Rice topdressing decision method and system based on dynamic growth rate of unmanned aerial vehicle group

Abstract

The application discloses a rice topdressing decision method and system based on dynamic growth and decline rate of unmanned aerial vehicle group, relates to the field of agricultural remote sensing technology, and comprises the following steps: aligning the crown layer nitrogen nutrition change time sequence by using the group dynamic growth and decline rate graph, distinguishing the natural convergence and the recoverable nitrogen limitation type attenuation according to the chronological relationship between the growth and decline turning point and the crown layer nitrogen nutrition state decline, and forming a group growth and decline driving judgment graph; screening the nitrogen limitation area with recovery potential according to the group growth and decline driving judgment graph, combining the degree of insufficient tillering expansion and the tillering attenuation trend of the rice group to determine the topdressing priority and the topdressing intensity, and forming a rice topdressing decision prescription graph. According to the chronological relationship between the growth and decline turning point and the crown layer nitrogen nutrition state decline, the natural convergence and the recoverable nitrogen limitation type attenuation are distinguished, the group growth and decline driving judgment graph is formed, and the topdressing priority and the topdressing intensity are determined accordingly, so that the pertinence and rationality of the rice topdressing decision are improved.

Description

Technical Field

[0001] This invention relates to the field of agricultural remote sensing technology, and in particular to a method and system for making decisions on topdressing rice based on the dynamic growth and decline rate of unmanned aerial vehicle (UAV) swarms. Background Technology

[0002] Topdressing decisions for rice are typically made by combining field growth monitoring, plant nitrogen nutrition diagnosis, and variable fertilization techniques. Existing methods utilize drones equipped with visible light or multispectral sensors to acquire canopy images. The growth status of the rice population is characterized by vegetation indices, canopy coverage, plant height, and chlorophyll-related features. These characteristics are then combined with the growth stage, target yield, or empirical thresholds to determine the topdressing area and amount, generating spatialized fertilization prescriptions that provide a basis for precision nutrient management in paddy fields.

[0003] Existing methods for determining continuous population changes still mainly focus on differences in quantity, spectrum, and growth at a certain moment or adjacent moments, making it difficult to distinguish between population states with similar net growth and decline but different degrees of synchronization between internal growth and decline. At the same time, when a population transitions from expansion to decline, judging solely based on a decrease in canopy nitrogen nutrition or a reduction in tillering makes it difficult to further distinguish between natural convergence and nitrogen-limited decline that still has the potential for recovery, thus affecting the determination of the target and intensity of topdressing. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a rice topdressing decision-making method based on the dynamic growth and decline rate of UAV populations to solve the problem that existing technologies are unable to accurately distinguish the true growth and decline state of rice populations and the type of decline drive.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for rice topdressing decision-making based on the dynamic growth and decline rate of unmanned aerial vehicle (UAV) swarms, comprising:

[0008] Collect drone field patrol image data, uniformly register the spatial location of rice population at different observation times, and complete cross-time alignment according to fixed population evaluation units to generate a temporal observation volume of rice population;

[0009] Read the time series observation data of rice populations, track the cross-time continuity relationship of rice populations within the same fixed population evaluation unit, identify the real increase and effective decline of rice tillers, and exclude false decline caused by shading, and generate population growth and decline correlation chains;

[0010] The process of addition and decline in the population growth and decline correlation chain is tracked simultaneously. By identifying the dynamic process of population expansion to balance and decline through the net growth and decline of rice tillers and the population turnover status, a population dynamic growth and decline rate map is generated.

[0011] By using the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes, and by distinguishing between natural convergence and recoverable nitrogen-limited decay based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, a population growth and decline driving determination map is formed.

[0012] Based on the population growth and decline driving determination map, nitrogen-limited regions with recovery potential were screened, and the priority and intensity of topdressing were determined by combining the degree of insufficient tiller expansion and tiller decline trend of rice population, thus forming a rice topdressing decision prescription map.

[0013] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms described in this invention, the UAV field inspection image data includes the spatial location of rice tillers, canopy boundary, canopy coverage, canopy height, red light reflectance, red edge reflectance, near-infrared reflectance, and corresponding acquisition time and geographical coordinates.

[0014] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms described in this invention, the steps of uniformly registering the spatial locations of rice swarms at different observation times and completing cross-time alignment according to fixed swarm evaluation units to generate a temporal observation volume of the paddy field swarm are as follows:

[0015] The drone field patrol image data at each observation time is converted to a unified geographic coordinate benchmark, and the overall positional offset between drone field patrol images of different sorties is corrected based on the geographic coordinate information.

[0016] Using the first valid observation time of the UAV field inspection image as the spatial reference, the boundary of the rice population and stable ground features in the subsequent UAV field inspection image data are matched to correct the local registration deviation of the same spatial location at different observation times.

[0017] Fixed population evaluation units are divided according to the spatial range of the target paddy field after unified registration, and the spatial boundaries and location markers of each fixed population evaluation unit are established without change.

[0018] Based on the location identifier of the fixed population evaluation unit, the rice population observation information within the corresponding range of each observation time is extracted, and a continuous correspondence is established according to the collection time to generate the rice field population time series observation body.

[0019] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV populations described in this invention, the step of tracking the cross-time continuity relationship of rice populations within the same fixed population evaluation unit includes the following steps:

[0020] Read the time-series observation data of the paddy field population, extract the rice population observation information corresponding to each observation time according to the location identifier of the fixed population evaluation unit, and form a continuous observation sequence according to the collection time;

[0021] Spatial correspondence is established for rice population observation information at adjacent observation times within the same fixed population evaluation unit, and the cross-temporal continuity relationship of the rice population is determined based on the continuous change of the spatial location of rice tillers.

[0022] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV populations described in this invention, the method distinguishes between the actual increase and effective decline of rice tillers, while excluding false declines caused by shading, and generates a population growth and decline correlation chain. The steps are as follows:

[0023] Based on the cross-time continuity relationship of rice populations, rice tillers that did not appear at the previous observation time but appeared at the next observation time and continued to exist in subsequent observations were identified as truly newly added rice tillers.

[0024] Backtracking was performed on rice tillers that existed at the previous observation time but were continuously missing at subsequent observation times, and rice tillers that were determined to have effectively disappeared were excluded due to false disappearance caused by spatial overlap of adjacent rice tillers.

[0025] The actual increase and effective decline of rice tillers are correlated with the corresponding fixed population evaluation units according to the time of occurrence, forming a population growth and decline correlation chain.

[0026] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of unmanned aerial vehicle (UAV) populations described in this invention, the steps of identifying the dynamic process of population expansion transitioning to equilibrium and decline by analyzing the net growth and decline changes in rice tillers and the population turnover status, and generating a population dynamic growth and decline rate map, are as follows:

[0027] Based on the population growth and decline correlation chain, the actual newly added rice tillers and effectively regressed rice tillers in each fixed population evaluation unit are extracted simultaneously to form a continuous growth and decline sequence;

[0028] The net change in rice tillering in each fixed population evaluation unit was determined by comparing the actual newly added rice tillers with the effectively regressed rice tillers in the continuous growth and decline sequence, based on the relative changes of the two.

