Methods, devices and systems for surveying trees in forest farms

Abstract

This application provides a method, apparatus, and system for surveying trees in a forest farm, relating to the field of forest farm monitoring technology. The method includes: acquiring the first tree scan results from a first scan of the forest farm conducted by a drone; drawing a heat map of the distribution of the first tree scan results in the forest farm based on the first tree scan results; identifying areas in the forest farm to be reviewed based on the heat map of the first tree scan results; acquiring the second tree scan results from a second scan of the areas to be reviewed by a robot dog; obtaining the survey results for each tree in the areas to be reviewed based on the second tree scan results; and assessing the damage to trees in the forest farm based on the survey results for each tree in the areas to be reviewed. This application employs a dual-modal collaboration between a drone and a robot dog to monitor tree damage in a forest farm. The drone performs large-scale screening, while the robot dog accurately reviews suspected areas, significantly improving detection accuracy and inspection efficiency.

Description

Technical Field

[0001] This application relates at least to the field of forest farm monitoring technology, and in particular to a method, apparatus and system for surveying trees in forest farms. Background Technology

[0002] Existing methods for surveying trees in forest farms may utilize aerial scanning by drones or whole-forest scanning by robotic dogs. Aerial scanning by drones has a missed detection rate of >15% due to factors such as viewing angle and obstruction. Whole-forest scanning by robotic dogs is time-consuming and costly, and using a single device is difficult to adapt to the changing environment such as young forests, mature forests, and complex terrain. Summary of the Invention

[0003] To address the aforementioned shortcomings, this application provides a method, apparatus, and system for surveying trees in forest farms, aiming to solve the following technical problem: how to combine drones and robot dogs to achieve an efficient and accurate survey of trees in forest farms.

[0004] Firstly, this application provides a method for conducting a general survey of trees in a forest farm, the method comprising:

[0005] Obtain the first tree scan results of the first scan of the forest farm by the drone, draw a heat map of the distribution of the first tree scan results in the forest farm based on the first tree scan results, and obtain the areas to be reviewed in the forest farm based on the heat map of the distribution of the first tree scan results.

[0006] Obtain the second tree scan results from the robot dog's second scan of the area to be reviewed. Based on the second tree scan results, obtain the survey results for each tree in the area to be reviewed. Based on the survey results for each tree in the area to be reviewed, assess the damage to the trees in the forest.

[0007] Furthermore, based on the distribution heatmap of the first tree scan results, the areas to be verified in the forest farm are obtained, specifically including:

[0008] Based on the distribution heatmap of the first tree scan results, obtain the coordinate set of suspected damaged trees in the forest area:

[0009] ;

[0010] Obtain the terrain slope z along the x-coordinate of the forest and the planting row spacing of the trees along the y-coordinate. ;

[0011] According to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed.

[0012] Furthermore, according to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed, specifically including:

[0013] Plan to use M robot dogs to perform a second scan of the area to be checked. Divide the coordinate set of suspected damaged trees in the forest into M areas to be checked. The paths of the M robot dogs to perform the second scan of trees in the M areas to be checked satisfy the following formula: ,in, To preset weights, These represent the maximum difference between the x and y coordinates of several suspected damaged trees assigned to the m-th area to be reviewed. It is the average topographic slope value of the m-th area to be reviewed.

[0014] Furthermore, based on the survey results of each tree in the area to be reviewed, the damage to the trees in the forest farm will be assessed, specifically including:

[0015] Based on the survey results of each tree in the area to be reviewed, obtain the number of damaged trees in the forest farm;

[0016] The survival rate of trees in the forest farm is obtained based on the number of damaged trees and the total number of trees in the forest farm.

[0017] If the forest farm has tree asset insurance / mortgage, the survival rate of the trees in the forest farm will be sent to the insurance / mortgage institution's claims / post-loan management system in real time.

[0018] Furthermore, in response to the method used to verify the illegal logging and damage of trees in a forest farm, and assuming that each tree in the forest farm is equipped with a unique electronic tag, the method includes:

[0019] The first signal strength attenuation gradient of the first scan of trees in the forest by the drone is obtained. The signal strength attenuation heat map of the electronic tags in the forest is drawn based on the first signal strength attenuation gradient. The suspected illegal logging areas in the forest are obtained based on the signal strength attenuation heat map of the electronic tags.

[0020] The system obtains the second electronic tag recognition result of the robot dog's second scan of the electronic tags of trees in the suspected illegal logging area. Based on the second electronic tag recognition result, it obtains the true electronic tag status of each tree in the area to be reviewed. Based on the true electronic tag status of each tree in the area to be reviewed, it assesses the illegal logging damage to the trees in the forest farm.

[0021] Furthermore, the first signal strength attenuation gradient of the electronic tags on the trees in the forest is obtained during the first scan by the drone. A heatmap of electronic tag signal strength attenuation in the forest is then drawn based on this gradient. Suspected illegal logging areas in the forest are then identified based on this heatmap. Specifically, this includes:

[0022] Obtain the first signal strength attenuation gradient from the first scan of the electronic tags on each tree in the forest by the drone. ,in, The signal strength of the electronic tag received by the drone. This represents the minimum effective reception strength of the electronic tag under accessible conditions. The straight-line distance between the drone and the target tree;

[0023] According to each tree Plot a heatmap of electronic tag signal intensity attenuation in the forest farm using coordinates, and obtain... The coordinates of the trees form a set of coordinates for suspected damaged trees. , To determine the threshold based on the tree species in the forest, the coordinates of suspected damaged trees are set according to the forest topography and tree planting row spacing. The area was divided into several suspected illegal logging zones.

