An intelligent citrus fruit sunburn monitoring and early warning method and system

Abstract

The present invention discloses an intelligent citrus fruit sunburn monitoring and early warning method and system, which relates to the technical field of citrus fruit. The method comprises: collecting periodic growth records of a target citrus orchard within a preset historical window, generating K periodic growth record sets; obtaining L divided grid areas; generating L first monitoring trajectories; obtaining L fruit growth monitoring data sets; obtaining L first extraction step lengths and L second extraction step lengths; configuring L sunburn monitoring and analysis sub-channels of a sunburn monitoring and analysis channel to obtain target sunburn monitoring results; judging whether the target sunburn monitoring results meet a preset threshold, and if so, obtaining an early warning instruction, and sending the early warning instruction to a user end for sunburn early warning. The present invention solves the technical problems in the prior art that the monitoring data is redundant during citrus fruit sunburn monitoring, resulting in a long monitoring and analysis cycle and low early warning reliability, and achieves the technical effect of improving the accuracy of citrus fruit sunburn monitoring and early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of citrus fruits, and in particular to an intelligent citrus fruit sunburn monitoring and early warning method and system. Background Art

[0002] The sunburn phenomenon of citrus fruits has a very important impact on citrus production, so the research on sunburn monitoring and early warning of citrus fruits has become a focus of attention. At present, the monitoring cycle of sunburn that occurs after citrus fruits are subjected to high temperature stress is too long, and the amount of data collected is too much. It is impossible to distinguish the monitoring data that truly conforms to the actual fruit growth conditions of fruit trees in the orchard, resulting in a long time for data analysis. In addition, due to the limitations of the capabilities of technical personnel, false alarms or no alarms are prone to occur during the analysis process, resulting in the citrus fruit yield failing to achieve the expected result. The existing technology has technical problems in the monitoring data of citrus fruit sunburn monitoring, which leads to a long monitoring and analysis cycle and low early warning reliability. Summary of the invention

[0003] The present application provides an intelligent citrus fruit sunburn monitoring and early warning method and system, which is used to solve the technical problems in the prior art of citrus fruit sunburn monitoring that the monitoring data is redundant, resulting in a long monitoring and analysis cycle and low early warning reliability.

[0004] In view of the above problems, the present application provides an intelligent citrus fruit sunburn monitoring and early warning method and system.

[0005] The first aspect of the present application provides an intelligent citrus fruit sunburn monitoring and early warning method, the method comprising: collecting periodic growth records of a target citrus orchard within a preset historical window to generate K periodic growth record sets; performing grid division and fusion on the target citrus orchard based on the K periodic growth record sets to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers; performing monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories; monitoring the fruit growth trend of the target citrus orchard according to the L first monitoring trajectories to obtain L A fruit growth monitoring data set; using L growth trend feature identifiers to perform multi-channel extraction and analysis on a target citrus orchard to obtain L first extraction steps and L second extraction steps; configuring L sunburn monitoring and analysis sub-channels of a sunburn monitoring and analysis channel according to the L first extraction steps and the L second extraction steps, using the configured L sunburn monitoring and analysis sub-channels to monitor and analyze the L fruit growth monitoring data sets to obtain a target sunburn monitoring result; determining whether the target sunburn monitoring result meets a preset threshold, and if so, obtaining an early warning instruction, and sending the early warning instruction to a user end for sunburn warning.

