Bird repelling device transfer planning method and system based on bird activity trajectory data analysis
Patent Information
- Application Number
- CN202611059674.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-16
- Publication Date
- 2026-10-09
Smart Images
Figure CN122886918A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of bird deterrence strategy planning technology, and in particular to a method and system for relocating bird deterrence devices based on bird activity trajectory data analysis. Background Technology
[0002] Birds, when active in areas requiring bird control, such as airports, substations, power lines, farms, and other locations, can negatively impact aviation operations, power facility safety, crop protection, and on-site work safety. Current technologies typically employ methods such as manual patrols, fixed bird deterrent devices, vehicle-mounted bird deterrent devices, radar bird detection equipment, video monitoring equipment, infrared monitoring equipment, acoustic bird deterrent devices, ultrasonic bird deterrent devices, laser bird deterrent devices, and drone inspections for bird monitoring and control. For example, existing solutions utilize radar, infrared, and video bird detection equipment to acquire real-time bird information around airports, assess bird strike risk levels based on this information, and then activate bird deterrent devices such as gas cannons, directional acoustic waves, and ultrasonic waves to carry out bird deterrence operations. Relevant airport bird strike prevention documents also mention that proven and effective radar-based photoelectric bird monitoring and deterrence equipment can reduce the need for manual patrols to some extent. However, existing bird control methods mostly focus on fixed-point deployment, manual judgment based on experience, or immediate driving away of flocks after they are discovered. They are difficult to fully utilize activity trajectory data such as flock location, flock size, and discovery time points obtained during multiple rounds of inspections to model and analyze the recurrence, activity stability, and changing trends of flocks at different time points and spatial locations. At the same time, when the number of bird control devices is limited, the transfer efficiency is restricted, or the activity area of the flock changes, existing methods cannot accurately determine the transfer location, timing, and priority of the bird control devices. This can easily lead to problems such as ineffective premature deployment of bird control devices, delayed arrival, repeated coverage of low-risk areas, or omission of high-risk areas, thereby reducing the utilization rate of bird control resources and the effectiveness of bird control. Summary of the Invention
[0003] The purpose of this invention is to provide a planning method and system for relatively efficient bird deterrence devices.
[0004] This invention discloses a method for relocation planning of bird deterrent devices based on bird activity trajectory data analysis, comprising: Step S100: Construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock size and discovery time nodes on the preset activity trajectory demonstration surface, and based on the activity trajectory demonstration surface and bird flock marking data, construct an initial bird distribution model. Step S200: Perform a holistic analysis on several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, optimize the initial UAV inspection route to obtain several optimized UAV inspection routes. Step S300: Bird inspection is carried out based on the optimized UAV inspection route, and a third-order bird distribution model is constructed based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are fused to obtain a reference bird distribution model. Step S400: Based on the reference bird distribution model, determine the bird deterrent device relocation planning strategy.
[0005] In some embodiments disclosed in this invention, the method for constructing an initial bird distribution model based on the activity trajectory demonstration surface and bird flock marking data includes: Step S101: Divide the activity trajectory demonstration surface into a grid and construct several layers of activity trajectory demonstration surfaces above the activity trajectory demonstration surface. Each layer of activity trajectory demonstration surface corresponds to a preset time node. Step S102: Determine the location of the bird flock discovered by the drone, and highlight the corresponding grid on the activity trajectory demonstration surface based on the location of the bird flock to obtain a highlighted grid group. Based on the discovery time node, map the highlighted grid group onto the corresponding activity trajectory demonstration surface and mark the size of the bird flock.
[0006] In some embodiments disclosed in this invention, the method for performing a holistic analysis of several initial bird distribution models includes: Step S201: Map the representation of each activity trajectory demonstration surface in all initial bird distribution models into the bird distribution model for analysis in a unified manner. Step S202: Analyze the overlapping of prominent grids in each layer of the activity trajectory demonstration surface in the bird distribution model for analysis, determine the number of overlaps of each grid, and calculate the overlap ratio between the number of overlaps and the number of initial bird distribution models. Step S203: Compare the overlap ratio of the protruding grids in each layer of the activity trajectory demonstration surface with a threshold interval, mark the protruding grids belonging to the same threshold interval, and associate the corresponding protruding grids that are less than or equal to a preset value and have the same interval marking to form several second-order grid clusters. The common performance of several layers of activity trajectory demonstration surfaces is recorded as a second-order bird distribution model.
[0007] In some embodiments disclosed in this invention, the method for optimizing the initial UAV inspection route based on a second-order bird distribution model includes: Step S2031 involves analyzing the second-order grid clusters in each activity estimation demonstration surface of the second-order bird distribution model, including the location node of the cluster, the threshold interval to which the cluster belongs, and the time node to which the cluster belongs. Step S2032: Using the time node to which the cluster belongs as a constraint, randomly select the location node to which the cluster belongs, and construct the UAV inspection route. If the arrival time node is earlier than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then add other inspection paths before the location node to which the cluster belongs. If the arrival time node is later than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then remove the corresponding location node to which the cluster belongs.
[0008] In some embodiments disclosed in this invention, the method for constructing a third-order bird distribution model includes: Step S301: Based on the bird flock marking data, identify the prominent grid groups on each activity trajectory demonstration surface, and mark the corresponding bird flock size on the prominent grid groups; Step S302: Calculate the number of overlaps for each protruding grid, and calculate the ratio of the number of overlaps to the number of routes in the optimized unmanned inspection route, denoted as the overlap ratio.