[0024] Furthermore, the method also includes:

[0025] It receives information on abnormal stress changes from a miniature strain sensor integrated into an electronic tag, and issues a logging theft alarm based on the abnormal stress change information.

[0026] Furthermore, in response to the method for verifying damage to trees in a forest farm, the method includes:

[0027] Obtain the first visual or infrared scan results of the first scan of trees in the forest by the drone, draw a heat map of the distribution of the first visual or infrared scan results in the forest, and obtain the accessible disaster areas in the forest based on the heat map of the distribution of the first visual or infrared scan results.

[0028] The system obtains the second visual or electronic tag recognition results of the robot dog's second scan of trees in the disaster-accessible area. Based on the second visual or electronic tag recognition results, it obtains the tree survival recognition results for each tree in the area to be reviewed. Based on the tree survival recognition results for each tree in the area to be reviewed, it assesses the disaster damage to the trees in the forest.

[0029] Secondly, this application provides a forestry tree survey device, the device comprising:

[0030] The first scanning analysis module is used to obtain the first tree scanning results of the first scan of the forest farm by the UAV, draw a heat map of the distribution of the first tree scanning results in the forest farm based on the first tree scanning results, and obtain the areas to be reviewed in the forest farm based on the heat map of the distribution of the first tree scanning results.

[0031] The second scanning analysis module, connected to the first scanning analysis module, is used to obtain the second tree scanning results of the robot dog performing a second scan of the area to be reviewed, obtain the survey results of each tree in the area to be reviewed based on the second tree scanning results, and assess the damage to the trees in the forest farm based on the survey results of each tree in the area to be reviewed.

[0032] Thirdly, this application provides a forest farm tree survey system, the system comprising:

[0033] A computer device for performing the forest tree survey method described above;

[0034] The drone, connected to the computer device, is used to perform the first scan of the forest and send the first tree scan results to the computer device.

[0035] The robot dog, connected to the computer device, is used to perform a second scan of the area to be reviewed according to the instructions of the computer device, and to send the second tree scan results to the computer device.

[0036] This application provides a method, device, and system for surveying trees in forest farms. It adopts a dual-modal collaboration of unmanned aerial vehicles (UAVs) and robotic dogs to monitor the damage to trees in forest farms. The UAVs conduct large-scale screening and generate a heat map of the screening results. Based on the heat map, suspected damaged areas that need to be verified are identified. The robotic dog performs precise verification of the suspected areas, thereby significantly improving the detection accuracy and inspection efficiency of the survey of trees in forest farms. Attached Figure Description

[0037] Figure 1 This is a flowchart of a forest tree survey method according to an embodiment of this application;

[0038] Figure 2 This is a flowchart of another forest farm tree survey method according to an embodiment of this application;

[0039] Figure 3 This is a schematic diagram of the structure of a forest tree survey device according to an embodiment of this application;

[0040] Figure 4 This is a schematic diagram of the structure of a forest tree survey system according to an embodiment of this application;

[0041] Figure 5 This is a schematic diagram of the structure of a computer-readable storage medium according to an embodiment of this application;

[0042] Figure 6 This is a schematic diagram of the structure of a computer device according to an embodiment of this application. Detailed Implementation

[0043] To enable those skilled in the art to better understand the technical solution of this application, the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0044] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining this application and are not intended to limit this application.

[0045] It is understood that, without conflict, the various embodiments and features in the embodiments of this application can be combined with each other.

[0046] It is understood that, for ease of description, only the parts relevant to this application are shown in the accompanying drawings, while parts unrelated to this application are not shown in the drawings.

[0047] It is understood that each module or unit involved in the embodiments of this application may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple modules or units may be integrated into one entity structure.

[0048] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this application may occur in a different order than that marked in the accompanying drawings.

[0049] It is understood that the flowcharts and block diagrams of this application illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, unit, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagrams and flowcharts may be implemented using a hardware-based device to implement the specified function, or using a combination of hardware and computer instructions.

[0050] It is understood that the modules and units involved in the embodiments of this application can be implemented by software or by hardware. For example, the modules and units can be located in the processor.

[0051] Example 1:

[0052] like Figure 1 As shown, this application provides a method for surveying trees in a forest farm, the method comprising:

[0053] S11. Obtain the first tree scan results of the first scan of the forest farm by the drone, draw a heat map of the distribution of the first tree scan results in the forest farm based on the first tree scan results, and obtain the area to be reviewed in the forest farm based on the heat map of the distribution of the first tree scan results.

[0054] S12. Obtain the second tree scan results of the robot dog performing a second scan of the area to be reviewed. Based on the second tree scan results, obtain the survey results of each tree in the area to be reviewed. Based on the survey results of each tree in the area to be reviewed, assess the damage to the trees in the forest farm.

[0055] In this embodiment, the provided method employs a dual-modal collaboration between a drone and a robot dog to monitor the damage to trees in a forest farm. The drone conducts a large-scale screening and generates a heat map of the screening results. Based on the heat map, suspected damaged areas that need to be reviewed are identified. The robot dog performs precise review of the suspected areas, thereby significantly improving the detection accuracy and inspection efficiency of the forest farm tree survey.

[0056] The first tree scan results reflect the survival and damage status of trees, the first tree scan result distribution heatmap reflects the distribution of tree survival and damage, and the second tree scan results reflect a more detailed tree damage situation.