[0006] According to a second aspect of the present application, there is provided an intelligent citrus fruit sunburn monitoring and early warning system, the system comprising: a first generating unit, for collecting periodic growth records of a target citrus orchard within a preset historical window, and generating K periodic growth record sets; a first obtaining unit, for performing grid division and fusion on the target citrus orchard based on the K periodic growth record sets, and obtaining L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers; a second generating unit, for performing monitoring trajectory analysis based on the L divided grid areas, and generating L first monitoring trajectories; a second obtaining unit, for performing fruit growth trend monitoring on the target citrus orchard according to the L first monitoring trajectories, Obtain L fruit growth monitoring data sets; a third obtaining unit, used to perform multi-channel extraction and analysis on the target citrus orchard using L growth trend feature identifiers, and obtain L first extraction steps and L second extraction steps; a fourth obtaining unit, used to configure L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel according to the L first extraction steps and the L second extraction steps, and use the configured L sunburn monitoring and analysis sub-channels to monitor and analyze the L fruit growth monitoring data sets to obtain a target sunburn monitoring result; a first sending unit, used to determine whether the target sunburn monitoring result meets a preset threshold, and if so, obtain an early warning instruction, and send the early warning instruction to the user end for sunburn warning.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The present application generates K sets of periodic growth records by collecting periodic growth records of a target citrus orchard within a preset historical window, and then performs grid division and fusion on the target citrus orchard based on the K sets of periodic growth records to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers, and then performs monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories; fruit growth trend monitoring is performed on the target citrus orchard according to the L first monitoring trajectories to obtain L fruit growth monitoring data sets; multi-channel extraction and analysis is performed on the target citrus orchard using the L growth trend feature identifiers to obtain L first extraction steps and L second extraction steps; L sunburn monitoring and analysis sub-channels of a sunburn monitoring and analysis channel are configured according to the L first extraction steps and the L second extraction steps, and the configured L sunburn monitoring and analysis sub-channels are used to monitor and analyze the L fruit growth monitoring data sets to obtain a target sunburn monitoring result; it is determined whether the target sunburn monitoring result meets a preset threshold, and if so, an early warning instruction is obtained, and the early warning instruction is sent to a user end for a sunburn warning. The technical effect of conducting dimensionality reduction analysis on citrus fruit sunburn monitoring data was achieved, improving monitoring accuracy, shortening the monitoring and analysis cycle, and improving early warning reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic flow chart of an intelligent citrus fruit sunburn monitoring and early warning method provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of a process for obtaining L divided grid areas in an intelligent citrus fruit sunburn monitoring and early warning method provided in an embodiment of the present application;

[0012] Figure 3 A schematic diagram of a process for updating a plurality of initially divided grid areas according to splitting results in an intelligent citrus fruit sunburn monitoring and early warning method provided in an embodiment of the present application;

[0013] Figure 4 A schematic diagram of the structure of an intelligent citrus fruit sunburn monitoring and early warning system provided in an embodiment of the present application.

[0014] Explanation of the reference numerals: a first generating unit 11 , a first obtaining unit 12 , a second generating unit 13 , a second obtaining unit 14 , a third obtaining unit 15 , a fourth obtaining unit 16 , a first sending unit 17 . DETAILED DESCRIPTION

[0015] The present application provides an intelligent citrus fruit sunburn monitoring and early warning method and system, which is used to solve the technical problems in the prior art of citrus fruit sunburn monitoring that the monitoring data is redundant, resulting in a long monitoring and analysis cycle and low early warning reliability.

[0016] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0017] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0018] Embodiment 1

[0019] like Figure 1 As shown, the present application provides an intelligent citrus fruit sunburn monitoring and early warning method, wherein the method further includes:

[0020] Collect the periodic growth records of the target citrus orchard within the preset historical window and generate K periodic growth record sets;

[0021] In a possible embodiment, the preset historical window is a preset time period for extracting the historical growth conditions of the target citrus orchard, which can be 3 years, 5 years, etc. The target citrus orchard is any citrus orchard that needs to be monitored and warned of fruit sunburn, regardless of its geographical location. The K cycle growth record sets are obtained by recording and retrieving the growth conditions of citrus fruits in the target citrus orchard in each citrus growth cycle within the preset historical window. Among them, the K cycle growth record sets reflect the growth conditions of citrus fruits in K citrus growth cycles, including the total number of fruits on the periphery of the plant crown (fruits that can be directly exposed to sunlight), the total fruit yield of fruit trees in the orchard, the number of sunburned fruits and other data. By obtaining K cycle growth record sets, a basis is provided for the subsequent reliable analysis of the growth conditions of citrus in the target citrus orchard, avoiding accidental errors caused by analyzing a single growth cycle, and achieving the technical effect of improving analysis reliability and ensuring data accuracy.