[0029] Based on the net changes in rice tillering, the synchronous changes in actual new tillering and effective decline within the same observation period are compared to identify the turnover degree in which new tillering and decline occur simultaneously. This distinguishes between population states with similar net changes but different internal turnover degrees, and generates population turnover status.

[0030] By continuously tracking the changes in net growth and decline of rice tillers and the changes in population turnover, we can identify the dynamic process of the rice population tiller growth and decline state from continuous expansion to high turnover equilibrium and then to continuous decline, and determine the turning point of growth and decline.

[0031] The net changes in rice tillering in each fixed population evaluation unit are converted into a rate of change according to the time interval between adjacent observation times, and then mapped with the population turnover status and the turning point of population change according to the location identifier to generate a dynamic rate of change of the population.

[0032] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV populations described in this invention, the steps for distinguishing between natural convergence and recoverable nitrogen-limited decay based on the sequential relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, and forming a population growth and decline-driven decision graph, are as follows:

[0033] Read the population dynamic growth rate map, continuously track the growth rate and population turnover status changes of each fixed population evaluation unit according to the observation time, and identify the moment when the growth and decline of rice population tillering occurs, forming a time series of population growth and decline changes;

[0034] Red reflectance, red edge reflectance, and near-infrared reflectance of the corresponding fixed population evaluation unit are extracted from the time series observation of the paddy field population to identify the canopy nitrogen nutrition status at each observation time and generate the time series of canopy nitrogen nutrition changes.

[0035] Align the population growth and decline time series with the canopy nitrogen nutrition change time series using the same fixed population evaluation unit and observation time to obtain the time correspondence between the population growth and decline turning point and the decline in canopy nitrogen nutrition status;

[0036] By comparing the timing of population growth and decline with the decline in canopy nitrogen nutrition status, when the population growth and decline precedes the decline in canopy nitrogen nutrition status, it is determined to be a natural convergence; when the decline in canopy nitrogen nutrition status precedes the population growth and decline and rice tillers continue to increase, it is determined to be a recoverable nitrogen-limiting decline.

[0037] The natural convergence and recoverable nitrogen-limited decay states of each fixed population evaluation unit are correlated and mapped with the corresponding population growth and decline inflection points to form a population growth and decline driving decision map.

[0038] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV populations described in this invention, the steps of screening nitrogen-limited regions with recovery potential based on the population growth and decline driving determination map, and determining the topdressing priority and intensity by combining the degree of insufficient tillering expansion and tillering decline trend of the rice population, are as follows:

[0039] Based on the population growth and decline-driven determination map, nitrogen-limited regions with recovery potential are screened. The current number of rice tillers in the nitrogen-limited regions with recovery potential is compared with the preset tillering expansion benchmark. The degree of insufficient tillering expansion of the rice population is characterized by the gap between the current number of tillers and the preset tillering expansion benchmark.

[0040] The growth and decline rates of nitrogen-limited regions with recovery potential were continuously compared at multiple observation times, and the tillering decline trend of rice population was characterized by the degree of change from slow growth to continuous decline.

[0041] Based on the degree of insufficient tiller expansion and the trend of tiller decline in rice populations, the topdressing needs of nitrogen-limited areas with recovery potential were ranked, and the corresponding topdressing priorities and intensities were determined according to the changes in the deepening of insufficient expansion and the acceleration of decline.

[0042] As a preferred embodiment of the rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms described in this invention, the rice topdressing decision prescription map is formed by mapping the topdressing priority and intensity of each fixed population evaluation unit according to the spatial location of the target paddy field.

[0043] Secondly, the present invention provides a rice topdressing decision system based on the dynamic growth and decline rate of UAV populations, including: a time-series observation module, which collects UAV field patrol image data, performs unified registration of the spatial position of rice populations at different observation times and completes cross-time alignment according to fixed population evaluation units to generate a time-series observation body of rice populations.

[0044] The growth and decline correlation module reads the time series observation data of rice populations, tracks the cross-time continuity relationship of rice populations within the same fixed population evaluation unit, identifies the real increase and effective decline of rice tillers, and excludes false decline caused by shading, generating a population growth and decline correlation chain.

[0045] The population growth rate module synchronously tracks the addition and decline processes in the population growth-decline correlation chain. By identifying the dynamic process of population expansion turning into equilibrium and decline through the net growth-decline changes of rice tillers and population turnover status, it generates a population dynamic growth-decline rate map.

[0046] The population dynamic growth and decline rate determination module uses the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes. Based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, it distinguishes between natural convergence and recoverable nitrogen-limited decay, thus forming a population growth and decline rate determination map.

[0047] The topdressing decision module screens nitrogen-limited regions with recovery potential based on the population growth and decline driving judgment map, and determines the priority and intensity of topdressing by combining the degree of insufficient tillering expansion and tillering decline trend of rice population, thus forming a rice topdressing decision prescription map.

[0048] The beneficial effects of this invention are as follows: By simultaneously tracking the actual increase and effective decline in the population growth and decline correlation chain, and combining the net growth and decline changes of rice tillers with the population turnover status to identify the expansion, equilibrium and decline processes, a population dynamic growth and decline rate map is generated, realizing continuous characterization of the internal replacement and growth and decline transitions of rice populations; by aligning the population growth and decline change sequence with the canopy nitrogen nutrition change sequence, and distinguishing between natural convergence and recoverable nitrogen-limited decline based on the order of growth and decline transitions and the decline in canopy nitrogen nutrition status, a population growth and decline driving judgment map is formed, and the priority and intensity of topdressing are determined accordingly, thereby improving the pertinence and rationality of rice topdressing decisions. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of a rice topdressing decision-making method based on the dynamic growth and decline rate of drone swarms.

[0051] Figure 2 The flowchart for determining the dynamics of a group.

[0052] Figure 3 The flowchart shows the generation of time-series observation data for rice paddy populations.

[0053] Figure 4 A flowchart for generating a population dynamic growth and decline rate diagram. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0057] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for rice topdressing decision-making based on the dynamic growth and decline rate of unmanned aerial vehicle (UAV) swarms, comprising the following steps:

[0058] S1. Collect UAV field patrol image data, uniformly register the spatial location of rice population at different observation times, and complete cross-time alignment according to fixed population evaluation units to generate rice population time series observation volume.

[0059] The drone field inspection image data includes the spatial location of rice tillers, canopy boundaries, canopy coverage, canopy height, red light reflectance, red edge reflectance, near-infrared reflectance, as well as the corresponding acquisition time and geographical coordinates.

[0060] The drone field patrol image data at each observation time is converted to a unified geographic coordinate benchmark, and the overall positional offset between drone field patrol images of different sorties is corrected based on the geographic coordinate information.