[0057] Specifically, this embodiment provides a method for automatically surveying trees in forest farms based on RFID (Radio Frequency Identification), constructing a "dual-modal adaptive inspection system of drones and robot dogs" to achieve intelligent inspection of forest areas through efficient collaboration, which has the following beneficial effects:

[0058] 1. In response to the fact that the false negative rate of high-altitude scanning by UAVs in the existing technology is >15% due to factors such as viewing angle and obstruction, this embodiment adopts dual-modal collaboration of UAV and robot dog. The UAV conducts large-scale screening, while the robot dog conducts precise verification of suspected areas, thereby reducing the false negative rate to <0.5% and significantly improving the detection accuracy.

[0059] 2. In view of the time-consuming and costly nature of full-forest scanning by robot dogs in the existing technology, this embodiment only dispatches robot dogs to re-examine some suspected areas identified by drones, realizing targeted operation, greatly shortening the path and time, and significantly improving inspection efficiency.

[0060] 3. In view of the fact that existing technologies using a single device are difficult to adapt to the changing environment such as young forests, mature forests, and complex terrain, this embodiment can dynamically match the characteristics of the forest area and intelligently adjust the collaborative strategy and parameters of the two devices to ensure efficient and stable operation in various environments and overcome the limitations of adaptability.

[0061] In one embodiment, the area to be reviewed in the forest farm is obtained based on the distribution heat map of the first tree scan results, specifically including:

[0062] Based on the distribution heatmap of the first tree scan results, obtain the coordinate set of suspected damaged trees in the forest area:

[0063] ;

[0064] Obtain the terrain slope z along the x-coordinate of the forest and the planting row spacing of the trees along the y-coordinate. ;

[0065] According to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed.

[0066] In this embodiment, the overall process of this solution is as follows: Figure 2 As shown, it includes:

[0067] (1) The UAV performs the first round of large-scale scanning and generates a signal attenuation heat map to preliminarily assess the status of the tags in the area.

[0068] (2) After the scan is completed, the system automatically analyzes the data and determines whether the signal attenuation value exceeds the preset safety threshold.

[0069] (3.1) If the signal attenuation value does not exceed the threshold, it indicates that the tag survival status in the area is good. At this time, the system will update the database to record the results of this inspection and end this task.

[0070] (3.2) If the signal attenuation value exceeds the set threshold, the system will mark these suspicious areas and prepare for further investigation.

[0071] (4) The robot dog is dispatched to the suspicious area identified by the drone for in-depth verification. It can scan each tag in the suspicious area at close range and from multiple angles to accurately verify the true status of the tag.

[0072] (5) Based on the review results provided by the robot dog, the system will take corresponding actions: if the tag is confirmed to exist, the relevant information in the database will be updated; otherwise, if the tag cannot be found, the system will trigger the illegal logging alarm mechanism and promptly notify relevant personnel to take measures.

[0073] In one embodiment, according to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed, specifically including:

[0074] Plan to use M robot dogs to perform a second scan of the area to be checked. Divide the coordinate set of suspected damaged trees in the forest into M areas to be checked. The paths of the M robot dogs to perform the second scan of trees in the M areas to be checked satisfy the following formula: ,in, To preset weights, These represent the maximum difference between the x and y coordinates of several suspected damaged trees assigned to the m-th area to be reviewed. It is the average topographic slope value of the m-th area to be reviewed.

[0075] In this embodiment, the core algorithm of this solution includes a robot dog optimal re-examination path planning algorithm. To improve ground re-examination efficiency and reduce energy consumption, this system designs a robot dog optimal re-examination path planning algorithm for complex terrain in forest areas. This algorithm achieves path optimization and dynamic obstacle avoidance while ensuring full coverage. Specifically, it includes:

[0076] Input parameters: Set of suspicious tag points identified by the drone And the planting row spacing in forest areas .

[0077]

[0078] in, These are the coordinates of suspicious points recorded by the decay gradient threshold triggering algorithm.

[0079] Clustering by Planting Row: Based on the planting patterns in the forest area, suspicious points are clustered according to their spatial distribution. If the distance between two points in the y-coordinate direction is less than 0.5... If they are identified as belonging to the same planting row, row-level classification of suspicious tag point sets is achieved, ensuring that path planning conforms to the actual forest row structure.

[0080] Inline point sorting: For each clustered inline point set, sort them in ascending order according to the x-coordinate to form an ordered scanning sequence along the row direction, reducing inline back-and-forth movement and improving scanning continuity.

[0081] Main path generation and cross-row movement optimization: The robot dog enters from the starting point of each row and performs a close scan along the sorted point sequence to complete the single-row re-check. During cross-row transitions, a "minimum energy path" strategy is adopted, with the objective function being:

[0082]

[0083] in, This represents the terrain slope value at the path point. The distance moved in the row direction. The horizontal movement distance between rows, with the weight preset as follows: =0.1, =0.2, prioritizing paths with gentle slopes and short lateral movements, significantly reducing the robot dog's energy consumption and traffic risks.

[0084] Dynamic obstacle avoidance and path update: The robot dog is equipped with a lidar and runs a SLAM (Simultaneous Localization and Mapping) system to perceive obstacles ahead (such as fallen trees, ditches, bushes, etc.) in real time. Once the path is blocked, the system immediately replans the local path to maintain mission continuity while ensuring safety.

