net conduction tension at the nodes with the uneven vector of the soil pressure distribution in the shared root zone, each plant within the clump is classified and labeled: When a plant is marked as the initiator, and the corrected net conduction tension amplitude of the corresponding shared binding node is greater than the average net conduction tension of the node, and the uneven vector of the soil pressure distribution in the shared root zone points to the location of the initiator corresponding to the initiator, the plant of the initiator is marked as the dominant plant of mutual push. When a plant is marked as the affected party, and the corrected net conduction tension amplitude of the corresponding shared binding node is greater than the average net conduction tension of the node, and the unbalanced vector of the soil pressure distribution in the shared root zone points from the direction of the applied party corresponding to the affected party to the direction of the affected party, the affected plant is marked as a mutually pushed plant. If the corrected net conduction tension amplitude of the node does not exceed the average net conduction tension of the node, or if the two determination results are inconsistent, the plant is marked as a neutral plant.

[0048] The arithmetic mean of the net conduction tension of all shared binding nodes within the current monitoring period is used as the mean of the net conduction tension of the nodes.

[0049] A dominant plant in a mutual push refers to the source plant within a cluster that exerts lateral pushing on its neighboring plants; a passive plant in a mutual push refers to a plant within a cluster that is passively pushed by its neighboring plants.

[0050] In step S4, the canopy projection offset of each tree in the grove of trees to be tested is extracted. The canopy projection offset refers to the horizontal offset vector of the center of the canopy projection on the horizontal ground relative to the initial projection position. It is obtained periodically by the overhead image acquisition device.

[0051] The overhead image acquisition device can use a drone to hover and shoot at a fixed point, a camera fixed on the top of a tower, or a telephoto device erected at a high point of a building. After the overhead image is transmitted to the image processing unit, the coordinates of the canopy projection center of each tree are extracted by the canopy projection center recognition algorithm and compared with the initial coordinates to obtain the offset vector.

[0052] The canopy projection offset includes two components: offset direction and offset magnitude. The direction reflects the direction of the canopy offset in the horizontal plane, and the magnitude reflects the distance of the canopy offset.

[0053] The canopy offset features are generated by combining the offset direction and offset amplitude of the canopy projection offset of each plant. The canopy offset features include the offset direction and offset amplitude of the canopy of each plant, as well as the offset pairing relationship between the dominant plant and the pushed plant.

[0054] The duality of canopy offset characteristics is defined as follows: the angle between the offset direction of the canopy projection of the mutually pushed plant and the direction of the line connecting the location of the corresponding mutually pushed dominant plant is greater than 90 degrees, and the offset amplitude of the mutually pushed plant is greater than the average of the canopy offset amplitudes of all plants in the tree clump to be tested.

[0055] The consistency of the canopy migration features with the internal cross-referencing relationship expressed by the classification labels output in step S3 is compared: The canopy shift characteristics were compared with the classification markers for consistency. Condition one was that the angle between the shift direction of the reciprocating plant and the direction of the corresponding dominant reciprocating plant was greater than 90 degrees; condition two was that the shift amplitude of the reciprocating plant was greater than the average of all shift amplitudes. When both conditions one and two are met, the canopy offset feature is determined to be consistent with the classification label, and the callback is not executed. If either condition one or condition two is not met, the canopy offset feature is determined to be inconsistent with the classification label. The classification label is then re-adjusted based on the offset direction: the azimuth angle between each plant and the offset in the opposite direction of the pushed plant is calculated, the plant with the smallest azimuth angle is re-labeled as the dominant plant of mutual push, the pushed plant is re-labeled as the pushed plant of mutual push, the remaining plants are re-labeled as neutral plants, and the consistency comparison is re-executed. If the re-comparison still fails to find a match and the cumulative number of callbacks reaches the maximum number of callbacks, the callback will be terminated, a review prompt will be output, and a manual review will be requested.

[0056] The upper limit of the callback count is set by using the median value of the convergence rounds of spurious classification samples in the pre-test. In this example, it is set to 2 times.

[0057] Based on the callback results, a mutual thrust assessment map is generated: using the top view of the tree clump to be tested as the base map, the position, role (dominant tree, receptive tree, or neutral tree), mutual thrust direction vector, and mutual thrust intensity level of each tree are marked on the base map; the mutual thrust intensity level is determined by the uneven distribution of soil pressure in the shared root zone and is divided into three levels: high, medium, and low.

[0058] The mutual thrust assessment map reflects the spatial distribution and intensity grading of mutual thrust within the clump. Based on the map, maintenance personnel can perform suppressive pruning on the dominant trees in the mutual thrust, external reinforcement on the trees affected by the mutual thrust, and maintain the existing maintenance on the neutral trees, thereby maintaining the overall support stability of the clump of trees to be tested.

[0059] Finally, it should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0060] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] In this document, the singular forms “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that terms such as “comprising / including” or “having” specify the presence of the stated features, integrals, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, integrals, steps, operations, components, parts, or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0062] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0063] The above description of the disclosed embodiments will enable those skilled in the art to make or use various modifications to these embodiments. It will be readily apparent to those skilled in the art that the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-sensor fusion-based tree support stability intelligent evaluation method, characterized by: Includes the following steps: Step S1: Set a monitoring period, collect the radial thickening of each tree in the clump of trees to be tested within the monitoring period, generate a growth difference spectrum between trees based on the radial thickening, and determine whether to enter the mutual push identification mechanism through the growth difference spectrum between trees. Step S2: When entering the mutual push identification mechanism, detect the ambient humidity of the shared binding node and generate a tension humidity correction baseline, retrieve the binding tension of the shared binding node, perform humidity removal on the binding tension according to the tension humidity correction baseline, and calculate the net conduction tension of the node. Step S3: Identify the trunk elevation angle of each tree in the tree clump to be tested, correct the net conduction tension of the nodes according to the trunk elevation angle, collect the soil pressure distribution of the common root zone in the tree clump to be tested, and classify and mark each tree in the tree clump to be tested in combination with the corrected net conduction tension of the nodes. Step S4: Extract the canopy projection offset of each tree in the clump of trees to be tested, generate canopy offset features based on the canopy projection offset, determine whether to perform a callback to the classification label based on the canopy offset features, and generate a mutual thrust assessment map using the callback results.

2. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 1, characterized in that: In step S1, the radial thickening of the main trunk is the diameter increase of each main trunk at the set calibration section during the monitoring period; The radial thickening of the main trunk of multiple plants was statistically analyzed to obtain the mean and standard deviation of the radial thickening of the main trunk. The standard deviation of the radial thickening of the main stem is normalized by using the mean and standard deviation of the radial thickening of the main stem to obtain the growth difference value between plants.

3. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 2, characterized in that: In step S1, the inter-plant growth difference spectrum includes the direction and intensity of the difference for each plant. Wherein, the difference intensity is the absolute value of the difference in growth between plants, and the difference direction is the sign direction of the difference in growth between plants. The maximum value of the difference intensity of each plant in the inter-plant growth difference spectrum is compared with the cross-inference trigger threshold: When the maximum value of the difference intensity in the inter-plant growth difference spectrum is greater than or equal to the mutual inference trigger threshold, the mutual inference recognition mechanism is entered. Conversely, it will not enter the mutual push identification mechanism.

4. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 1, characterized in that: In step S2, the shared binding node is the mechanical coupling point formed by binding multiple main trunks of trees in the grove of trees to be tested with the same hemp rope at the intersection. The environmental humidity and binding tension of the shared binding node are detected. The tension humidity correction baseline is calculated using linear regression based on the ambient humidity of the shared binding node. Humidity rejection is performed on the retrieved binding tension, and the net conduction tension of the node is obtained by subtracting the binding tension from the humidity correction baseline.

5. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 1, characterized in that: In step S3, the trunk elevation tilt angle refers to the deviation angle of the upper end of each trunk from its initial vertical position relative to the initial installation position, including the direction vector; For each shared binding node, retrieve the trunk high-position tilt angle vector of the two adjacent plants, and calculate the tilt angle and deviation angle of the two plants based on the trunk high-position tilt angle vector; Compare the principal tilt angles and deflection angles of two adjacent plants. The plant with the smaller principal tilt angle and deflection angle is marked as the initiating plant, and the plant with the larger principal tilt angle and deflection angle is marked as the receiving plant. The net conduction tension at the node is corrected by adjusting the direction vector from the actuator to the receiver to obtain the corrected net conduction tension at the node.

6. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 5, characterized in that: In step S3, the pressure values ​​in the four symmetrical directions of the common root domain are collected and the earth pressure distribution of the common root domain is constructed. Based on the earth pressure distribution in the common root domain, the difference in earth pressure components in the symmetrical direction is calculated separately. The difference in earth pressure components in the two axes is combined into an unbalanced vector of earth pressure distribution in the common root domain. The amplitude of the unbalanced vector is used as the unbalanced amplitude. The rules for classifying and labeling each tree in the grove of trees to be tested are as follows: the tree that is labeled as the initiator and whose corrected net conduction tension amplitude is greater than the average net conduction tension amplitude of the node, and whose imbalance vector points to the location of the initiator, is labeled as the dominant tree in mutual push; the tree that is labeled as the initiator and whose corrected net conduction tension amplitude is greater than the average net conduction tension amplitude of the node, and whose imbalance vector points from the location of the initiator to the location of the corresponding tree, is labeled as the initiated tree in mutual push. The remaining plants were marked as neutral plants.

7. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 6, characterized in that: In step S4, the canopy projection offset refers to the horizontal offset vector of the center of the canopy projection on the horizontal ground relative to the initial projection position. The canopy projection offset includes two components: offset direction and offset magnitude. The canopy offset features are generated by combining the offset direction and offset amplitude of the canopy projection offset of each plant. The canopy offset features include the offset direction, offset amplitude and offset duality of the canopy of each plant. The duality of canopy offset characteristics is defined as follows: the angle between the offset direction of the canopy projection of the mutually pushed plant and the direction of the line connecting the location of the corresponding dominant plant is greater than 90 degrees, and the amplitude of the offset of the canopy projection of the mutually pushed plant is greater than the average amplitude of the offset of the canopy projection of all plants in the clump of trees to be tested.

8. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 7, characterized in that: In step S4, the canopy shift features are compared with the classification markers for consistency. Condition one is that the angle between the shift direction of the reciprocating plant and the direction of the line connecting the location of the corresponding dominant reciprocating plant is greater than 90 degrees; condition two is that the shift amplitude of the reciprocating plant is greater than the average of all shift amplitudes. If both conditions one and two are met, the callback will not be executed. When either condition one or condition two is not met, a callback is executed on the classification label based on the offset direction: calculate the azimuth angle between each plant and the opposite offset direction of the pushed plant, and re-label the plant with the smallest azimuth angle. If the re-comparison still fails to find a match, and the cumulative number of callbacks reaches the maximum number of callbacks, the callback will be terminated, and a review prompt will be output.

9. The intelligent assessment method for tree support stability based on multi-sensor fusion according to claim 8, characterized in that: In step S4, a mutual push force evaluation map is generated based on the callback results: using the top view of the tree clump to be tested as the base map, the position, role (dominant tree, reciprocated tree, or neutral tree), mutual push direction vector, and mutual push intensity level of each tree are marked on the base map. The mutual push strength level is determined by the uneven distribution of earth pressure in the shared root zone and is divided into three levels: high, medium, and low.