[0022] Based on the K periodic growth record sets, the target citrus orchard is grid-divided and merged to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers;

[0023] Further, such as Figure 2 As shown, based on the K periodic growth record sets, the target citrus orchard is grid-divided and fused to obtain L divided grid areas. The embodiment of the present application also includes:

[0024] Taking each fruit tree in the target citrus orchard as an index, searching the K periodic growth record sets to obtain Q periodic growth record sets of Q fruit trees, wherein the Q fruit trees have position coordinate identifiers;

[0025] Performing sunburn granularity analysis on the Q period growth record sets, and calculating the mean of the analysis results to obtain Q sunburn granularities, wherein the sunburn granularity is the degree of sunburn caused by the fruit trees being subjected to high temperature and strong light stress;

[0026] The target citrus orchard is grid-divided and fused based on the Q sunburn granularities to obtain the L divided grid areas.

[0027] Furthermore, based on the Q sunburn granularities, the target citrus orchard is grid-divided and merged to obtain the L divided grid areas. The embodiment of the present application further includes:

[0028] According to the position coordinates of the Q fruit trees corresponding to the Q sunburn granularities, the landform of the simulated basin is constructed in combination with the Q sunburn granularities;

[0029] The Q sunburn granularities are sorted in order from small to large, the sunburn granularity at the first position is used as the starting water injection point, and the simulated water injection is performed from bottom to top on the simulated basin. When the water surface after water injection exceeds the starting water injection point and submerges the first sunburn granularity, it is used as the first sunburn granularity. It is determined whether the difference between the first sunburn granularity and the sunburn granularity at the starting water injection point meets a preset threshold. If not, a grid dividing ridge line is generated at the corresponding position according to the position coordinate identifier of the first sunburn granularity, and then water injection is continued;

[0030] If yes, continue to inject water. When the water surface after injection submerges the next sunburn particle size, take it as the second sunburn particle size. Determine whether the difference between the second sunburn particle size and the sunburn particle size at the starting water injection point meets the preset threshold. If not, generate a grid dividing ridge line at the corresponding position according to the position coordinate mark of the second sunburn particle size, and then continue to inject water.

[0031] If yes, continue to inject water until the water surface submerges the position of the maximum solar scorch particle size in the simulated basin, stop injecting water into the simulated basin, and obtain multiple grid division ridge lines;

[0032] Grid division and fusion are performed according to the multiple grid division ridge lines to obtain L divided grid areas.

[0033] In one possible embodiment, after obtaining the K periodic growth record sets, the growth conditions of each fruit tree in the target citrus orchard are extracted and analyzed, and then the target citrus orchard is grid-divided and fused according to the growth conditions of the fruit trees based on the analysis results. That is, multiple fruit trees with similar growth conditions are divided into one grid area, thereby obtaining multiple divided grid areas. Among them, the L divided grid areas have L growth trend feature identifiers. The L growth trend feature identifiers are used to describe the growth conditions of the fruit trees in the L divided grid areas, respectively. By performing grid division and fusion, the target citrus orchard is divided into refined grid areas on the basis of ensuring that the divided grids fit the growth conditions of the fruit trees in the target citrus orchard, thereby achieving the technical effect of constructing a reasonable monitoring trajectory for the subsequent construction and improving monitoring efficiency.

[0034] In a possible embodiment, after obtaining the L divided grid areas, the mean number of mild sunburns of fruit trees in the L areas, the mean number of moderate sunburns of fruit trees in the L areas, and the mean number of severe sunburns of fruit trees in the L areas are extracted in combination with the K periodic growth record sets. The mean number of mild sunburns of fruit trees in the L areas, the mean number of moderate sunburns of fruit trees in the L areas, and the mean number of severe sunburns of fruit trees in the L areas are used as L growth trend features to identify the L divided grid areas.