[0085] This embodiment proposes the above-mentioned structured clustering and energy-optimal cross-row model. Considering the strong regularity of planting patterns in forest areas, a composite path planning algorithm of "row clustering + minimum energy path" is proposed. First, suspicious points are clustered according to planting rows based on their y-coordinates (row spacing tolerance < 0.5). Then, the points in each row are sorted by their x-coordinates to form a structured scanning sequence. Finally, an objective function is constructed to achieve the optimal balance between energy consumption and traffic safety. This model fully considers the forest slope and the robot dog's obstacle-crossing ability, significantly reducing ineffective movement and energy consumption.

[0086] In one implementation, the damage to trees in the forest farm is assessed based on the survey results of each tree in the area to be reviewed, specifically including:

[0087] Based on the survey results of each tree in the area to be reviewed, obtain the number of damaged trees in the forest farm;

[0088] The survival rate of trees in the forest farm is obtained based on the number of damaged trees and the total number of trees in the forest farm.

[0089] If the forest farm has tree asset insurance / mortgage, the survival rate of the trees in the forest farm will be sent to the insurance / mortgage institution's claims / post-loan management system in real time.

[0090] In this embodiment, the system includes a financial risk control interface. The system can convert inspection results into financial-grade data assets and output a standardized "Forest Asset Survival Rate" API (Application Programming Interface). This interface connects in real-time to the post-loan management system of banks or financial institutions, dynamically reflecting the health status of mortgaged forest assets. When the survival rate falls below a warning threshold, a risk alert is automatically triggered, assisting financial institutions in post-loan risk monitoring and decision-making, improving the transparency and risk control capabilities of forestry credit, and promoting the digital implementation of "green finance." Similarly, it can be used for loss assessment in insurance claims.

[0091] In one embodiment, in response to the method for verifying illegal logging and damage to trees in a forest farm, wherein each tree in the forest farm is equipped with a unique electronic tag, the method includes:

[0092] The first signal strength attenuation gradient of the first scan of trees in the forest by the drone is obtained. The signal strength attenuation heat map of the electronic tags in the forest is drawn based on the first signal strength attenuation gradient. The suspected illegal logging areas in the forest are obtained based on the signal strength attenuation heat map of the electronic tags.

[0093] The system obtains the second electronic tag recognition result of the robot dog's second scan of the electronic tags of trees in the suspected illegal logging area. Based on the second electronic tag recognition result, it obtains the true electronic tag status of each tree in the area to be reviewed. Based on the true electronic tag status of each tree in the area to be reviewed, it assesses the illegal logging damage to the trees in the forest farm.

[0094] In this embodiment, the solution can be applied to an intelligent anti-theft logging mode. The drone performs a wide-area, rapid scan at a safe height of approximately 5 meters above the tree canopy, efficiently activating and initially locating target tags. For areas suspected of being missed due to drone obstruction or signal attenuation, a robotic dog performs close-range, precise verification on the ground, completing the review through close-range, multi-angle scanning between rows to ensure identification accuracy. The system's core is equipped with a dynamic task allocation engine, which can intelligently assess detection confidence based on real-time signal attenuation data, automatically switching or coordinating aerial and ground scanning modes to achieve optimal task allocation. This solution balances efficiency and accuracy and has the ability to dynamically adapt to complex forest conditions.

[0095] In one embodiment, the first signal strength attenuation gradient of the electronic tags on trees in the forest area is obtained during the first scan by the UAV; a heat map of electronic tag signal strength attenuation in the forest area is drawn based on the first signal strength attenuation gradient; and suspected illegal logging areas in the forest area are identified based on the electronic tag signal strength attenuation heat map. Specifically, this includes:

[0096] Obtain the first signal strength attenuation gradient from the first scan of the electronic tags on each tree in the forest by the drone. ,in, The signal strength of the electronic tag received by the drone. This represents the minimum effective reception strength of the electronic tag under accessible conditions. The straight-line distance between the drone and the target tree;

[0097] According to each tree Plot a heatmap of electronic tag signal intensity attenuation in the forest farm using coordinates, and obtain... The coordinates of the trees form a set of coordinates for suspected damaged trees. , To determine the threshold based on the tree species in the forest, the coordinates of suspected damaged trees are set according to the forest topography and tree planting row spacing. The area was divided into several suspected illegal logging zones.

[0098] In this embodiment, the core algorithm of this solution includes an attenuation gradient threshold triggering algorithm. To accurately identify areas of signal anomalies caused by factors such as vegetation obstruction and terrain shading, this system introduces an attenuation gradient threshold triggering algorithm to intelligently determine whether a robot dog needs to be activated for ground verification. Specifically, it includes:

[0099] Define the signal attenuation gradient ∇A, and its calculation formula is as follows:

[0100]

[0101] in, The strength of the tag signal received by the drone. The minimum effective reception strength of the tag under unobstructed conditions, such as -18 dBm. The straight-line distance between the drone and the target tree trunk.

[0102] If ∇A (like =0.4dB / m (empirical value for pine forests), then the signal is determined to be blocked, and a set of coordinates of suspicious tag points for drone identification is generated. Add it to the robot dog task queue.

[0103] If ∇A (like If the effective value range is 0.2 ~ 0.6 dB / m (applicable to typical plantation forests such as pine and fir), then the label is confirmed to be valid and the trees are alive and have not been illegally logged or damaged.

[0104] This embodiment proposes the above-mentioned intelligent scheduling based on signal attenuation gradient. The signal attenuation gradient ∇A is used as the trigger criterion for the collaborative operation of UAV and robot dog. By calculating the signal attenuation rate per unit distance, the system can accurately identify signal abnormality areas caused by vegetation obstruction. When ∇A exceeds the set threshold α, the robot dog's re-examination task is automatically triggered.