[0009] In some embodiments disclosed in this invention, the method for fusing a second-order bird distribution model and a third-order bird distribution model includes: Step S303: Compare the protruding grids on the relative activity trajectory demonstration surfaces in the second-order bird distribution model and the third-order bird distribution model. If the difference in the overlap ratio between the two is less than or equal to a preset value, then the relative protruding grids are fused and associated. If the difference in the overlap ratio between the two is greater than the preset value, then analyze the preset difference range to which the difference in the overlap ratio belongs, and determine the fusion strength between the relative protruding grids based on the preset difference range to which it belongs. Step S304: Mark the protruding grids whose fusion intensity is determined, calculate the average value of the overlap ratio of the corresponding protruding grids, and map it to the protruding grids corresponding to the activity trajectory demonstration surface to obtain the reference bird distribution model.
[0010] In some embodiments disclosed in this invention, the method for determining a bird deterrent device relocation planning strategy based on a reference bird distribution model includes: Step S401: Analyze the overlap ratio and fusion intensity of the prominent grids on each activity trajectory demonstration surface in the reference bird distribution model, generate several bird deterrent device transfer strategies, and verify the actual transfer of bird deterrent devices based on the bird deterrent device transfer strategies, count the number of effective bird deterrents, and select transfer strategies based on the number of effective bird deterrents.
[0011] In some embodiments disclosed in this invention, the method for generating a transfer strategy for a bird deterrent device includes: Step S4011: Lock the protruding grids with an overlap ratio greater than or equal to a preset value on different activity trajectory demonstration surfaces, and randomly select a protruding grid from them as the position reference protruding grid. When selecting the protruding grid, it is determined that only one protruding grid can be selected within the preset range. Step S4012: Based on the overlap ratio and fusion strength of the reference protruding mesh at each location, determine the degree of strategy optimization of the transfer strategy. The expression for calculating the optimization degree of the strategy is as follows: ; Where Y represents the degree of strategy optimization. The overlap ratio of the i-th reference protruding grid. The blending strength of the i-th reference protruding mesh, This is the preset maximum fusion strength.
[0012] In some embodiments disclosed in this invention, the method for generating a transfer strategy for a bird deterrent device further includes: Step S4013: Based on the transfer strategy, perform transfer simulation on the bird deterrent device to obtain several transfer simulation results. The method for performing transfer simulation on the bird deterrent device includes: determining the number of bird deterrent devices and the transfer efficiency of the bird deterrent device; determining the time node to which each position reference protruding grid belongs in the transfer strategy; randomly selecting different position reference protruding grids for all bird deterrent devices; if the time of the bird deterrent device at the position reference protruding grid exceeds the time node and the time difference is greater than or equal to a preset value, then the bird deterrent device is transferred; when selecting the transfer, based on the transfer efficiency, calculate the arrival time to the next position reference protruding grid; if the time difference between the arrival time and the time node of the position reference protruding grid is greater than or equal to a preset value, then change to the next position protruding grid until the time difference is less than or equal to the preset value, then determine to transfer the bird deterrent device to the corresponding position protruding grid. Step S4014: Perform actual transfer verification on each transfer simulation performance of the bird deterrent device, count the number of effective bird deterrents, and determine the transfer strategy and specific transfer simulation performance based on the number of effective bird deterrents.
[0013] In some embodiments disclosed in this invention, a bird deterrent device relocation planning system based on bird activity trajectory data analysis is also disclosed, including: The first module is used to construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock size and discovery time nodes on the preset activity trajectory demonstration surface, and construct an initial bird distribution model based on the activity trajectory demonstration surface and bird flock marking data. The second module is used to perform a holistic analysis of several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, the initial UAV inspection route is optimized to obtain several optimized UAV inspection routes. The third module is used to conduct bird inspections based on the optimized UAV inspection route, and to construct a third-order bird distribution model based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are then fused to obtain a reference bird distribution model. The fourth module is used to determine the relocation planning strategy for bird deterrent devices based on a reference bird distribution model.
[0014] This invention discloses a method and system for planning the relocation of bird deterrent devices based on bird activity trajectory data analysis, belonging to the field of bird deterrent strategy planning technology. The method includes constructing an initial drone inspection route and collecting data on bird flock locations, flock sizes, and discovery times; marking bird flock data on an activity trajectory demonstration surface to construct an initial bird distribution model; performing a holistic analysis of multiple initial bird distribution models to construct a second-order bird distribution model and optimizing the drone inspection route; conducting a second inspection based on the optimized route to construct a third-order bird distribution model, which is then fused with the second-order bird distribution model to obtain a reference bird distribution model; finally, determining the bird deterrent device relocation planning strategy based on the reference bird distribution model, thereby improving the targeting and effectiveness of the bird deterrent device relocation planning.
[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the steps of the bird relocation planning method based on bird activity trajectory data analysis disclosed in this embodiment of the invention. Detailed Implementation
[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0018] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.
[0019] Example: The purpose of this invention is to provide a planning method and system for relatively efficient bird deterrence devices.
[0020] This invention discloses a method for relocation planning of bird deterrent devices based on bird activity trajectory data analysis, see reference. Figure 1 ,include: Step S100: Construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock sizes and discovery time nodes on the preset activity trajectory demonstration surface, and based on the activity trajectory demonstration surface and bird flock marking data, construct an initial bird distribution model.