[0061] Furthermore, the geographic coordinate information corresponding to the UAV field patrol image data is extracted, and the geographic coordinate information used at different observation times is converted to a unified geographic coordinate benchmark. Based on the unified geographic coordinate benchmark, the spatial positioning position within the coverage area of ​​each UAV field patrol image data is determined, and spatial positioning positions with consistent geographic coordinates in UAV field patrol image data from different sorties are selected to establish cross-sortie position correspondence. The planar coordinate deviation of the same spatial positioning position in the cross-sortie position correspondence is compared, and the overall translation direction and overall translation distance between different UAV field patrol image data are determined based on the planar coordinate deviation. The spatial position of the corresponding UAV field patrol image data is adjusted according to the overall translation direction and overall translation distance to ensure that the same geographic location in different UAV field patrol image data remains consistent, thus completing the correction of the overall positional offset between different UAV field patrol images.

[0062] Using the first valid observation time of the UAV field inspection image as a spatial reference, the boundary of the rice population and stable ground features in the subsequent UAV field inspection image data are matched to correct the local registration deviation of the same spatial location at different observation times.

[0063] Furthermore, using the UAV field survey imagery from the first valid observation time as a spatial reference, rice population boundaries and stable ground features are extracted from the UAV field survey imagery from the first valid observation time and subsequent UAV field survey imagery data. The corresponding matching range is determined based on the geographical coordinate proximity relationship, and the positional relationship between the outline position of the rice population boundary and the stable ground features within the corresponding range is compared to establish the spatial correspondence between different observation times. Based on the spatial correspondence, the local positional offset of the same spatial location at different observation times is calculated, and the position of subsequent UAV field survey imagery data is corrected according to the local positional offset to ensure that the corresponding rice population boundary and stable ground features are consistent in space, thus correcting the local registration deviation of the same spatial location at different observation times.

[0064] It should be noted that stable ground features include field ridge boundaries, ditch intersections, field road edges, and corners of fixed structures. These features are obtained by extracting the edges of non-rice areas from UAV field patrol images at each observation time and selecting edge locations with consistent geographic coordinates across consecutive observation times.

[0065] Fixed population evaluation units are divided according to the unified registration of the target paddy field spatial range, and the spatial boundaries and location markers of each fixed population evaluation unit are established without change.

[0066] Furthermore, a planar coordinate grid is established according to the spatial range of the target paddy field after unified registration. The grid area is continuously divided along the length and width of the target paddy field using a preset evaluation scale. The complete grid area within the boundary of the target paddy field and the edge grid area after being truncated by the boundary of the target paddy field are respectively used as fixed group evaluation units. The coordinates of each boundary vertex of each fixed group evaluation unit under a unified geographic coordinate reference are recorded and connected to form a spatial boundary. At the same time, a unique position identifier is assigned according to the row and column position of the fixed group evaluation unit in the planar coordinate grid. After the spatial boundary and position identifier are established, they remain unchanged at each subsequent observation time, so that the same position identifier always corresponds to the same spatial range in the target paddy field.

[0067] It should be noted that a fixed population evaluation unit refers to a spatial area formed by dividing the target paddy field space into fixed regions according to a preset evaluation scale under a unified geographic coordinate benchmark. Each fixed population evaluation unit has an unchanging spatial boundary and a unique location identifier, used to continuously correspond to changes in the rice population at the same location at different observation times. The evaluation scale is set based on the actual row spacing and plant spacing of the target paddy field. First, continuous rice planting rows and the center positions of adjacent rice plants within each row are identified from the uniformly registered UAV field inspection image data. The average row spacing and average plant spacing are calculated. Then, the lateral distance covered by a fixed number of continuous planting rows is used as the evaluation scale width, and the longitudinal distance covered by a fixed number of continuous plant positions is used as the evaluation scale length, thereby determining the evaluation scale of the fixed population evaluation unit.

[0068] Based on the location identifier of the fixed population evaluation unit, the rice population observation information within the corresponding range of each observation time is extracted, and a continuous correspondence is established according to the collection time to generate the rice field population time series observation body.

[0069] Furthermore, based on the location identifiers of fixed population evaluation units, the corresponding spatial boundaries are invoked. Image ranges falling within these spatial boundaries are extracted from the UAV field survey imagery data at each observation time. Within these extracted ranges, rice tillering locations are identified. Canopy boundaries are extracted along the outer edge of the rice vegetation area, and canopy coverage is statistically analyzed. Simultaneously, canopy height, red reflectance, red-edge reflectance, and near-infrared reflectance at corresponding locations are read. These extracted data, along with the corresponding acquisition time and geographic coordinates, constitute rice population observation information. Using the location identifiers of fixed population evaluation units as cross-time correspondence markers, rice population observation information with the same location identifiers is arranged according to the chronological order of acquisition time. This allows the same fixed population evaluation unit to form a continuous observation sequence at different observation times. The continuous observation sequences of all fixed population evaluation units are then organized according to their location identifiers to generate a temporal observation volume of the rice population. This temporal observation volume of the rice population maintains continuous correspondence between rice population observation information of the same fixed population evaluation unit at different observation times, providing a temporal basis for subsequent identification of the actual increase and effective decline of rice tillers.

[0070] S2. Read the time series observation data of the paddy field population, track the cross-time continuity relationship of the rice population within the same fixed population evaluation unit, identify the real increase and effective decline of rice tillers, and exclude false decline caused by shading, and generate the population growth and decline correlation chain.

[0071] Read the time-series observation data of the paddy field population, extract the rice population observation information corresponding to each observation time according to the location identifier of the fixed population evaluation unit, and form a continuous observation sequence according to the collection time.

[0072] Spatial correspondence is established for rice population observation information at adjacent observation times within the same fixed population evaluation unit, and the cross-temporal continuity relationship of the rice population is determined based on the continuous change of the spatial location of rice tillers.

[0073] Furthermore, spatial correspondence is established for rice population observation information at adjacent observation times within the same fixed population evaluation unit. Using the spatial location of each rice tiller at the previous observation time as a reference, the spatial location of the rice tillers at the subsequent observation time is projected onto the same geographical coordinate range. The average spacing between adjacent rice tillers at the previous observation time is used to limit the corresponding search range. Within the corresponding search range, the positional distance between rice tillers at different observation times, the relative arrangement of surrounding rice tillers, and the degree of overlap of canopy boundaries are compared. A one-to-one correspondence is established between rice tillers at different observation times with the smallest positional distance and consistent surrounding arrangement. This one-to-one correspondence is repeated along consecutive observation times, connecting rice tillers that can continuously correspond at adjacent observation times into a continuous tracking sequence. This continuous tracking sequence is used to characterize the cross-time continuity relationship of the rice population within the same fixed population evaluation unit.

[0074] Based on the cross-temporal continuity of rice populations, rice tillers that did not appear at the previous observation time but appeared at the next observation time and continued to exist in subsequent observations were identified as truly newly added rice tillers.