[0105] In one embodiment, the method further includes:

[0106] It receives information on abnormal stress changes from a miniature strain sensor integrated into an electronic tag, and issues a logging theft alarm based on the abnormal stress change information.

[0107] In this embodiment, the tag can be upgraded to integrate a strain sensor, such as upgrading an existing RFID tag to a smart composite tag that integrates a miniature strain sensor. When a tree undergoes severe physical deformation such as falling or being felled, the sensor detects abnormal stress changes, immediately triggering an audible and visual alarm and uploading a "physical logging theft" alarm via a wireless network. This mechanism achieves a second-level response to logging theft, compensating for the delay in signal obstruction identification and significantly improving the real-time performance and deterrent effect of protection.

[0108] In one embodiment, in response to the method for verifying damage to trees in a forest, the method includes:

[0109] Obtain the first visual or infrared scan results of the first scan of trees in the forest by the drone, draw a heat map of the distribution of the first visual or infrared scan results in the forest, and obtain the accessible disaster areas in the forest based on the heat map of the distribution of the first visual or infrared scan results.

[0110] The system obtains the second visual or electronic tag recognition results of the robot dog's second scan of trees in the disaster-accessible area. Based on the second visual or electronic tag recognition results, it obtains the tree survival recognition results for each tree in the area to be reviewed. Based on the tree survival recognition results for each tree in the area to be reviewed, it assesses the disaster damage to the trees in the forest.

[0111] In this embodiment, the solution can be applied to disaster emergency response modes. For example, in sudden disaster scenarios such as forest fires, drones can take off to perform large-scale thermal infrared monitoring, tracking the spread of fire lines and high-temperature areas in real time. Simultaneously, the system delineates safe zones based on a fire prediction model and dispatches a drone to these zones to quickly verify tree tags. By confirming the tags' survival status, the extent of damage can be assessed immediately, providing accurate data support for emergency command and post-disaster claims, achieving a highly efficient "rescue while assessing damage" response. Similarly, the above functions can be implemented based on a visual analysis solution.

[0112] The following is a more specific implementation example:

[0113] Scenario: A 2-hectare pine forest with an average tree height of 20 meters and a planting row spacing of... =5 meters.

[0114] Implementation process:

[0115] First scan by drone

[0116] The drone, hovering 25 meters above the ground at a cruising speed of 8 m / s and a transmit power of 30 dBm, conducted a full-coverage aerial scan of the forest area. The system collected the RSSI (Received Signal Strength Indicator) signal strength of each tag in real time, combining this data with the straight-line distance between the drone and the tree trunks. Calculate the signal attenuation gradient ∇A and connect it to the signal attenuation gradient. A comparison of 0.4 dB / m yielded a set of suspicious point coordinates. The first round of scanning identified 182 suspicious points, and the system automatically added their geographical coordinates to the robot dog's task queue.

[0117] Robot dog scheduling and ground review

[0118] Based on the "robot dog optimal re-inspection path planning algorithm," the system performed row clustering on 182 suspicious points (identifying planting rows with a y-coordinate difference of < 2.5 m) and arranged the points in ascending order of x-coordinate. A "minimum energy path" strategy was adopted for cross-row paths, comprehensively considering terrain slope and lateral movement distance to optimize energy consumption. The robot dog traveled along a 5-meter row spacing path, equipped with a laser SLAM system for real-time obstacle avoidance, and performed close-up scanning of the final suspicious points (approximately 1-2 meters from the tag). The entire re-inspection process took 15 minutes, ultimately confirming that 12 tags, despite signal attenuation due to new shrub obstruction, were still alive, thus ruling out false alarms about illegal logging.

[0119] Overall effect

[0120] The inspection took a total of 38 minutes (23 minutes for drone scanning + 15 minutes for robot dog re-inspection), achieving a 0% miss rate (compared to a 22% miss rate for pure drone solutions). The accuracy of locating illegal logging sites reached ±1.5 meters, which is significantly better than traditional single-mode inspections.

[0121] This embodiment proposes the above-mentioned dual-device error mutual compensation mechanism, constructing a dual-device error mutual compensation system between the UAV and the robot dog. The UAV, leveraging its high-altitude field of view, performs a wide-area scan, effectively covering areas inaccessible to ground equipment and providing global guidance; the robot dog, through close-range inspection, accurately identifies targets missed by the UAV due to foliage obstruction or signal attenuation, enabling ground verification. The two work together to fill in high-altitude omissions on the ground and cover ground blind spots from the air, forming a closed-loop detection system and significantly improving the overall integrity, robustness, and inspection reliability of the system.

[0122] This embodiment leverages clear physical advantages to achieve significant performance improvements at minimal cost. Furthermore, it relies on mature hardware and an open ecosystem, with a clear technical path, simple integration, and strong potential for practical application and commercialization.

[0123] 1. High level of technological maturity: Modules can be directly integrated.

[0124] It adopts mature commercial RFID read / write chips (such as Impinj R2000), supports dynamic beamforming and high-sensitivity reception, and can be seamlessly integrated into the robot dog platform to improve tag capture capabilities;

[0125] Drone platforms (such as DJI M300) and mainstream robot dogs (such as Doosan DOGS) have opened up their underlying control and communication APIs, supporting task scheduling, path coordination and data feedback, which facilitates the rapid development of air-ground collaborative control systems and significantly shortens the R&D cycle and engineering difficulty.

[0126] 2. Economic balance: Controllable costs and outstanding benefits.