[0021] The principle of step S100 is to first obtain basic observation data on bird activity within the target area through several initial drone inspection routes, and then uniformly map the bird flock locations, flock sizes, and discovery time points discovered by the drones during the inspection process onto a preset activity trajectory demonstration surface. This transforms the originally scattered and discrete bird flock observation results into analyzable data with spatial location and time node attributes. Specifically, the initial drone inspection routes cover different locations within the monitored area to avoid bias in the observation area caused by a single route; the activity trajectory demonstration surface carries bird activity information and can be understood as a spatial representation surface corresponding to the actual inspection area. The bird flock location determines the spatial unit where bird activity occurs, the flock size reflects the degree of bird aggregation at that location, and the discovery time point reflects the temporal distribution characteristics of bird activity. By continuously recording the above bird flock marking data on the activity trajectory demonstration surface, an initial bird distribution model can be formed. This model can initially express the location, time, and aggregation intensity of bird flocks within the target area, providing basic data support for subsequent identification of high-incidence bird areas, optimization of drone inspection routes, and planning of bird deterrent device relocation locations.
[0022] Step S200: Perform a holistic analysis on several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, optimize the initial UAV inspection route to obtain several optimized UAV inspection routes.
[0023] The principle of step S200 lies in performing a holistic analysis of multiple initial bird distribution models to extract relatively stable bird activity patterns from bird activity data obtained from multiple drone inspections. Initial bird distribution models obtained from a single inspection may be affected by factors such as weather, noise, temporary food sources, the timing of drone inspections, and occasional bird migrations, thus the bird distribution results they reflect may contain randomness and local biases. By uniformly comparing several initial bird distribution models, the recurrence of bird flock marker areas in spatial location and time node in different inspection results can be determined, and areas with high bird activity frequency and relatively stable occurrence times can be identified, thereby constructing a second-order bird distribution model. Compared to a single initial bird distribution model, the second-order bird distribution model can better reflect the stable distribution characteristics and regular activity trends of bird flocks within the target area. Based on this, the initial UAV inspection route is further optimized according to the second-order bird distribution model, so that the UAV inspection route prioritizes covering areas with a high probability of repeated bird flocks, large bird flocks, or high matching degree with specific time nodes, thereby reducing repeated inspections of low-value areas, improving the effectiveness of UAV in detecting bird flocks and the targeting of subsequent data collection.
[0024] Step S300: Bird inspection is carried out based on the optimized UAV inspection route, and a third-order bird distribution model is constructed based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are fused to obtain a reference bird distribution model.
[0025] The principle of step S300 is to collect bird activity data again using the optimized drone inspection route and integrate the new inspection results with the previously formed second-order bird distribution model to improve the accuracy and real-time adaptability of the bird distribution judgment results. Since the second-order bird distribution model mainly comes from several initial inspection data points, it can reflect the stable patterns of bird flock activity in historical inspections. However, bird flock activity itself has dynamic characteristics and may shift with changes in season, time of day, environmental disturbances, food sources, or bird deterrence interventions. Therefore, relying solely on historical models may lead to delays in the planning of bird deterrence device relocation. By conducting a second inspection based on the optimized drone inspection route, bird flock marking data that better reflects the current bird activity status can be obtained, and a third-order bird distribution model can be constructed accordingly. The third-order bird distribution model reflects the bird flock distribution results re-acquired under optimized inspection conditions. The fusion of second-order and third-order bird distribution models essentially involves a comprehensive assessment of historical stable distribution patterns and current updated distribution results. When the two models show consistency in the same or similar spatial locations and time points, the credibility of the corresponding area as a high-incidence area for bird populations can be improved. When the two models differ, the degree of difference can be used to determine the trend of bird activity changes, thereby obtaining a more suitable reference bird distribution model for bird deterrence device relocation decisions.
[0026] Step S400: Based on the reference bird distribution model, determine the bird deterrent device relocation planning strategy.
[0027] The principle of step S400 lies in determining the relocation planning strategy for bird deterrent devices based on a reference bird distribution model. This allows the deployment or relocation of bird deterrent devices to move beyond simply relying on fixed locations, human experience, or passive responses after bird flocks are discovered. Instead, it involves proactive planning based on the activity patterns of bird flocks at different spatial locations and time points. The reference bird distribution model contains information such as the location of bird flock appearances, flock size, time points, historical stability, and current trends. Therefore, it can be used to determine which areas are key areas where bird flocks are likely to reappear or concentrate their activities, and when these areas require bird deterrent device coverage. When determining the relocation planning strategy for bird deterrent devices, candidate relocation locations can be identified based on areas with strong bird activity in the reference bird distribution model, and candidate relocation timings can be determined based on corresponding time points. Furthermore, by combining the number of bird deterrent devices, their current locations, relocation efficiency, coverage area, and the importance of different candidate areas, the relocation sequence and target locations of the bird deterrent devices are generated. This method allows bird deterrence devices to be prioritized for relocation to areas with a higher probability of bird activity, larger bird populations, or more stable model fusion results. This reduces instances of ineffective device deployment, delayed arrival, or repeated coverage of low-risk areas, thereby improving the accuracy, timeliness, and utilization rate of bird deterrence device relocation planning.
[0028] In some embodiments disclosed in this invention, the method for constructing an initial bird distribution model based on the activity trajectory demonstration surface and bird flock marking data includes: Step S101: Divide the activity trajectory demonstration surface into a grid and construct several layers of activity trajectory demonstration surfaces above the activity trajectory demonstration surface. Each layer of activity trajectory demonstration surface corresponds to a preset time node.