[0035] In an embodiment of the present application, each fruit tree in the target citrus orchard is used as an index, and the K periodic growth record sets are retrieved to obtain Q periodic growth record sets of Q fruit trees, wherein the Q periodic growth record sets reflect the growth status of citrus fruits of Q fruit trees in different growth cycles. Wherein, the Q fruit trees have a position coordinate identifier. The position coordinate identifier is used to uniquely identify the position of the Q fruit trees in the target citrus orchard. Then, the number of whole fruits and the number of sunburned fruits of the Q fruit trees in each growth cycle in the Q periodic growth record sets are extracted respectively, and then the ratio of the number of sunburned fruits of the Q fruit trees in each growth cycle to the number of whole fruits of the Q fruit trees is calculated, and the ratio calculation results of multiple growth cycles of the Q fruit trees are averaged to obtain the Q sunburn granularity. Wherein, the sunburn granularity is the degree of sunburn caused by the fruit tree after being subjected to high temperature and strong light stress. The higher the sunburn granularity, the higher the degree of sunburn caused by the corresponding fruit tree after being subjected to high temperature and strong light stress. Furthermore, the target citrus orchard is divided into and fused into grids with the Q sunburn granularities as unique variables, thereby obtaining the L divided grid regions.

[0036] In one possible embodiment, the target citrus orchard is used as a simulated basin, the Q fruit trees in the target citrus orchard are used as basins with undulating topography in the simulated basin, and the Q sunburn granularities are used as the highest points of each undulating basin in the simulated basin to complete the construction of the topography of the simulated basin. Thus, accurate and efficient grid division and fusion of orchard areas based on the topography of the simulated basin are performed in the subsequent process. A hole is opened at the lowest point of the simulated basin, and water is poured into the basin from this hole. The water will flow evenly and equally to each basin. When the amount of water increases, the basin will be filled with water. When the water in two or more basins is filled, it is necessary to construct a grid division ridge line between the two basins to avoid the merging of the two basins, thereby achieving the goal of efficiently and accurately dividing multiple basins in the simulated basin. Based on the above principle, the Q sunburn granularities are sorted in order from small to large, and then the sunburn granularity at the first position is used as the starting water injection point, that is, a hole is dug at the lowest point in the basin corresponding to the minimum value of the Q sunburn granularities for water injection. Since the remaining multiple sunburn granularities are larger than the sunburn granularity at the starting water injection point, the sunburn granularity at the first position is used as the starting water injection point, and the simulated water injection is performed from bottom to top in the simulated basin. The water surface will first submerge the starting water injection point. When the water surface after water injection exceeds the starting water injection point and submerges the first sunburn granularity, it is used as the first sunburn granularity. Preferably, in order to avoid too fine grid division, resulting in too loose division of the target citrus orchard, it is necessary to determine whether the difference between the first sunburn granularity and the sunburn granularity at the starting water injection point meets the preset threshold, so as to determine whether it is necessary to set up a grid division ridge line to separate the two areas according to the judgment result. Among them, the preset threshold is the minimum sunburn granularity difference when the growth conditions of two fruit trees can be analyzed together, which is set by technicians in this field. If not, it indicates that the difference between the first sunburn particle size and the sunburn particle size at the starting water injection point is higher than the preset threshold, and the corresponding fruit trees need to be monitored separately. At this time, a grid division ridge line is generated at the fruit tree corresponding to the first sunburn particle size. The goal of regional division of fruit trees in the target citrus orchard is achieved.

[0037] Preferably, if yes, it indicates that the difference between the first sunburn particle size and the sunburn particle size at the starting water injection point is lower than the preset threshold value, and the corresponding fruit trees can be monitored together. At this time, water injection continues. When the water surface after injection submerges the next sunburn particle size, it is used as the second sunburn particle size, and it is determined whether the difference between the second sunburn particle size and the sunburn particle size at the starting water injection point meets the preset threshold value. The analysis is performed based on the same principle. If not, a grid dividing ridge line is generated at the corresponding position according to the position coordinate mark of the second sunburn particle size, and then water injection continues. If yes, water injection continues until the water surface submerges the position of the maximum value of the sunburn particle size in the simulated basin, and water injection into the simulated basin is stopped. The multiple grid division ridge lines obtained during the water injection process are counted, and the target citrus orchard is initially divided into regions using the multiple grid division ridge lines, thereby achieving the goal of dividing the regions according to the actual growth conditions of the fruit trees in the target citrus orchard. The division results are then subjected to grid division fusion, and grid areas that are too large or too small after division are split or merged, thereby obtaining the L divided grid areas.