[0127] The robot dog only needs to re-inspect a few high-attenuation, suspicious areas, resulting in a minimal workload and avoiding the high energy consumption of full-forest patrols. Although the introduction of the robot dog increases the cost of a single inspection by about 15%, the missed detection rate drops from 22% with pure drone solutions to below 0.5%, improving detection reliability by more than 22 times. In high-value scenarios such as forestry finance and carbon sequestration, the risk avoidance and management benefits brought by accurate data far outweigh the cost increase, demonstrating an excellent return on investment.

[0128] 3. Physical layer advantages: Significant signal attenuation advantages

[0129] When a drone scans at an altitude of 18 meters, the signal needs to penetrate dense tree canopy, resulting in a measured attenuation of 15-25 dB, severely impacting recognition reliability. In contrast, a robotic dog can operate within 3 meters of the target, with the same tag signal attenuating by only about 3 dB, significantly improving the signal-to-noise ratio. This leap in signal quality due to the increased physical distance provides a fundamental guarantee for accurate ground-based verification and is the core physical basis for achieving high read rates.

[0130] The advantages of this embodiment include:

[0131] This system enables automated, high-frequency status inspections and theft monitoring of forest assets, eliminating reliance on manual sampling and emphasizing autonomous system operation capabilities, making it suitable for large-scale asset management scenarios. It constructs a fully automated, system-level, multi-target synchronous monitoring platform, achieving unmanned and continuous inspections through air-ground collaboration, significantly improving operational efficiency and management precision. As a mobile, fully covered, and scalable dynamic inspection system, it is suitable for the daily management of large forest areas, possessing greater practicality and economy, and is more suitable for routine operation and maintenance needs.

[0132] A dual-modal collaborative mechanism combining UAV aerial scanning and robot dog ground verification, along with an attenuation gradient triggering algorithm, enables large-scale, high-precision, and adaptive intelligent inspection, significantly reducing the missed detection rate and improving system automation and response speed. Through UAV-robot dog collaboration, ground verification is dynamically triggered based on the attenuation amount ∇A, achieving all-terrain, schedulable dynamic inspection, better suited to the needs of large-scale plantation forest management. Unmanned rapid inventory of forest farms is achieved through collaborative aerial scanning by UAVs and ground scanning by robot dogs.

[0133] Employing a "tag signal attenuation gradient" algorithm combined with data fusion from drones and robotic dogs, this system enables dynamic inventory management of single-tree assets. Utilizing low-cost RFID tags and dual-device collaborative identification, it focuses on asset viability assessment and theft alerts, offering fast response, low cost, and easy deployment, making it ideal for commercial applications such as financial risk control and insurance loss assessment. The robotic dog's close-range scanning obtains the precise coordinates (±1.5m) of missing tags, supporting the tracking of theft trajectories. Based on RFID signal attenuation triggering collaborative inspections by drones and robotic dogs, it achieves forest theft prevention and status verification, emphasizing asset security and closed-loop alarm systems.

[0134] By employing a collaborative approach between drones and robotic dogs, and based on RFID identification of trees, this system utilizes signal analysis and dynamic path planning to achieve autonomous inspection and closed-loop alarm systems for asset theft prevention. It boasts enhanced mobility, intelligence, and targeted capabilities. Requiring only the survival status and location of RFID tags, it employs a lightweight decision tree algorithm to determine illegal logging, making it more convenient, efficient, and suitable for low-cost, high-frequency asset inspections. Using low-cost commercial RFID tags, it requires no special equipment and can simultaneously achieve asset inventory and illegal logging monitoring, demonstrating strong versatility and ease of large-scale application.

[0135] Example 2:

[0136] like Figure 3 As shown, this application provides a forestry tree survey device, the device comprising:

[0137] The first scanning analysis module 11 is used to obtain the first tree scanning results of the first scan of the forest farm by the UAV, draw a heat map of the distribution of the first tree scanning results of the forest farm based on the first tree scanning results, and obtain the area to be reviewed in the forest farm based on the heat map of the distribution of the first tree scanning results.

[0138] The second scanning analysis module 12, connected to the first scanning analysis module 11, is used to obtain the second tree scanning results of the robot dog performing a second scan of the area to be reviewed, obtain the survey results of each tree in the area to be reviewed based on the second tree scanning results, and assess the damage to the trees in the forest farm based on the survey results of each tree in the area to be reviewed.

[0139] In one embodiment, the first scanning analysis module 11 includes a region to be verified determination unit, specifically including:

[0140] The damaged coordinate set unit is used to obtain the coordinate set of suspected damaged trees in the forest based on the distribution heatmap of the first tree scan results:

[0141] ;

[0142] The forest farm data unit is used to obtain the terrain slope z along the x-coordinate direction and the planting row spacing of trees along the y-coordinate direction. ;

[0143] The energy-saving partitioning unit, connected to the damaged coordinate set unit and the forest farm data unit, is used to determine the data based on z and... The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed.

[0144] In one embodiment, the energy-saving partitioning unit is specifically used for:

[0145] Plan to use M robot dogs to perform a second scan of the area to be checked. Divide the coordinate set of suspected damaged trees in the forest into M areas to be checked. The paths of the M robot dogs to perform the second scan of trees in the M areas to be checked satisfy the following formula: ,in, To preset weights, These represent the maximum difference between the x and y coordinates of several suspected damaged trees assigned to the m-th area to be reviewed. It is the average topographic slope value of the m-th area to be reviewed.