[0029] The principle of step S101 lies in converting the actual inspection area into a spatiotemporal grid structure that facilitates calculation and comparison. Specifically, the activity trajectory demonstration surface can be understood as a spatial representation surface corresponding to the area to be driven away from birds. By dividing the activity trajectory demonstration surface into grids, the continuous actual geographical area can be divided into multiple grid units with definite location ranges, so that the location of bird flocks discovered by UAVs can be mapped to specific grids, thereby facilitating subsequent statistical analysis, overlap comparison, and regional correlation analysis. At the same time, several layers of activity trajectory demonstration surfaces are constructed above the activity trajectory demonstration surface, with each layer corresponding to a preset time node. In essence, this expands the bird activity data from a single spatial distribution to a multi-layer spatiotemporal distribution structure of "spatial location + time node". Since bird flock activities usually have obvious temporal variation characteristics, the probability of bird flocks appearing at the same location at different time nodes may be different. Therefore, by setting up multiple layers of activity trajectory demonstration surfaces, the appearance of bird flocks in each grid at different time nodes can be recorded separately, avoiding the judgment bias caused by mixing bird flock activities from different time periods in the same plane. This processing enables subsequent steps to compare the activity intensity of bird flocks in different grids within the same time frame, and to analyze the migration changes of bird flock activity areas between different time frames, thus providing a unified data carrying structure for the construction of the initial bird distribution model.
[0030] Step S102: Determine the location of the bird flock discovered by the drone, and highlight the corresponding grid on the activity trajectory demonstration surface based on the location of the bird flock to obtain a highlighted grid group. Based on the discovery time node, map the highlighted grid group onto the corresponding activity trajectory demonstration surface and mark the size of the bird flock.
[0031] The principle of step S102 is to transform the bird flock observation results detected by the UAV into prominent grid data that can be used for overlap ratio calculation. Specifically, when the UAV detects a bird flock during its inspection, it first determines the location of the flock in the actual inspection area. Then, based on the correspondence between this location and the grid cells in the activity trajectory demonstration surface, it identifies one or more grids where the flock is located and highlights the corresponding grids, resulting in a prominent grid group. This prominent grid group indicates that bird activity has been detected in the corresponding spatial area, and marking the flock size further reflects the degree of bird aggregation in that area. At the same time, based on the discovery time node corresponding to the bird flock, the prominent grid group is mapped to the activity trajectory demonstration surface layer corresponding to the corresponding time node, so that the location, size, and time of the bird flock are correlated. In this way, each drone inspection can generate an initial bird distribution model with a time layer and prominent grid markings. When multiple initial bird distribution models are analyzed in a unified manner, if the same grid in the same time layer is marked as a prominent grid multiple times, the overlap ratio of the grid increases. This indicates that the reproducibility of bird flocks in the area at the corresponding time node is relatively strong, suggesting that the bird flocks in the area are not just occasional occurrences, but more likely belong to stable or high-frequency activity areas of bird flocks. This provides a basis for the subsequent construction of second-order bird distribution models, optimization of drone inspection routes, and determination of bird deterrent device relocation planning strategies.
[0032] In some embodiments disclosed in this invention, the method for performing a holistic analysis of several initial bird distribution models includes: Step S201: Map the representation of each activity trajectory demonstration surface in all initial bird distribution models into the bird distribution model for analysis.
[0033] The principle of step S201 is to unify the bird activity performance in multiple initial bird distribution models into a single analytical benchmark, so as to compare the recurrence between different inspection results. Since each initial bird distribution model originates from one or a group of drone inspections, it has already formed multiple layers of activity trajectory demonstration surfaces according to time nodes, and the location and size of bird flocks are marked in the corresponding grids. However, without a unified mapping between different initial bird distribution models, it is difficult to determine whether the same time layer and the same spatial grid have been repeatedly marked in multiple inspections. Therefore, this step maps the performance of each layer of activity trajectory demonstration surfaces in all initial bird distribution models to the bird distribution model used for analysis, so that the same time nodes and the same grid positions under different inspection batches can be within the same analytical coordinate system. Through this unified mapping process, the prominent grid data scattered in multiple initial bird distribution models can be centrally expressed, providing a unified data foundation for subsequent statistics on the overlap of each grid, calculation of the overlap ratio, and judgment of the recurrence of bird flock appearances.
[0034] Step S202: Analyze the overlapping of prominent grids in each layer of the activity trajectory demonstration surface in the bird distribution model for analysis, determine the number of overlaps for each grid, and calculate the overlap ratio between the number of overlaps and the number of initial bird distribution models.
[0035] The principle of step S202 is to quantify the reproducibility of bird flocks appearing in each grid area at the corresponding time node by statistically analyzing the number of times prominent grids appear in multiple initial bird distribution models. Specifically, in the bird distribution model used for analysis, each layer of activity trajectory demonstration surface is analyzed separately to determine the number of times each grid in the same time layer is marked as a prominent grid in multiple initial bird distribution models, and this number is recorded as the overlap number. Subsequently, the overlap number is calculated as a ratio to the number of initial bird distribution models involved in the analysis to obtain the overlap ratio of the grid. The higher the overlap ratio, the more times the grid is repeatedly marked as a prominent grid in multiple inspections or multiple initial bird distribution models, that is, the stronger the reproducibility of bird flocks appearing in the actual area corresponding to the grid at the corresponding time node, and the more likely it belongs to a stable or high-frequency activity area of bird flocks; conversely, the lower the overlap ratio, the more likely the appearance of bird flocks in the area is an accidental occurrence or a short-term disturbance. Therefore, this step can improve the bird flock activity from a single observation result to a reproducible judgment result under multiple observations, providing a quantitative basis for subsequent selection of key grid areas and construction of a second-order bird distribution model.
[0036] Step S203: Compare the overlap ratio of the protruding grids in each layer of the activity trajectory demonstration surface with a threshold interval, mark the protruding grids belonging to the same threshold interval, and associate the corresponding protruding grids that are less than or equal to a preset value and have the same interval marking to form several second-order grid clusters. The common performance of several layers of activity trajectory demonstration surfaces is recorded as a second-order bird distribution model.