[0038] Further, such as Figure 3 As shown, grid division and fusion are performed according to the multiple grid division ridge lines to obtain L divided grid areas. The embodiment of the present application also includes:

[0039] Dividing the target citrus orchard into a plurality of initial divided grid areas using the plurality of grid dividing ridge lines;

[0040] Traversing the plurality of initial divided grid regions to perform grid region density analysis, and taking a plurality of initial grid regions having a density higher than a preset grid region density threshold as a plurality of initial divided grid regions to be split;

[0041] The multiple initial grid areas to be split are split according to a preset splitting method, and the multiple initial divided grid areas are updated according to the splitting results, wherein the preset splitting method is to calculate the difference between the grid area density of the multiple initial grid areas to be split and the preset grid area density threshold, and the ratio of the multiple calculated differences to the preset grid area density threshold is used as the proportion of the split areas.

[0042] Furthermore, the embodiment of the present application also includes:

[0043] Performing grid area density analysis on the updated multiple initial divided grid areas again, and taking multiple initial divided grid areas with density below a preset grid area density threshold as multiple initial divided grid areas to be fused;

[0044] The multiple to-be-fused initial divided grid regions are respectively merged into the initial divided grid region with the lowest grid region density in the adjacent multiple initial divided grid regions to obtain L divided grid regions.

[0045] In an embodiment of the present application, the fruit trees in the target citrus orchard are divided along the coordinate positions corresponding to the plurality of grid division ridge lines, thereby obtaining a plurality of initial division grid areas. There are a plurality of fruit trees in each initial division grid area. The grid area density analysis is performed on the plurality of initial division grid areas respectively, that is, by counting the number of fruit trees in the plurality of initial division grid areas, and then comparing the grid area areas in the plurality of initial division grid areas, a plurality of grid area densities of the plurality of initial division grid areas are obtained. Among them, the plurality of grid area densities reflect the density of fruit trees in the plurality of initial division grid areas. The plurality of grid area densities of the plurality of initial division grid areas are traversed, and are compared with the preset grid area density thresholds respectively, and the plurality of initial division grid areas higher than the preset grid area density thresholds are used as a plurality of initial division grid areas to be split. That is, when the grid area density is higher than the preset grid area density threshold, it indicates that there are too many fruit trees gathered in the grid area, which is likely to cause an excessive amount of monitoring data and prolong the monitoring and analysis cycle, so splitting is required. Preferably, the multiple initial grid areas to be split are split according to a preset splitting method, and the multiple initial divided grid areas are updated according to the splitting results. The preset splitting method is to calculate the difference between the grid area density of the multiple initial grid areas to be split and the preset grid area density threshold, and the ratio of the calculated multiple differences to the preset grid area density threshold is used as the proportion of the split area. Thus, the multiple initial grid areas to be split are split with minimum loss, and the monitoring resources are fully utilized as a foundation.

[0046] In one possible embodiment, the grid area density analysis is performed again on the updated multiple initial divided grid areas, and the multiple initial divided grid areas that are lower than the preset grid area density threshold are used as multiple initial divided grid areas to be fused. When the grid area density is too low, it is easy to cause a waste of monitoring resources, and it is easy to cause the monitoring equipment in multiple initial divided grid areas to be fused to have completed their work while the monitoring equipment in other grid areas is still working, thereby making the monitoring resources unevenly distributed and unable to be fully utilized to improve the monitoring efficiency. Then, the multiple initial divided grid areas to be fused are respectively integrated into the initial divided grid area with the lowest grid area density among the multiple adjacent initial divided grid areas, and the grid area fusion is completed to obtain L divided grid areas. The technical effect of dividing the grid area according to the actual situation of the target citrus orchard, improving the efficiency and accuracy of monitoring and analysis, and improving the reliability of early warning is achieved.