[0075] Furthermore, based on the cross-temporal continuity of rice populations, rice tillers that appear at a later observation time but cannot establish a correspondence with any existing rice tillers at a previous observation time are considered as new candidate objects. The spatial location of the new candidate objects is then tracked along subsequent observation times. If the new candidate object can continuously establish a cross-temporal correspondence with rice tillers at adjacent locations at subsequent observation times, it is confirmed that the new candidate object is not a false addition caused by instantaneous identification error or short-term occlusion removal, and the new candidate object is determined as a real new rice tiller. For example, if a fixed population evaluation unit does not detect a rice tiller at location P at observation time T1, but a new rice tiller appears near location P at observation time T2, and the corresponding rice tiller can still be tracked at observation time T3, then the rice tiller that appears at observation time T2 is determined as a real new rice tiller.

[0076] Rice tillers that existed at the previous observation time but were continuously missing at subsequent observation times were backtracked, and rice tillers that were determined to have effectively disappeared were excluded due to false disappearance caused by spatial overlap of adjacent rice tillers.

[0077] Furthermore, based on the cross-temporal continuity of rice populations, rice tillers that existed in the previous observation time but could not be established in the subsequent observation time are located to the last spatial position of their appearance. The existence status of rice tillers at the corresponding spatial positions is continuously checked along subsequent observation times, while the changes in the canopy boundaries of adjacent rice tillers around the corresponding spatial position are compared. When the missing position is covered by the canopy expanded by adjacent rice tillers and the corresponding rice tiller reappears in subsequent observation times, the missing position is determined to be a false regression caused by the spatial coverage of adjacent rice tillers. When the missing position is not covered by the canopy of adjacent rice tillers and the corresponding rice tiller continues to be missing in subsequent observation times, the corresponding rice tiller is determined to be a effectively regressed rice tiller.

[0078] The actual increase and effective decline of rice tillers are correlated with the corresponding fixed population evaluation units according to the time of occurrence, forming a population growth and decline correlation chain.

[0079] Furthermore, the actual newly added rice tillers and the effectively regressed rice tillers are respectively bound to the location identifiers of the corresponding fixed population evaluation units, and the observation time of the first occurrence of actual new tillers and the observation time of the continuous absence of effective regression are taken as the corresponding occurrence time. The actual new tillers and effective regression are collected according to the fixed population evaluation units, and then arranged in chronological order of occurrence time, so that the actual new tillers and effective regression of the same fixed population evaluation unit in adjacent observation times are continuously connected, while retaining the spatial location and growth and decline type of the corresponding rice tillers, thereby forming a population growth and decline correlation chain that can continuously characterize the rice tillering new and regression process of each fixed population evaluation unit.

[0080] S3. Synchronously track the new growth and decline processes in the population growth and decline correlation chain, identify the dynamic process of population expansion to equilibrium and decline through the net growth and decline changes of rice tillers and population turnover status, and generate a population dynamic growth and decline rate map.

[0081] Based on the population growth and decline correlation chain, the actual newly added rice tillers and effectively regressed rice tillers in each fixed population evaluation unit are extracted simultaneously to form a continuous growth and decline sequence.

[0082] The net change in rice tillering in each fixed population evaluation unit was determined by comparing the actual newly added rice tillers with the effectively regressed rice tillers in the continuous growth and decline sequence, based on the relative changes of the two.

[0083] Furthermore, the continuous growth and decline sequence was divided into observation periods according to adjacent observation times. Within each fixed population evaluation unit, the actual number of newly added rice tillers and the number of effectively regressed rice tillers in the same observation period were counted separately. The results were then normalized using the number of rice tillers at the beginning of the observation period to obtain the actual increase and the effective decrease. The actual increase and the effective decrease were compared concurrently. When the actual increase was greater than the effective decrease, it indicated a net increase in rice tillers. When the two were close, it indicated a balance in the growth and decline of rice tillers. When the actual increase was less than the effective decrease, it indicated a net decrease in rice tillers. The net growth and decline of rice tillers in each fixed population evaluation unit was determined by the difference between the two and their positive or negative direction.

[0084] Based on the net changes in rice tillering, the synchronous changes in actual new tillering and effective decline within the same observation period are compared to identify the turnover degree in which new tillering and decline occur simultaneously. This distinguishes between population states with similar net changes but different internal replacement degrees, and generates population turnover status.

[0085] Furthermore, based on the net changes in rice tillering, the actual number of newly added rice tillers and the number of effectively regressed rice tillers were counted in each fixed population evaluation unit within the same observation period. These numbers were then normalized using the number of rice tillers at the beginning of the observation period to obtain the actual increase ratio and the effective regression ratio. The turnover rate was calculated by dividing twice the product of the actual increase ratio and the effective regression ratio by the sum of the two. A decrease in either actual increase or effective regression resulted in a simultaneous decrease in turnover rate, while an increase in both resulted in an increase in turnover rate. The net increase in rice tillering was then used to calculate the turnover rate. The rice population is grouped according to the direction of its growth and decline. Fixed population evaluation units with a net growth and decline difference of no more than one normalized change corresponding to rice tillers are identified as comparison objects with similar net growth and decline. The turnover degree corresponding to the fixed population evaluation units with similar net growth and decline is then compared. The larger the turnover degree value, the more the actual increase and effective decline occur at the same time under approximately the same net growth and decline, and the stronger the internal replacement degree of the rice population. The smaller the turnover degree value, the weaker the internal replacement degree. Based on this, the population states with similar net growth and decline but different internal replacement degrees are distinguished, and the population turnover state is generated.

[0086] It should be noted that similar net changes mean that the difference between the actual increase rate and the effective decrease rate is close within the same observation period. For example, if the net change of fixed population evaluation unit A is 0.10 and the net change of fixed population evaluation unit B is 0.09, they can be regarded as having similar net changes.

[0087] The degree of internal replacement refers to the strength of the simultaneous occurrence of real new additions and effective declines when net growth and decline are similar. For example, if the real new addition ratio of fixed group evaluation unit A is 0.20 and the effective decline ratio is 0.10, and the real new addition ratio of fixed group evaluation unit B is 0.30 and the effective decline ratio is 0.21, then the degree of internal replacement of fixed group evaluation unit B is stronger.

[0088] Population turnover status refers to the internal turnover status of rice populations, which is reflected by the degree of simultaneous occurrence of actual new growth and effective decline based on the net change in rice tillering. For example, when the net change in tillering is close to zero but both actual new growth and effective decline occur significantly, it indicates that the rice population is in a high turnover state.

[0089] The turnover rate, where additions and deletions occur simultaneously, is expressed as:

[0090] ;

[0091] in, Indicates the first The degree of group turnover over a single observation period This represents the proportion of the actual number of newly added rice tillers to the initial number of rice tillers in that observation period. This indicates the proportion of effectively regressed rice tillers to the initial number of rice tillers in that observation period. This indicates the use of extremely small positive numbers to avoid having a denominator of zero;

[0092] By continuously tracking the changes in net tillering and population turnover in rice, we can identify the dynamic process of the rice population's tillering state changing from continuous expansion to high turnover equilibrium and then to continuous decline, and determine the turning point of the tillering and turnover.