[0037] The principle of step S203 lies in classifying and spatially associating prominent grids with similar bird flock reproducibility based on overlap ratio, thereby forming a second-order grid cluster that reflects the stable activity area of bird flocks. Specifically, in each layer of the activity trajectory demonstration surface, the overlap ratio of different prominent grids may vary. Grids with a higher overlap ratio indicate stronger bird flock reproducibility in that area, while grids with a lower overlap ratio indicate weaker bird flock reproducibility. By comparing the overlap ratio of each prominent grid with a preset threshold interval, the prominent grids can be divided into different reproducibility levels, and prominent grids belonging to the same threshold interval are marked, so that grids with similar bird flock reproducibility are assigned the same or corresponding analysis labels. On this basis, the spatial distance between prominent grids is further determined. When the distance between them is less than or equal to a preset value, and the corresponding prominent grids have the same interval label, it indicates that these grids are not only spatially adjacent to each other, but also have similar performance in terms of bird flock reproducibility. Therefore, they can be associated to form a second-order grid cluster. Second-order grid clusters are not simply single, randomly appearing grids of bird flocks, but rather consist of multiple spatially adjacent salient grids with similar reproducibility levels. Therefore, they can more stably represent high-frequency or stable activity areas of bird flocks at specific time points. By summarizing the second-order grid clusters formed in several layers of activity trajectory demonstration surfaces and their temporal representations, a second-order bird distribution model is obtained. This model can simultaneously reflect the spatial aggregation characteristics, temporal distribution characteristics, and reproducibility characteristics of bird flock activities under multiple inspections, providing a more reliable basis for subsequent UAV inspection route optimization and bird deterrence device relocation planning.
[0038] In some embodiments disclosed in this invention, the method for optimizing the initial UAV inspection route based on a second-order bird distribution model includes: Step S2031: Analyze the second-order grid clusters in each activity estimation demonstration surface of the second-order bird distribution model, including the location node to which the cluster belongs, the threshold interval to which the cluster belongs, and the time node to which the cluster belongs.
[0039] The principle of step S2031 lies in extracting key constraint information from the second-order bird distribution model that can be used for UAV inspection route optimization. The second-order grid clusters in the second-order bird distribution model are formed by the association of prominent grids that are spatially close and have the same threshold interval for overlap. Essentially, they represent a set of regions where bird flocks exhibit similar reproducibility at corresponding time points. Therefore, when analyzing the second-order grid clusters in each layer of the activity trajectory demonstration surface, it is necessary to determine the location node, threshold interval, and time node to which the cluster belongs. Specifically, the location node indicates the spatial target location that the UAV needs to focus on inspecting; the threshold interval indicates the strength of bird flock reproducibility within the cluster; and the time node indicates the time range within which birds are more likely to appear in the corresponding area. By simultaneously extracting the location node, threshold interval, and time node, the bird flock reproducibility information in the second-order bird distribution model can be transformed into target point information, priority information, and time constraint information required for UAV route optimization. This allows subsequent UAV inspection routes to no longer simply cover fixed areas but to prioritize planning around areas with high bird flock reproducibility and clear time characteristics.
[0040] Step S2032: Using the time node to which the cluster belongs as a constraint, randomly select the location node to which the cluster belongs, and construct the UAV inspection route. If the arrival time node is earlier than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then add other inspection paths before the location node to which the cluster belongs. If the arrival time node is later than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then remove the corresponding location node to which the cluster belongs.
[0041] The principle of step S2032 is to use the time node corresponding to the second-order grid cluster as the inspection timing constraint to filter, sort, and correct the position nodes in the UAV inspection route, so that the UAV arrives at the corresponding cluster area during the time period with a high probability of bird recurrence. Specifically, since the time node of the second-order grid cluster reflects the time characteristics of repeated bird recurrence in the area, when constructing the UAV inspection route, it is not only necessary to consider the spatial distance between position nodes, but also whether the arrival time of the UAV to the position node of each cluster matches the time node of that cluster. When the expected arrival time of the UAV at a position node of a certain cluster is earlier than the time node of that cluster, and the time difference between the two is greater than or equal to a preset value, it means that the UAV has arrived in the area too early. At this time, the bird flock has not yet entered the high recurrence activity period, and direct inspection may lead to invalid inspection. Therefore, other inspection paths are added before the position node of that cluster, so that the UAV inspects other candidate areas first, and then arrives at that position when it is closer to the time node of that cluster. Conversely, if the estimated arrival time of a drone at a node belonging to a certain bird cluster is later than the time node to which the cluster belongs, and the time difference between the two is greater than or equal to a preset value, it indicates that the period of high reproducibility of bird activity may have passed by the time the drone arrives. This means the inspection value of that node decreases, and therefore the corresponding node is removed to avoid wasting inspection resources. Through this process, the optimized drone inspection route can spatially prioritize covering second-order grid clusters with high bird reproducibility and temporally match the corresponding bird activity time nodes as much as possible. This improves the effectiveness of drones in detecting bird flocks and provides more accurate and targeted inspection data for the subsequent construction of third-order bird distribution models.
[0042] In some embodiments disclosed in this invention, the method for constructing a third-order bird distribution model includes: Step S301: Based on the bird flock marking data, identify the prominent grid groups on each activity trajectory demonstration surface, and mark the corresponding bird flock size on the prominent grid groups.