[0047] Perform monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories;

[0048] Monitor the fruit growth trend of the target citrus orchard according to the L first monitoring trajectories to obtain L fruit growth monitoring data sets;

[0049] In one possible embodiment, the fruit tree position coordinates of L divided grid areas are collected to generate L fruit tree position coordinate sets. Monitoring trajectory analysis is performed based on the L fruit tree position coordinate sets respectively, so as to obtain L first monitoring trajectories. Wherein, the L first monitoring trajectories are the routes taken when the citrus trees in the L divided grid areas are monitored for sunburn. Preferably, by obtaining multiple sample fruit tree position coordinate sets and multiple sample first monitoring trajectories as training data, supervised training is performed on the network layer constructed based on the convolutional neural network until the output reaches convergence, so as to obtain a trained trajectory analysis network layer. Wherein, the trajectory analysis network layer is used to perform intelligent monitoring trajectory analysis based on the distribution of fruit tree coordinate positions. The L fruit tree position coordinate sets are transmitted to the trajectory analysis network layer to obtain L first monitoring trajectories.

[0050] In one embodiment, after obtaining the L first monitoring tracks, an image acquisition device such as a drone is used to acquire images of the fruit conditions outside the fruit trees in the target citrus orchard according to the L first monitoring tracks, and the fruit images acquired by monitoring are used as fruit growth monitoring data, thereby obtaining L fruit growth monitoring data sets. The L fruit growth monitoring data sets are used to describe the fruit growth conditions of the citrus trees in the L divided grid areas.

[0051] Using L growth trend feature identifiers, a multi-channel extraction analysis is performed on the target citrus orchard to obtain L first extraction steps and L second extraction steps;

[0052] Furthermore, L growth trend feature identifiers are used to perform multi-channel extraction analysis on the target citrus orchard to obtain L first extraction steps and L second extraction steps. The embodiment of the present application also includes:

[0053] Retrieve according to the L growth trend characteristic identifiers respectively to obtain L citrus fruit sunburn first densities and L citrus fruit sunburn second densities, wherein the L citrus fruit sunburn first densities are the aggregation degrees of citrus fruits with mild sunburn in the sample citrus orchard with the L growth trend characteristic identifiers, and the L citrus fruit sunburn second densities are the aggregation degrees of citrus fruits with severe sunburn in the sample citrus orchard with the L growth trend characteristic identifiers;

[0054] Respectively calculating the ratio of the L citrus fruit sunburn first densities to the sum of the L citrus fruit sunburn first densities, and multiplying the ratio by a preset extraction step length, to obtain L first extraction step lengths, wherein the first extraction step length is smaller than the second extraction step length, and the first extraction step length is the minimum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard;

[0055] Calculate the ratio of the L citrus fruit sunburn second densities to the sum of the L citrus fruit sunburn second densities respectively, and multiply the ratio by the preset extraction step length to obtain L second extraction step lengths, wherein the second extraction step length is the maximum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard.

[0056] In one possible embodiment, the monitoring data of the L grid-divided areas in the target citrus orchard are subjected to multi-channel extraction using the L growth trend feature identifiers to obtain L first extraction step sizes and L second extraction step sizes, thereby achieving the goal of reducing the amount of monitoring data and improving the efficiency of monitoring analysis while ensuring the accuracy of the monitoring data. The first extraction step size is the minimum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard. The second extraction step size is the maximum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard.

[0057] In one embodiment, the search is performed according to the L growth trend characteristic identifiers, and the mean number of mild sunburn of fruit trees in L regions and the mean number of severe sunburn of fruit trees in L regions are extracted. Then, the extracted data are respectively compared with the total amount of fruit in the L divided grid areas to obtain L citrus fruit sunburn first densities and L citrus fruit sunburn second densities. The L citrus fruit sunburn first densities are the aggregation degree of citrus fruits with mild sunburn in the sample citrus orchards with L growth trend characteristic identifiers, and the L citrus fruit sunburn second densities are the aggregation degree of citrus fruits with severe sunburn in the sample citrus orchards with L growth trend characteristic identifiers. Furthermore, the ratio of the L citrus fruit sunburn first densities to the sum of the L citrus fruit sunburn first densities is calculated respectively, and the ratio is multiplied by the preset extraction step length, and L first extraction step lengths are obtained, wherein the first extraction step length is smaller than the second extraction step length. The preset extraction step length is a time period pre-set by a person skilled in the art for extracting monitoring data obtained by trend monitoring of the target citrus orchard and is the interval between two adjacent monitoring data.