[0146] In one embodiment, the second scanning analysis module 12 includes a damage assessment unit, specifically comprising:

[0147] The damage quantity unit is used to obtain the number of damaged trees in the forest farm based on the survey results of each tree in the area to be verified;

[0148] The damage ratio unit, connected to the damage quantity unit, is used to obtain the asset survival rate of trees in the forest farm based on the number of damaged trees and the total number of trees in the forest farm.

[0149] The insurance / mortgage connection unit, connected to the damage ratio unit, is used to send the survival rate of trees in the forest farm to the claims / loan management system of the insurance / mortgage institution in real time if there is tree asset insurance / mortgage in the forest farm.

[0150] In one embodiment, in response to the method for verifying illegal logging and damage to trees in a forest farm, wherein each tree in the forest farm is equipped with a unique electronic tag, the device includes:

[0151] The first scanning analysis module 11 is a signal strength attenuation analysis module, used to obtain the first signal strength attenuation gradient of the first scan of the electronic tags of trees in the forest by the UAV, draw the electronic tag signal strength attenuation heat map of the forest based on the first signal strength attenuation gradient, and obtain the suspected illegal logging area in the forest based on the electronic tag signal strength attenuation heat map.

[0152] The second scanning analysis module 12 is an electronic tag recognition module, used to obtain the second electronic tag recognition result of the robot dog performing a second scan of the electronic tags of trees in the suspected illegal logging area, obtain the real status of the electronic tags of each tree in the area to be reviewed based on the second electronic tag recognition result, and assess the illegal logging damage of trees in the forest farm based on the real status of the electronic tags of each tree in the area to be reviewed.

[0153] In one embodiment, the signal strength attenuation analysis module specifically includes:

[0154] The signal strength attenuation gradient unit is used to obtain the first signal strength attenuation gradient of the electronic tags on each tree in the forest during the first scan by the UAV. ,in, The signal strength of the electronic tag received by the drone. This represents the minimum effective reception strength of the electronic tag under accessible conditions. The straight-line distance between the drone and the target tree;

[0155] The heatmap analysis unit, connected to the signal intensity attenuation gradient unit, is used to analyze the heatmap data of each tree. Plot a heatmap of electronic tag signal intensity attenuation in the forest farm using coordinates, and obtain... The coordinates of the trees form a set of coordinates for suspected damaged trees. , To determine the threshold based on the tree species in the forest, the coordinates of suspected damaged trees are set according to the forest topography and tree planting row spacing. The area was divided into several suspected illegal logging zones.

[0156] In one embodiment, the apparatus further includes:

[0157] The electronic tag strain analysis module is used to receive abnormal stress change information from the miniature strain sensor integrated in the electronic tag, and issue a logging theft alarm based on the abnormal stress change information.

[0158] In one embodiment, in response to the method for verifying damage to trees in a forest, the apparatus includes:

[0159] The first scanning analysis module 11 is a first visual or infrared scanning analysis module, used to obtain the first visual or infrared scanning results of the first scan of trees in the forest by the UAV, draw a heat map of the distribution of the first visual or infrared scanning results in the forest based on the first visual or infrared scanning results, and obtain the accessible disaster areas in the forest based on the heat map of the distribution of the first visual or infrared scanning results.

[0160] The second scanning analysis module 12 is a second vision or electronic tag recognition module, used to obtain the second vision or electronic tag recognition results of the robot dog's second scan of trees that can enter the disaster area, obtain the tree survival recognition results of each tree in the area to be reviewed based on the second vision or electronic tag recognition results, and assess the disaster damage to trees in the forest farm based on the tree survival recognition results of each tree in the area to be reviewed.

[0161] Example 3:

[0162] like Figure 4 As shown, Embodiment 3 of this application provides a forest farm tree survey system, the system comprising:

[0163] Computer device 1 is used to perform the forest tree survey method as described in Example 1;

[0164] The drone 2 is connected to the computer device 1 and is used to perform the first scan of the forest and send the first tree scan results to the computer device 1.

[0165] The robot dog 3 is connected to the computer device 1 and is used to perform a second scan of the area to be reviewed according to the instructions of the computer device 1, and send the second tree scan results to the computer device 1.

[0166] like Figure 5 As shown, this application can also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the forest tree survey method as described in Embodiment 1 or as described in Embodiment 2. Figure 4 The computer device 1 shown.

[0167] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program units, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0168] like Figure 6 As shown, this application may also provide a computer device 1 (such as...) Figure 4 As shown, it includes a memory and a processor. The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the forest tree survey method as described in Example 1.

[0169] The memory is connected to the processor. The memory can be flash memory, read-only memory or other types of memory. The processor can be a central processing unit or a microcontroller.

[0170] Embodiments 1-3 of this application provide a method, apparatus, and system for surveying trees in forest farms. The system employs a dual-modal collaboration of unmanned aerial vehicles (UAVs) and robotic dogs to monitor the damage to trees in forest farms. The UAVs conduct large-scale screening and generate a heat map of the screening results. Based on the heat map, suspected damaged areas that need to be verified are identified. The robotic dog performs precise verification of the suspected areas, thereby significantly improving the detection accuracy and inspection efficiency of the survey of trees in forest farms.

[0171] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of this application, and this application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this application, and these modifications and improvements are also considered to be within the scope of protection of this application.

Claims

1. A method for conducting a general survey of trees in a forest farm, characterized in that, The method includes: Obtain the first tree scan results of the first scan of the forest farm by the drone, draw a heat map of the distribution of the first tree scan results in the forest farm based on the first tree scan results, and obtain the areas to be reviewed in the forest farm based on the heat map of the distribution of the first tree scan results. Obtain the second tree scan results from the robot dog's second scan of the area to be reviewed. Based on the second tree scan results, obtain the survey results for each tree in the area to be reviewed. Based on the survey results for each tree in the area to be reviewed, assess the damage to the trees in the forest.