[0043] The principle of step S301 lies in re-acquiring bird flock marking data based on the optimized UAV inspection route and transforming this bird flock marking data into prominent grid cluster data required by the third-order bird distribution model. Specifically, the UAV inspection route optimized by the second-order bird distribution model has prioritized second-order grid clusters, corresponding location nodes, and corresponding time nodes with high bird flock reproducibility. Therefore, when conducting bird inspections again based on this optimized route, the UAV can more effectively cover areas with high bird flock reproducibility and key time periods. When the UAV discovers a bird flock during the optimized inspection, it determines the corresponding grid on each activity trajectory demonstration surface based on the bird flock's location and highlights the corresponding grid, forming prominent grid clusters. Simultaneously, the corresponding bird flock size is marked on the prominent grid clusters to characterize the degree of bird aggregation in the actual area corresponding to the prominent grid cluster. Through this processing, the third-order bird distribution model can not only record the location and time of bird flock appearance during the optimized inspection but also record the bird flock size of different prominent grid clusters, thereby reflecting the current distribution performance of bird flocks obtained after route optimization and providing a data basis for subsequent calculation of the prominent grid overlap ratio.
[0044] Step S302: Calculate the number of overlaps for each protruding grid, and calculate the ratio of the number of overlaps to the number of routes in the optimized unmanned inspection route, denoted as the overlap ratio.
[0045] In some embodiments disclosed in this invention, the method for fusing a second-order bird distribution model and a third-order bird distribution model includes: Step S303: Compare the protruding grids on the relative activity trajectory demonstration surfaces in the second-order bird distribution model and the third-order bird distribution model. If the difference in the overlap ratio between the two is less than or equal to a preset value, then the relative protruding grids are fused and associated. If the difference in the overlap ratio between the two is greater than the preset value, then analyze the preset difference range to which the difference in the overlap ratio belongs, and determine the fusion strength between the relative protruding grids based on the preset difference range to which it belongs.
[0046] Step S303 involves comparing the historical reproducibility results reflected in the second-order bird distribution model with the current reproducibility results after optimized inspection reflected in the third-order bird distribution model to determine whether bird flock activities in the same spatial area and at the same time point have stable continuity. Specifically, the overlap ratio of prominent grids in the second-order bird distribution model mainly comes from the overall analysis of multiple initial bird distribution models. The higher the overlap ratio, the stronger the reproducibility of bird flocks in the corresponding area in previous rounds of inspections. The overlap ratio of prominent grids in the third-order bird distribution model comes from the re-inspection results under the optimized UAV inspection route, which can reflect the reproducibility of bird flocks in the corresponding area under the current optimized inspection conditions. By comparing the prominent grids on the relative activity trajectory demonstration surfaces in the second-order and third-order bird distribution models, it can be determined whether the reproducibility of the same or corresponding grids in historical inspection results is consistent with the current inspection results. When the difference in overlap ratio between the two is less than or equal to a preset value, it indicates that the area corresponding to the grid shows similar bird flock recurrence in both historical and current optimized inspections, and the bird activity patterns in this area are relatively stable. Therefore, the relatively prominent grids are fused and correlated to indicate that this area can be considered a highly reliable candidate area in the subsequent bird deterrence device relocation plan. When the difference in overlap ratio between the two is greater than a preset value, it indicates that there is a significant difference between the historical and current recurrence in this area, which may be affected by seasonality, weather, food source changes, environmental disturbances, or previous bird deterrence measures. In this case, it is not advisable to make a judgment directly based on a single model. Instead, it is necessary to further analyze the preset difference range to which the difference in overlap ratio belongs, and determine the fusion strength between the relatively prominent grids based on the difference range. The fusion strength is used to represent the degree of consistency or reliable fusion between the second-order model and the third-order model on the corresponding grids. The smaller the difference, the higher the fusion strength; the larger the difference, the lower the fusion strength. This process avoids delays in relocation planning due to reliance solely on historical models, and also avoids accidental misjudgments due to reliance solely on current or limited optimization inspection results, thereby improving the stability and real-time adaptability of the reference bird distribution model.
[0047] Step S304: Mark the protruding grids whose fusion intensity is determined, calculate the average value of the overlap ratio of the corresponding protruding grids, and map it to the protruding grids corresponding to the activity trajectory demonstration surface to obtain the reference bird distribution model.
[0048] The principle of step S304 is to write the fusion result of the fusion intensity and the corresponding overlap ratio of the protruding grids determined in step S303 into a unified activity trajectory demonstration surface to form a reference bird distribution model for the bird deterrent device's transfer decision. Specifically, in step S303, the corresponding protruding grids in the second-order and third-order bird distribution models have completed consistency judgment, and the fusion intensity has been determined based on the difference in overlap ratio. This step further marks the protruding grids whose fusion intensity has been determined, so that each fused protruding grid not only has positional and time node attributes, but also a fusion intensity attribute representing the consistency or credible fusion degree of the model. At the same time, the average overlap ratio of the corresponding protruding grids in the second-order and third-order bird distribution models is calculated, and this average value is mapped to the protruding grids corresponding to the activity trajectory demonstration surface as a reference overlap ratio of the grid in the reference bird distribution model. The reference overlap ratio integrates the reproducibility of bird flocks under historical multi-round inspections and the current reproducibility of bird flocks under optimized inspections, providing a more balanced representation of the probability of bird flocks appearing in the corresponding area at the corresponding time point. The fusion strength further reflects the consistency between the second-order and third-order models, ensuring that subsequent bird deterrence device relocation planning considers not only the reproducibility of bird flocks in a particular grid but also the reliability of that reproducibility assessment. By mapping the reference overlap ratio, fusion strength, spatial grid location, and time point to the activity trajectory demonstration surface, a reference bird distribution model can be obtained. This model provides a more stable and accurate data basis for subsequently determining the relocation location, timing, and priority of bird deterrence devices.
[0049] In some embodiments disclosed in this invention, the method for determining a bird deterrent device relocation planning strategy based on a reference bird distribution model includes: Step S401: Analyze the overlap ratio and fusion intensity of the prominent grids on each activity trajectory demonstration surface in the reference bird distribution model, generate several bird deterrent device transfer strategies, and verify the actual transfer of bird deterrent devices based on the bird deterrent device transfer strategies, count the number of effective bird deterrents, and select transfer strategies based on the number of effective bird deterrents.