[0058] configuring L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel according to the L first extraction steps and the L second extraction steps, and using the configured L sunburn monitoring and analysis sub-channels to monitor and analyze the L fruit growth monitoring data sets to obtain target sunburn monitoring results;

[0059] Furthermore, the embodiment of the present application also includes:

[0060] Constructing L sunburn monitoring and analysis sub-channels, wherein the L sunburn monitoring and analysis sub-channels respectively have L first monitoring branches, L second monitoring branches and L monitoring data analysis network layers;

[0061] configuring the L first monitoring branches and the L second monitoring branches according to the L first extraction step lengths and the L second extraction step lengths respectively;

[0062] Acquire multiple sample fruit growth monitoring data sets, multiple sample first fruit monitoring data sets, and multiple sample second fruit monitoring data sets to perform supervised training on the L first monitoring branches and the L second monitoring branches, and use multiple sample first fruit monitoring data sets, multiple sample second fruit monitoring data sets, and multiple sample target sunburn monitoring results to perform supervised training on the L monitoring data analysis network layers, and update parameters according to the output results during the training process until the training is completed at convergence;

[0063] The L sunburn monitoring and analysis sub-channels are fully connected in parallel to generate the sunburn monitoring and analysis channel.

[0064] In a possible embodiment, L sunburn monitoring and analysis subchannels of the sunburn monitoring and analysis channel are configured according to the L first extraction step lengths and the L second extraction step lengths, so that the L sunburn monitoring and analysis subchannels can fit the growth conditions of fruit trees in the L divided grid areas. The configured L sunburn monitoring and analysis subchannels are used to monitor and analyze the L fruit growth monitoring data sets to obtain target sunburn monitoring results. Among them, the target sunburn monitoring results are the sunburn monitoring conditions of fruit trees in the target citrus orchard. The sunburn monitoring and analysis channel is a functional channel for intelligently analyzing the sunburn conditions of fruits in the target citrus orchard due to high temperature and strong light stress, including L sunburn monitoring and analysis subchannels. The L sunburn monitoring and analysis subchannels are functional channels for intelligently analyzing the sunburn conditions of fruits in the L divided grid areas. The L sunburn monitoring and analysis subchannels respectively have L first monitoring branches, L second monitoring branches and L monitoring data analysis network layers.

[0065] Among them, the L first monitoring branches and the L second monitoring branches are respectively connected in parallel, and the output layers of the first monitoring branch and the second monitoring branch are connected to the input layer of the monitoring data analysis network layer. The first monitoring branch is used to extract the fruit growth monitoring data set according to the first extraction step length, and analyze the mild sunburn fruit situation that is not obvious. The second monitoring branch is used to extract the fruit growth monitoring data set according to the second extraction step length, and analyze the severe sunburn fruit situation that is not obvious. Since the moderate sunburn fruit is between the two, it is extracted in the first monitoring branch. The first monitoring branch extracts a large amount of data and has more analysis details, which can accurately identify the mild and moderate fruit sunburn situation, while the second monitoring branch extracts a small amount of data, which can quickly analyze the severe fruit sunburn situation that is more obvious. In this way, the goal of ensuring that all fruit conditions are monitored and analyzed, ensuring comprehensive monitoring, and having targeted monitoring is achieved. The monitoring data analysis network layer is used to perform intelligent analysis of the fruit sunburn situation on the monitoring data extracted by the first monitoring branch and the second monitoring branch after being configured according to the first extraction step length and the second extraction step length, and output the target sunburn monitoring result.