2. The method according to claim 1, characterized in that, Based on the distribution heatmap of the first tree scan results, the areas in the forest farm to be reviewed were identified, specifically including: Based on the distribution heatmap of the first tree scan results, obtain the coordinate set of suspected damaged trees in the forest area: ； Obtain the terrain slope z along the x-coordinate of the forest and the planting row spacing of the trees along the y-coordinate. ; According to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed.

3. The method according to claim 2, characterized in that, According to z and The minimum energy path for the robot dog to perform a second scan of the area to be reviewed is planned to divide the coordinate set of suspected damaged trees in the forest into several areas to be reviewed, specifically including: Plan to use M robot dogs to perform a second scan of the area to be checked. Divide the coordinate set of suspected damaged trees in the forest into M areas to be checked. The paths of the M robot dogs to perform the second scan of trees in the M areas to be checked satisfy the following formula: ,in, To preset weights, These represent the maximum difference between the x and y coordinates of several suspected damaged trees assigned to the m-th area to be reviewed. It is the average topographic slope value of the m-th area to be reviewed.

4. The method according to claim 1, characterized in that, The damage to trees in the forest farm will be assessed based on the results of a survey of every tree in the area to be reviewed, specifically including: Based on the survey results of each tree in the area to be reviewed, obtain the number of damaged trees in the forest farm; The survival rate of trees in the forest farm is obtained based on the number of damaged trees and the total number of trees in the forest farm. If the forest farm has tree asset insurance / mortgage, the survival rate of the trees in the forest farm will be sent to the insurance / mortgage institution's claims / post-loan management system in real time.

5. The method according to any one of claims 1-4, characterized in that, In response to the method used to verify the illegal logging and damage of trees in a forest farm, wherein each tree in the forest farm is equipped with a unique electronic tag, the method includes: The first signal strength attenuation gradient of the first scan of trees in the forest by the drone is obtained. The signal strength attenuation heat map of the electronic tags in the forest is drawn based on the first signal strength attenuation gradient. The suspected illegal logging areas in the forest are obtained based on the signal strength attenuation heat map of the electronic tags. The system obtains the second electronic tag recognition result of the robot dog's second scan of the electronic tags of trees in the suspected illegal logging area. Based on the second electronic tag recognition result, it obtains the true electronic tag status of each tree in the area to be reviewed. Based on the true electronic tag status of each tree in the area to be reviewed, it assesses the illegal logging damage to the trees in the forest farm.

6. The method according to claim 5, characterized in that, The process involves obtaining the first signal strength attenuation gradient from the first scan of trees in the forest by a drone using electronic tags, creating a heatmap of electronic tag signal strength attenuation based on this gradient, and identifying suspected illegal logging areas within the forest based on this heatmap. Specifically, this includes: Obtain the first signal strength attenuation gradient from the first scan of the electronic tags on each tree in the forest by the drone. ,in, The signal strength of the electronic tag received by the drone. This represents the minimum effective reception strength of the electronic tag under accessible conditions. The straight-line distance between the drone and the target tree; According to each tree Plot a heatmap of electronic tag signal intensity attenuation in the forest farm using coordinates, and obtain... The coordinates of the trees form a set of coordinates for suspected damaged trees. , To determine the threshold based on the tree species in the forest, the coordinates of suspected damaged trees are set according to the forest topography and tree planting row spacing. The area was divided into several suspected illegal logging zones.

7. The method according to claim 5, characterized in that, The method further includes: It receives information on abnormal stress changes from a miniature strain sensor integrated into an electronic tag, and issues a logging theft alarm based on the abnormal stress change information.

8. The method according to any one of claims 1-4, characterized in that, In response to the method for verifying damage to trees in a forest farm, the method includes: Obtain the first visual or infrared scan results of the first scan of trees in the forest by the drone, draw a heat map of the distribution of the first visual or infrared scan results in the forest, and obtain the accessible disaster areas in the forest based on the heat map of the distribution of the first visual or infrared scan results. The system obtains the second visual or electronic tag recognition results of the robot dog's second scan of trees in the disaster-accessible area. Based on the second visual or electronic tag recognition results, it obtains the tree survival recognition results for each tree in the area to be reviewed. Based on the tree survival recognition results for each tree in the area to be reviewed, it assesses the disaster damage to the trees in the forest.

9. A forestry tree survey device, characterized in that, The device includes: The first scanning analysis module is used to obtain the first tree scanning results of the first scan of the forest farm by the UAV, draw a heat map of the distribution of the first tree scanning results in the forest farm based on the first tree scanning results, and obtain the areas to be reviewed in the forest farm based on the heat map of the distribution of the first tree scanning results. The second scanning analysis module, connected to the first scanning analysis module, is used to obtain the second tree scanning results of the robot dog performing a second scan of the area to be reviewed, obtain the survey results of each tree in the area to be reviewed based on the second tree scanning results, and assess the damage to the trees in the forest farm based on the survey results of each tree in the area to be reviewed.

10. A forest farm tree survey system, characterized in that, The system includes: A computer device for performing the forest tree survey method as described in any one of claims 1-8; The drone, connected to the computer device, is used to perform the first scan of the forest and send the first tree scan results to the computer device. The robot dog, connected to the computer device, is used to perform a second scan of the area to be reviewed according to the instructions of the computer device, and to send the second tree scan results to the computer device.

Images