[0050] In some embodiments disclosed in this invention, the method for generating a transfer strategy for a bird deterrent device includes: Step S4011: Lock the protruding grids with an overlap ratio greater than or equal to a preset value on different activity trajectory demonstration surfaces, and randomly select a protruding grid from them as the position reference protruding grid. When selecting the protruding grid, it is determined that only one protruding grid can be selected within the preset range. Step S4012: Based on the overlap ratio and fusion strength of the reference protruding mesh at each location, determine the degree of strategy optimization of the transfer strategy. The expression for calculating the optimization degree of the strategy is as follows: ; Where Y represents the degree of strategy optimization. The overlap ratio of the i-th reference protruding grid. The blending strength of the i-th reference protruding mesh, This is the preset maximum fusion strength.
[0051] In some embodiments disclosed in this invention, the method for generating a transfer strategy for a bird deterrent device further includes: Step S4013: Based on the transfer strategy, a transfer simulation is performed on the bird-repelling device to obtain several transfer simulation results. The method for performing the transfer simulation on the bird-repelling device includes: determining the number of bird-repelling devices and the transfer efficiency of the bird-repelling devices; determining the time node to which each position reference protruding grid belongs in the transfer strategy; randomly selecting different position reference protruding grids for all bird-repelling devices; if the time of the bird-repelling device at the position reference protruding grid exceeds the time node and the time difference is greater than or equal to a preset value, then the bird-repelling device is transferred; when selecting the transfer, the arrival time to the next position reference protruding grid is calculated based on the transfer efficiency; if the time difference between the arrival time and the time node of the position reference protruding grid is greater than or equal to a preset value, then the next position protruding grid is changed until the time difference is less than or equal to the preset value, then it is determined that the bird-repelling device is transferred to the corresponding position protruding grid.
[0052] Step S4014: Perform actual transfer verification on each transfer simulation performance of the bird deterrent device, count the number of effective bird deterrents, and determine the transfer strategy and specific transfer simulation performance based on the number of effective bird deterrents.
[0053] In some embodiments disclosed in this invention, a bird deterrent device relocation planning system based on bird activity trajectory data analysis is also disclosed, including: The first module is used to construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock sizes and discovery time points on the preset activity trajectory demonstration surface, and construct an initial bird distribution model based on the activity trajectory demonstration surface and bird flock marking data.
[0054] The second module is used to perform a holistic analysis of several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, the initial UAV inspection route is optimized to obtain several optimized UAV inspection routes.
[0055] The third module is used to conduct bird inspections based on the optimized UAV inspection route and to construct a third-order bird distribution model based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are then fused to obtain a reference bird distribution model.
[0056] The fourth module is used to determine the relocation planning strategy for bird deterrent devices based on a reference bird distribution model.
[0057] This invention discloses a method and system for planning the relocation of bird deterrent devices based on bird activity trajectory data analysis, belonging to the field of bird deterrent strategy planning technology. The method includes constructing an initial drone inspection route and collecting data on bird flock locations, flock sizes, and discovery times; marking bird flock data on an activity trajectory demonstration surface to construct an initial bird distribution model; performing a holistic analysis of multiple initial bird distribution models to construct a second-order bird distribution model and optimizing the drone inspection route; conducting a second inspection based on the optimized route to construct a third-order bird distribution model, which is then fused with the second-order bird distribution model to obtain a reference bird distribution model; finally, determining the bird deterrent device relocation planning strategy based on the reference bird distribution model, thereby improving the targeting and effectiveness of the bird deterrent device relocation planning.
[0058] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for relocation planning of bird deterrent devices based on bird activity trajectory data analysis, characterized in that, include: Step S100: Construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock size and discovery time nodes on the preset activity trajectory demonstration surface, and based on the activity trajectory demonstration surface and bird flock marking data, construct an initial bird distribution model. Step S200: Perform a holistic analysis on several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, optimize the initial UAV inspection route to obtain several optimized UAV inspection routes. Step S300: Bird inspection is carried out based on the optimized UAV inspection route, and a third-order bird distribution model is constructed based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are fused to obtain a reference bird distribution model. Step S400: Based on the reference bird distribution model, determine the bird deterrent device relocation planning strategy.
2. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 1, characterized in that, Methods for constructing an initial bird distribution model based on activity trajectory demonstration surfaces and bird flock marking data include: Step S101: Divide the activity trajectory demonstration surface into a grid and construct several layers of activity trajectory demonstration surfaces above the activity trajectory demonstration surface. Each layer of activity trajectory demonstration surface corresponds to a preset time node. Step S102: Determine the location of the bird flock discovered by the drone, and highlight the corresponding grid on the activity trajectory demonstration surface based on the location of the bird flock to obtain a highlighted grid group. Based on the discovery time node, map the highlighted grid group onto the corresponding activity trajectory demonstration surface and mark the size of the bird flock.
3. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 2, characterized in that, Methods for conducting a holistic analysis of several initial bird distribution models include: Step S201: Map the representation of each activity trajectory demonstration surface in all initial bird distribution models into the bird distribution model for analysis in a unified manner. Step S202: Analyze the overlapping of prominent grids in each layer of the activity trajectory demonstration surface in the bird distribution model for analysis, determine the number of overlaps of each grid, and calculate the overlap ratio between the number of overlaps and the number of initial bird distribution models. Step S203: Compare the overlap ratio of the protruding grids in each layer of the activity trajectory demonstration surface with a threshold interval, mark the protruding grids belonging to the same threshold interval, and associate the corresponding protruding grids that are less than or equal to a preset value and have the same interval marking to form several second-order grid clusters. The common performance of several layers of activity trajectory demonstration surfaces is recorded as a second-order bird distribution model.
4. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 3, characterized in that, Methods for optimizing the initial UAV inspection route based on a second-order bird distribution model include: Step S2031 involves analyzing the second-order grid clusters in each activity estimation demonstration surface of the second-order bird distribution model, including the location node of the cluster, the threshold interval to which the cluster belongs, and the time node to which the cluster belongs. Step S2032: Using the time node to which the cluster belongs as a constraint, randomly select the location node to which the cluster belongs, and construct the UAV inspection route. If the arrival time node is earlier than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then add other inspection paths before the location node to which the cluster belongs. If the arrival time node is later than the time node to which the cluster belongs and the time difference between the two is greater than or equal to a preset value, then remove the corresponding location node to which the cluster belongs.
5. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 3, characterized in that, Methods for constructing third-order bird distribution models include: Step S301: Based on the bird flock marking data, identify the prominent grid groups on each activity trajectory demonstration surface, and mark the corresponding bird flock size on the prominent grid groups; Step S302: Calculate the number of overlaps for each protruding grid, and calculate the ratio of the number of overlaps to the number of routes in the optimized unmanned inspection route, denoted as the overlap ratio.
6. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 5, characterized in that, Methods for fusing second-order and third-order bird distribution models include: Step S303: Compare the protruding grids on the relative activity trajectory demonstration surfaces in the second-order bird distribution model and the third-order bird distribution model. If the difference in the overlap ratio between the two is less than or equal to a preset value, then the relative protruding grids are fused and associated. If the difference in the overlap ratio between the two is greater than the preset value, then analyze the preset difference range to which the difference in the overlap ratio belongs, and determine the fusion strength between the relative protruding grids based on the preset difference range to which it belongs. Step S304: Mark the protruding grids whose fusion intensity is determined, calculate the average value of the overlap ratio of the corresponding protruding grids, and map it to the protruding grids corresponding to the activity trajectory demonstration surface to obtain the reference bird distribution model.
7. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 6, characterized in that, Methods for determining bird deterrence device relocation planning strategies based on reference bird distribution models include: Step S401: Analyze the overlap ratio and fusion intensity of the prominent grids on each activity trajectory demonstration surface in the reference bird distribution model, generate several bird deterrent device transfer strategies, and verify the actual transfer of bird deterrent devices based on the bird deterrent device transfer strategies, count the number of effective bird deterrents, and select transfer strategies based on the number of effective bird deterrents.
8. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 7, characterized in that, Methods for generating transfer strategies for bird deterrent devices include: Step S4011: Lock the protruding grids with an overlap ratio greater than or equal to a preset value on different activity trajectory demonstration surfaces, and randomly select a protruding grid from them as the position reference protruding grid. When selecting the protruding grid, it is determined that only one protruding grid can be selected within the preset range. Step S4012: Based on the overlap ratio and fusion strength of the reference protruding mesh at each location, determine the degree of strategy optimization of the transfer strategy. The expression for calculating the optimization degree of the strategy is as follows: ； Where Y represents the degree of strategy optimization. The overlap ratio of the i-th reference protruding grid. The blending strength of the i-th reference protruding mesh, This is the preset maximum fusion strength.
9. The bird deterrent device relocation planning method based on bird activity trajectory data analysis according to claim 8, characterized in that, Methods for generating transfer strategies for bird deterrent devices also include: Step S4013: Based on the transfer strategy, perform transfer simulation on the bird deterrent device to obtain several transfer simulation results. The method for performing transfer simulation on the bird deterrent device includes: determining the number of bird deterrent devices and the transfer efficiency of the bird deterrent device; determining the time node to which each position reference protruding grid belongs in the transfer strategy; randomly selecting different position reference protruding grids for all bird deterrent devices; if the time of the bird deterrent device at the position reference protruding grid exceeds the time node and the time difference is greater than or equal to a preset value, then the bird deterrent device is transferred; when selecting the transfer, calculate the arrival time to the next position reference protruding grid based on the transfer efficiency; if the time difference between the arrival time and the time node of the position reference protruding grid is greater than or equal to a preset value, then change to the next position protruding grid until the time difference is less than or equal to the preset value, then determine to transfer the bird deterrent device to the corresponding position protruding grid. Step S4014: Perform actual transfer verification on each transfer simulation performance of the bird deterrent device, count the number of effective bird deterrents, and determine the transfer strategy and specific transfer simulation performance based on the number of effective bird deterrents.
10. A bird deterrent device relocation planning system based on bird activity trajectory data analysis, characterized in that, A method for implementing the bird deterrent device relocation planning method according to any one of claims 1-9, comprising: The first module is used to construct several initial drone inspection routes, and based on the initial drone inspection routes, drive the drone to conduct bird inspections, and mark the discovered bird flocks, flock size and discovery time nodes on the preset activity trajectory demonstration surface, and construct an initial bird distribution model based on the activity trajectory demonstration surface and bird flock marking data. The second module is used to perform a holistic analysis of several initial bird distribution models, and based on the analysis results, construct a second-order bird distribution model. Based on the second-order bird distribution model, the initial UAV inspection route is optimized to obtain several optimized UAV inspection routes. The third module is used to conduct bird inspections based on the optimized UAV inspection route, and to construct a third-order bird distribution model based on bird flock marking data. The second-order bird distribution model and the third-order bird distribution model are then fused to obtain a reference bird distribution model. The fourth module is used to determine the relocation planning strategy for bird deterrent devices based on a reference bird distribution model.