[0093] Furthermore, the net increase and decrease of rice tillers and the population turnover status of each fixed population evaluation unit were continuously tracked in adjacent observation periods. When the net increase and decrease of rice tillers was positive for two consecutive observation periods and the population turnover status showed that new growth was dominant, the corresponding stage was identified as continuous expansion. When the net increase and decrease of rice tillers gradually decreased from a positive value to close to zero, and the actual new growth and effective decline continued to occur synchronously and showed a high-frequency replacement state, the corresponding stage was identified as high turnover equilibrium. When the net increase and decrease of rice tillers was negative for two consecutive observation periods and the population turnover status showed that decline was dominant, the corresponding stage was identified as continuous decline. By continuously connecting the three stages according to the observation time, the dynamic process of the rice population tillering status changing from continuous expansion to high turnover equilibrium and then to continuous decline was identified. The observation positions where continuous expansion first turned into high turnover equilibrium and the observation positions where high turnover equilibrium first turned into continuous decline were determined as the turning points of the increase and decrease.

[0094] It should be noted that high turnover balance refers to the simultaneous and continuous occurrence of actual new growth and effective decline within the same observation period, with their quantities offsetting each other, so that the net change in rice tillering is close to zero, but the population still maintains a state of significant replacement.

[0095] The net changes in rice tillering in each fixed population evaluation unit are converted into a rate of change according to the time interval between adjacent observation times, and then mapped with the population turnover status and the turning point of population change according to the location identifier to generate a dynamic rate of change of the population.

[0096] Furthermore, the net change in rice tillering between adjacent observation times in each fixed population evaluation unit is divided by the actual time interval between the two corresponding observation times to obtain the rate of rice tillering change per unit time. The rates of rice tillering change for each fixed population evaluation unit are then continuously arranged according to the observation cycle. Using the location identifier of the fixed population evaluation unit as the correlation benchmark, the rate of rice tillering change for the same fixed population evaluation unit within the corresponding observation cycle is bound to the population turnover status. Simultaneously, the turning point of population change is mapped to the observation time at which the rice tillering change rate transitions. Then, using the spatial boundary of the fixed population evaluation unit as the mapping area, the rice tillering growth rate, population turnover status, and population growth-reversal position corresponding to each observation period are written into the spatial area corresponding to the same position marker. They are then continuously organized according to the order of observation time to generate a population dynamic growth rate map that can reflect the tillering growth rate, internal turnover changes, and growth-reversal process of each fixed population evaluation unit. This map is used to continuously characterize the rice tillering growth rate, internal turnover status, and growth-reversal process of each fixed population evaluation unit, providing a basis for subsequent determination of whether rice population decline is driven by nitrogen nutrition limitation.

[0097] S4. Using the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes, and based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, distinguish between natural convergence and recoverable nitrogen-limited decay, and form a population growth and decline driving determination map.

[0098] Read the population dynamic growth rate map, continuously track the growth rate and population turnover status changes of each fixed population evaluation unit according to the observation time, and identify the turning point of rice population tillering growth and decline to form a time sequence of population growth and decline changes.

[0099] Furthermore, the population dynamic growth rate map is read, and the growth rate and population turnover status corresponding to each fixed population evaluation unit are extracted sequentially according to the observation time. When the growth rate remains positive and the population turnover status shows that new growth is dominant, it is marked as a continuous expansion stage. When the growth rate drops from a positive value to close to zero and the population turnover status changes to high frequency of change, the observation time when the first high turnover balance appears is identified as the first growth-decline turning point. The process continues to track along subsequent observation times. When the growth rate changes from close to zero to a continuous negative value and the population turnover status shows that decline is dominant, the observation time when the first continuous decline begins is identified as the second growth-decline turning point. The growth rate and population turnover status are arranged in the order of observation time, and the first and second growth-decline turning points are marked at the corresponding positions to form a time sequence of population growth-decline changes.

[0100] It should be noted that "regression dominance" means that the number of effectively regressed rice tillers is consistently greater than the number of actually newly grown rice tillers within the same observation period, resulting in a negative net change in rice tiller growth.

[0101] Red reflectance, red edge reflectance, and near-infrared reflectance of the corresponding fixed population evaluation unit are extracted from the time-series observation of the paddy field population to identify the canopy nitrogen nutrient status at each observation time and generate a time-series of canopy nitrogen nutrient changes.

[0102] Furthermore, red reflectance, red edge reflectance, and near-infrared reflectance of each fixed population evaluation unit at the corresponding observation time were extracted from the time-series observation data of the paddy field population. First, red reflectance was used to characterize changes in chlorophyll absorption, and then red edge reflectance was used to characterize the sensitivity of chlorophyll content changes. In combination with near-infrared reflectance, changes in canopy structure and biomass were reflected. The three types of reflectance at the same observation time were combined according to the fixed population evaluation units to obtain the canopy nitrogen nutrition characterization value of each fixed population evaluation unit. Then, the canopy nitrogen nutrition characterization value at the current observation time was continuously compared with the corresponding value at the previous observation time. The canopy nitrogen nutrition status at the corresponding observation time was determined based on the rise, stabilization, or decline of the canopy nitrogen nutrition characterization value. The canopy nitrogen nutrition status at each observation time was arranged in the order of collection time to generate a time-series of canopy nitrogen nutrition changes.

[0103] It should be noted that using red edge reflectance to characterize the sensitivity of chlorophyll content changes specifically involves comparing the range of red edge reflectance changes in the same fixed population evaluation unit at adjacent observation times. When the red edge reflectance continues to increase, it indicates an increasing trend in canopy chlorophyll content; when the red edge reflectance continues to decrease, it indicates a decreasing trend in canopy chlorophyll content. The range of change reflects the sensitivity of chlorophyll content changes.

[0104] By aligning the population growth and decline time series with the canopy nitrogen nutrient change time series using the same fixed population evaluation unit and observation time, the time correspondence between the population growth and decline turning point and the decline in canopy nitrogen nutrient status can be obtained.

[0105] Furthermore, the population growth and decline time series and the canopy nitrogen nutrient change time series were mapped one-to-one according to the location identifier of the fixed population evaluation unit, and the same observation time was used as the time reference. The growth rate, population turnover status, and canopy nitrogen nutrient status of each fixed population evaluation unit were written into the same time series. In the population growth and decline time series, the turning points from continuous expansion to high turnover equilibrium and from high turnover equilibrium to continuous decline were marked. In the canopy nitrogen nutrient change time series, the first observation time when the canopy nitrogen nutrient characterization value changed from stable or rising to falling and continued to fall in subsequent observations was determined as the time when the canopy nitrogen nutrient status declined. Then, the turning points of population growth and decline and the time when the canopy nitrogen nutrient status declined were compared according to the observation time, and the order and interval of observation were recorded to obtain the time correspondence between the population growth and decline turning points and the canopy nitrogen nutrient status decline.

[0106] By comparing the timing of population growth and decline with the decline in canopy nitrogen nutrition status, when the population growth and decline precedes the decline in canopy nitrogen nutrition status, it is determined to be a natural convergence; when the decline in canopy nitrogen nutrition status precedes the population growth and decline and rice tillering continues to increase, it is determined to be a recoverable nitrogen-limiting decline.

[0107] Furthermore, based on the time correspondence between the population growth and decline transition and the decline in canopy nitrogen nutrition status, the timing of the population growth and decline transition and the timing of the decline in canopy nitrogen nutrition status were compared for each fixed population evaluation unit. When the population growth and decline transition occurred earlier than the canopy nitrogen nutrition status decline, and the canopy nitrogen nutrition status remained stable before the population growth and decline transition, it indicates that the rice population first transitioned from continuous expansion to high turnover equilibrium or continuous decline, and the canopy nitrogen nutrition status declined subsequently. The corresponding fixed population evaluation unit was judged as naturally converged. When the canopy nitrogen nutrition status decline occurred earlier than the population growth and decline transition, and real new rice tillers continued to appear after the canopy nitrogen nutrition status decline but before the population growth and decline transition, it indicates that the canopy nitrogen nutrition status decline occurred before the population expansion capacity disappeared, and the rice population still retained the tiller expansion capacity. The corresponding fixed population evaluation unit was judged as a recoverable nitrogen-limited decline.

[0108] The natural convergence and recoverable nitrogen-limited decay states of each fixed population evaluation unit are correlated and mapped with the corresponding population growth and decline inflection points to form a population growth and decline driving decision map.

[0109] Furthermore, using the location identifier of the fixed population evaluation unit as the association benchmark, a one-to-one correspondence is established between the natural convergence state or recoverable nitrogen-limited decay state corresponding to each fixed population evaluation unit and the corresponding population growth and decline inflection point in the population growth and decline time series, and the spatial boundary of the corresponding fixed population evaluation unit is recorded. According to the spatial location of the fixed population evaluation unit in the target paddy field, the natural convergence state and the recoverable nitrogen-limited decay state are written into the corresponding spatial boundary, and the population growth and decline inflection point is marked in the corresponding spatial boundary, so that each fixed population evaluation unit retains both decay type and growth and decline inflection point information, forming a population growth and decline driving determination map that can characterize the population growth and decline driving type and inflection point distribution in different areas of the target paddy field.

[0110] S5. Based on the population growth and decline driving determination map, nitrogen-limited regions with recovery potential are screened, and the priority and intensity of topdressing are determined by combining the degree of insufficient tillering expansion and tillering decline trend of rice population, thus forming a rice topdressing decision prescription map.

[0111] Based on the population growth and decline-driven determination map, nitrogen-limited regions with recovery potential are screened. The current number of rice tillers in the nitrogen-limited regions with recovery potential is compared with the preset tillering expansion benchmark. The degree of insufficient tillering expansion of the rice population is characterized by the gap between the current number of tillers and the preset tillering expansion benchmark.

[0112] Furthermore, based on the population growth and decline driving determination map, the corresponding population growth and decline driving state is read one by one according to the location identifier of the fixed population evaluation unit. Fixed population evaluation units marked as recoverable nitrogen-limited decay state are screened out, and nitrogen-limited regions with recovery potential are determined according to the spatial boundary of the fixed population evaluation unit. The number of rice tillers still existing in each nitrogen-limited region with recovery potential at the current observation time is counted, and a preset tillering benchmark consistent with the evaluation scale of the fixed population evaluation unit is called for corresponding comparison. When the current number of rice tillers reaches or exceeds the preset tillering benchmark, the difference that does not reach the preset tillering benchmark is recorded as zero. When the current number of rice tillers is lower than the preset tillering benchmark, the difference is obtained by subtracting the current number of rice tillers from the preset tillering benchmark. The ratio of the difference to the preset tillering benchmark is used to represent the difference that does not reach the preset tillering benchmark, and the size of the difference represents the degree of insufficient tillering of the rice population.

[0113] It should be noted that the tillering expansion benchmark is set based on the rice variety, planting density, and the target effective tillering number per unit area corresponding to the current growth stage of the target paddy field. Specifically, the target effective tillering number to be achieved per unit area is first determined, and then the corresponding actual area is calculated based on the spatial boundary of the fixed population evaluation unit. The target effective tillering number per unit area is multiplied by the actual area of ​​the fixed population evaluation unit to obtain the target rice tillering number to be achieved by the corresponding fixed population evaluation unit at the current growth stage. The target rice tillering number is then used as the tillering expansion benchmark.

[0114] The difference between the pre-set tillering amplification baseline and the target is expressed as:

[0115] ;

[0116] in, Indicates the first The gap between the fixed population evaluation units and the preset tillering expansion benchmark was not met. This indicates the preset tillering and amplification baseline. Indicates the first The current number of rice tillers in a fixed population evaluation unit;

[0117] The growth and decline rates of nitrogen-limited regions with recovery potential were continuously compared at multiple observation times, and the tillering decline trend of rice population was characterized by the degree of change from slow growth to continuous decline.

[0118] Furthermore, the tillering rate of rice in nitrogen-limited regions with recovery potential was continuously read at each observation time, and a tillering rate sequence was formed according to the order of observation time. From the tillering rate sequence, observation periods that were still positive but continuously decreasing were identified. The last tillering rate that remained positive was taken as the terminal rate of the slow tillering growth stage. The tillering rate was then continuously checked to see if it turned negative and remained negative for multiple observation periods. The observation period in which consecutive negative values ​​first appeared was determined as the starting point of the continuous tillering reduction stage. The average change amplitude of the terminal rate of the slow tillering growth stage and the tillering rate of each observation period in the continuous tillering reduction stage were compared. The larger the change amplitude and the longer the duration of consecutive negative values, the stronger the change from slow growth to continuous reduction in rice tillering. Based on this, the tillering decline trend of the rice population was characterized.

[0119] The degree of change in tillering from slow growth to continuous decline is expressed as:

[0120]

[0121] in, Indicates the first The degree of change in tillering from slow growth to continuous decline in a fixed population evaluation unit. Indicates the first The fixed group evaluation unit from the first The first observation period to the first The sequence of rice tillering rate changes over a continuous period of one observation cycle. Indicates the first The fixed group evaluation unit in the first The first observation period before the turning point of tillering and growth occurs Rice tillering rate over one observation period This indicates the preset growth and decline rate of the benchmark.

[0122] It should be noted that the benchmark for the rate of change in tillering is based on historical observation data during the period of stable canopy nitrogen nutrition and continuous expansion of rice tillers. It is set by calculating the change range of the rate of change in rice tillering between adjacent observation periods in each fixed population evaluation unit, and taking the average value of the change range during the continuous and stable expansion phase.

[0123] Based on the degree of insufficient tiller expansion and the trend of tiller decline in rice populations, the topdressing needs of nitrogen-limited areas with recovery potential were ranked, and the corresponding topdressing priorities and intensities were determined according to the changes in the deepening of insufficient expansion and the acceleration of decline.

[0124] Furthermore, the degree of insufficient tillering and the tillering decline trend of rice populations corresponding to each nitrogen-limited region with recovery potential are ranked from largest to smallest. The two ranking positions are added together to form the topdressing demand order. For example, if region A ranks first in tillering insufficiency and second in tillering decline trend, and region B ranks third and first respectively, then the sum of the ranking positions of region A is 3, and that of region B is 4, meaning that region A's topdressing demand is higher than that of region B. Then, the topdressing priority is divided according to the order of topdressing demand, with regions ranked higher being determined as high topdressing priority, regions ranked in the middle as medium topdressing priority, and regions ranked lower as low topdressing priority. Within the same topdressing priority, the basic topdressing intensity is determined by the degree of insufficient tillering, and then incremental adjustments are made according to the tillering decline trend. For example, if two regions have the same degree of insufficient tillering, but region A has a higher tillering decline rate than region B, then region A's topdressing intensity is higher than that of region B. This determines the topdressing priority and intensity corresponding to each nitrogen-limited region with recovery potential.

[0125] The topdressing priority and intensity of each fixed group evaluation unit are mapped according to the spatial location of the target paddy field to form a rice topdressing decision prescription map.

[0126] Furthermore, using the location identifiers of fixed-group evaluation units as spatial mapping indexes, the spatial boundaries corresponding to each fixed-group evaluation unit under a unified geographical coordinate benchmark are invoked, and the determined topdressing priority and intensity are bound to the corresponding location identifiers. Then, according to the actual coordinates of each fixed-group evaluation unit in the target paddy field, the topdressing priority and intensity are written into the corresponding spatial boundaries, ensuring that each fixed-group evaluation unit forms a topdressing decision area with a clearly defined spatial range. For spatially adjacent fixed-group evaluation units with consistent topdressing priorities and intensities, adjacent spatial boundaries are continuously connected to form the same topdressing area. For fixed-group evaluation units with different topdressing priorities or intensities, the original spatial boundaries are retained. This creates a spatial distribution relationship within the target paddy field that can characterize the topdressing sequence and intensity on a zone-by-zone basis, generating a rice topdressing decision prescription map. The rice topdressing decision prescription map reflects the spatial distribution of topdressing priorities and intensities corresponding to each fixed-group evaluation unit in the target paddy field, providing a direct basis for subsequent differentiated topdressing implementation by region.

[0127] This embodiment also provides a rice topdressing decision system based on the dynamic growth and decline rate of UAV swarms, including:

[0128] The temporal observation module collects image data from UAV field patrols, uniformly registers the spatial location of rice populations at different observation times, and completes cross-time alignment according to fixed population evaluation units to generate a temporal observation volume of rice populations.

[0129] The growth and decline correlation module reads the time series observation data of rice populations, tracks the cross-time continuity relationship of rice populations within the same fixed population evaluation unit, identifies the real increase and effective decline of rice tillers, and excludes false decline caused by shading, generating a population growth and decline correlation chain.

[0130] The population growth rate module synchronously tracks the addition and decline processes in the population growth-decline correlation chain. By identifying the dynamic process of population expansion turning into equilibrium and decline through the net growth-decline changes of rice tillers and population turnover status, it generates a population dynamic growth-decline rate map.

[0131] The population dynamic growth and decline rate determination module uses the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes. Based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, it distinguishes between natural convergence and recoverable nitrogen-limited decay, thus forming a population growth and decline rate determination map.

[0132] The topdressing decision module screens nitrogen-limited regions with recovery potential based on the population growth and decline driving judgment map, and determines the priority and intensity of topdressing by combining the degree of insufficient tillering expansion and tillering decline trend of rice population, thus forming a rice topdressing decision prescription map.

[0133] In summary, this invention achieves continuous characterization of rice population turnover and transition by simultaneously tracking real new growth and effective decline in the population growth and decline correlation chain, identifying the expansion, equilibrium, and decline processes by combining net tillering changes and population turnover status, generating a population dynamic growth and decline rate map, and aligning the population growth and decline timeline with the canopy nitrogen nutrition change timeline, distinguishing between natural convergence and recoverable nitrogen-limited decline based on the order of growth and decline transitions and canopy nitrogen nutrition status decline, forming a population growth and decline driving determination map, and determining the priority and intensity of topdressing accordingly, thereby improving the pertinence and rationality of rice topdressing decisions.

[0134] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. 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 be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for rice topdressing decision-making based on the dynamic growth and decline rate of unmanned aerial vehicle (UAV) swarms, characterized in that, include: Collect drone field patrol image data, uniformly register the spatial location of rice population at different observation times, and complete cross-time alignment according to fixed population evaluation units to generate a temporal observation volume of rice population; Read the time series observation data of rice populations, track the cross-time continuity relationship of rice populations within the same fixed population evaluation unit, identify the real increase and effective decline of rice tillers, and exclude false decline caused by shading, and generate population growth and decline correlation chains; The process of addition and decline in the population growth and decline correlation chain is tracked simultaneously. By identifying the dynamic process of population expansion to balance and decline through the net growth and decline of rice tillers and the population turnover status, a population dynamic growth and decline rate map is generated. By using the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes, and by distinguishing between natural convergence and recoverable nitrogen-limited decay based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, a population growth and decline driving determination map is formed. Based on the population growth and decline driving determination map, nitrogen-limited regions with recovery potential were screened, and the priority and intensity of topdressing were determined by combining the degree of insufficient tiller expansion and tiller decline trend of rice population, thus forming a rice topdressing decision prescription map.

2. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 1, characterized in that: The drone field inspection image data includes the spatial location of rice tillers, canopy boundaries, canopy coverage, canopy height, red light reflectance, red edge reflectance, near-infrared reflectance, as well as the corresponding acquisition time and geographical coordinates.

3. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 2, characterized in that, The steps for uniformly registering the spatial locations of rice populations at different observation times and aligning them across time periods according to fixed population evaluation units to generate a temporal observation volume of the rice paddy population are as follows: The drone field patrol image data at each observation time is converted to a unified geographic coordinate benchmark, and the overall positional offset between drone field patrol images of different sorties is corrected based on the geographic coordinate information. Using the first valid observation time of the UAV field inspection image as the spatial reference, the boundary of the rice population and stable ground features in the subsequent UAV field inspection image data are matched to correct the local registration deviation of the same spatial location at different observation times. Fixed population evaluation units are divided according to the spatial range of the target paddy field after unified registration, and the spatial boundaries and location markers of each fixed population evaluation unit are established without change. Based on the location identifier of the fixed population evaluation unit, the rice population observation information within the corresponding range of each observation time is extracted, and a continuous correspondence is established according to the collection time to generate the rice field population time series observation body.

4. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 1, characterized in that, The steps for tracking the cross-time continuity relationship of rice populations within the same fixed population evaluation unit are as follows: Read the time-series observation data of the paddy field population, extract the rice population observation information corresponding to each observation time according to the location identifier of the fixed population evaluation unit, and form a continuous observation sequence according to the collection time; Spatial correspondence is established for rice population observation information at adjacent observation times within the same fixed population evaluation unit, and the cross-temporal continuity relationship of the rice population is determined based on the continuous change of the spatial location of rice tillers.

5. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 4, characterized in that, The process involves identifying the true increase and effective decline of rice tillers, while excluding false decline caused by shading, to generate a population dynamics correlation chain. The steps are as follows: Based on the cross-time continuity relationship of rice populations, rice tillers that did not appear at the previous observation time but appeared at the next observation time and continued to exist in subsequent observations were identified as truly newly added rice tillers. Backtracking was performed on rice tillers that existed at the previous observation time but were continuously missing at subsequent observation times, and rice tillers that were determined to have effectively disappeared were excluded due to false disappearance caused by spatial overlap of adjacent rice tillers. The actual increase and effective decline of rice tillers are correlated with the corresponding fixed population evaluation units according to the time of occurrence, forming a population growth and decline correlation chain.

6. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 1, characterized in that, The process of identifying the dynamic process of population expansion transitioning to equilibrium and decline by analyzing the net changes in rice tillering and population turnover status, and generating a population dynamic growth rate map, is described in the following steps: Based on the population growth and decline correlation chain, the actual newly added rice tillers and effectively regressed rice tillers in each fixed population evaluation unit are extracted simultaneously to form a continuous growth and decline sequence; The net change in rice tillering in each fixed population evaluation unit was determined by comparing the actual newly added rice tillers with the effectively regressed rice tillers in the continuous growth and decline sequence, based on the relative changes of the two. Based on the net changes in rice tillering, the synchronous changes in actual new tillering and effective decline within the same observation period are compared to identify the turnover degree in which new tillering and decline occur simultaneously. This distinguishes between population states with similar net changes but different internal turnover degrees, and generates population turnover status. By continuously tracking the changes in net growth and decline of rice tillers and the changes in population turnover, we can identify the dynamic process of the rice population tiller growth and decline state from continuous expansion to high turnover equilibrium and then to continuous decline, and determine the turning point of growth and decline. The net changes in rice tillering in each fixed population evaluation unit are converted into a rate of change according to the time interval between adjacent observation times, and then mapped with the population turnover status and the turning point of population change according to the location identifier to generate a dynamic rate of change of the population.

7. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 6, characterized in that, The process of distinguishing between natural convergence and recoverable nitrogen-limited decay based on the relationship between the population growth and decline in the canopy nitrogen nutrient status, and forming a population growth and decline-driven determination diagram, is as follows: Read the population dynamic growth rate map, continuously track the growth rate and population turnover status changes of each fixed population evaluation unit according to the observation time, and identify the moment when the growth and decline of rice population tillering occurs, forming a time series of population growth and decline changes; Red reflectance, red edge reflectance, and near-infrared reflectance of the corresponding fixed population evaluation unit are extracted from the time series observation of the paddy field population to identify the canopy nitrogen nutrition status at each observation time and generate the time series of canopy nitrogen nutrition changes. Align the population growth and decline time series with the canopy nitrogen nutrition change time series using the same fixed population evaluation unit and observation time to obtain the time correspondence between the population growth and decline turning point and the decline in canopy nitrogen nutrition status; By comparing the timing of population growth and decline with the decline in canopy nitrogen nutrition status, when the population growth and decline precedes the decline in canopy nitrogen nutrition status, it is determined to be a natural convergence; when the decline in canopy nitrogen nutrition status precedes the population growth and decline and rice tillers continue to increase, it is determined to be a recoverable nitrogen-limiting decline. The natural convergence and recoverable nitrogen-limited decay states of each fixed population evaluation unit are correlated and mapped with the corresponding population growth and decline inflection points to form a population growth and decline driving decision map.

8. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 1, characterized in that, The steps for screening nitrogen-limited regions with recovery potential based on the population growth and decline-driven determination map, and determining the priority and intensity of topdressing by combining the degree of insufficient tillering expansion and tillering decline trend of the rice population, are as follows: Based on the population growth and decline-driven determination map, nitrogen-limited regions with recovery potential are screened. The current number of rice tillers in the nitrogen-limited regions with recovery potential is compared with the preset tillering expansion benchmark. The degree of insufficient tillering expansion of the rice population is characterized by the gap between the current number of tillers and the preset tillering expansion benchmark. The growth and decline rates of nitrogen-limited regions with recovery potential were continuously compared at multiple observation times, and the tillering decline trend of rice population was characterized by the degree of change from slow growth to continuous decline. Based on the degree of insufficient tiller expansion and the trend of tiller decline in rice populations, the topdressing needs of nitrogen-limited areas with recovery potential were ranked, and the corresponding topdressing priorities and intensities were determined according to the changes in the deepening of insufficient expansion and the acceleration of decline.

9. The rice topdressing decision-making method based on the dynamic growth and decline rate of UAV swarms as described in claim 8, characterized in that, The rice topdressing decision prescription map is formed by mapping the topdressing priority and intensity of each fixed population evaluation unit according to the spatial location of the target paddy field.

10. A rice topdressing decision system based on the dynamic growth and decline rate of UAV swarms, based on the rice topdressing decision method based on the dynamic growth and decline rate of UAV swarms as described in any one of claims 1 to 9, characterized in that, include: The temporal observation module collects image data from UAV field patrols, uniformly registers the spatial location of rice populations at different observation times, and completes cross-time alignment according to fixed population evaluation units to generate a temporal observation volume of rice populations. The growth and decline correlation module reads the time series observation data of rice populations, tracks the cross-time continuity relationship of rice populations within the same fixed population evaluation unit, identifies the real increase and effective decline of rice tillers, and excludes false decline caused by shading, generating a population growth and decline correlation chain. The population growth rate module synchronously tracks the addition and decline processes in the population growth-decline correlation chain. By identifying the dynamic process of population expansion turning into equilibrium and decline through the net growth-decline changes of rice tillers and population turnover status, it generates a population dynamic growth-decline rate map. The population dynamic growth and decline rate determination module uses the population dynamic growth and decline rate map to align the time sequence of canopy nitrogen nutrient changes. Based on the relationship between the growth and decline turning points and the decline in canopy nitrogen nutrient status, it distinguishes between natural convergence and recoverable nitrogen-limited decay, thus forming a population growth and decline rate determination map. The topdressing decision module screens nitrogen-limited regions with recovery potential based on the population growth and decline driving judgment map, and determines the priority and intensity of topdressing by combining the degree of insufficient tillering expansion and tillering decline trend of rice population, thus forming a rice topdressing decision prescription map.

Images