[0066] In one embodiment, the L first monitoring branches and the L second monitoring branches are configured according to the L first extraction step lengths and the L second extraction step lengths, respectively. After the configuration is completed, the obtained multiple sample fruit growth monitoring data sets are used as input constraints of the L first monitoring branches and the L second monitoring branches, and the multiple sample first fruit monitoring data sets and the multiple sample second fruit monitoring data sets are used as output constraints of the L first monitoring branches and the L second monitoring branches, respectively, and multiple frameworks constructed based on the convolutional neural network are supervised and trained until the output reaches convergence, thereby obtaining the L first monitoring branches and the L second monitoring branches. Furthermore, with the multiple sample first fruit monitoring data sets and the multiple sample second fruit monitoring data sets as input constraints, and the multiple sample target sunburn monitoring results as output constraints, multiple frameworks based on the convolutional neural network framework are supervised and trained, and parameters are updated according to the output results during the training process until the output reaches convergence, and the trained L monitoring data analysis network layers are obtained. The sunburn monitoring and analysis channel is generated by parallelly fully connecting the L sunburn monitoring and analysis subchannels.

[0067] It is determined whether the target sunburn monitoring result meets a preset threshold value. If so, an early warning instruction is obtained and sent to the user end for sunburn warning.

[0068] In an embodiment of the present application, it is determined whether the target sunburn monitoring result meets the minimum sunburn monitoring result for normal growth of fruit trees in a citrus orchard set by a person skilled in the art. If so, it indicates that the sunburn of citrus fruits in the target citrus orchard exceeds the allowable range, and the early warning instruction is obtained. Among them, the early warning instruction is used to remind the staff that the sunburn of citrus fruits in the target citrus orchard needs to be dealt with. Then, the early warning instruction is sent to the user end for sunburn warning. The technical effect of performing refined and reliable monitoring of the sunburn of citrus fruits, improving the reliability of fruit sunburn monitoring and early warning, and improving monitoring efficiency is achieved.

[0069] In summary, the embodiments of the present application have at least the following technical effects:

[0070] The present application collects periodic growth records of a target citrus orchard within a preset historical window to generate K periodic growth record sets, thereby achieving the goal of providing basic analysis data for subsequent grid division and fusion. Then, grid division and fusion are performed on the target citrus orchard based on the K periodic growth record sets to obtain L divided grid areas, thereby achieving the goal of dividing the orchard into areas that fit the actual fruit conditions. Monitoring trajectory analysis is then performed based on the L divided grid areas to generate L first monitoring trajectories. Fruit growth trend monitoring is performed on the target citrus orchard based on the L first monitoring trajectories, thereby achieving the goal of providing monitoring data for subsequent fruit sunburn analysis. The L growth trend features are used to generate L first monitoring trajectories. The target citrus orchard is subjected to multi-channel extraction and analysis based on the signature, and L first extraction steps and L second extraction steps are obtained, achieving the goal of performing differentiated monitoring data extraction and improving the efficiency of monitoring and analysis. Then, L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel are configured according to the L first extraction steps and L second extraction steps. The configured L sunburn monitoring and analysis sub-channels are used to monitor and analyze the L fruit growth monitoring data sets, and the target sunburn monitoring results are obtained, achieving the goal of intelligent fruit sunburn monitoring and analysis. It is determined whether the target sunburn monitoring results meet the preset threshold. If so, an early warning instruction is obtained and sent to the user end for sunburn warning. The technical effect of performing multi-channel extraction and analysis on citrus fruit sunburn monitoring data, improving monitoring efficiency, improving the fit between monitoring and early warning and the actual situation of the citrus orchard, and reducing false warnings is achieved.

[0071] Embodiment 2

[0072] Based on the same inventive concept as the intelligent citrus fruit sunburn monitoring and early warning method in the aforementioned embodiment, Figure 4 As shown, the present application provides an intelligent citrus fruit sunburn monitoring and early warning system, and the system and method embodiments in the present application embodiments are based on the same inventive concept. The system includes:

[0073] The first generating unit 11 is used to collect periodic growth records of the target citrus orchard within a preset historical window and generate K periodic growth record sets;

[0074] A first obtaining unit 12 is used to perform grid division and fusion on the target citrus orchard based on the K periodic growth record sets to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers;

[0075] A second generating unit 13 is used to perform monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories;

[0076] The second obtaining unit 14 is used to monitor the fruit growth trend of the target citrus orchard according to the L first monitoring trajectories to obtain L fruit growth monitoring data sets;

[0077] The third obtaining unit 15 is used to perform multi-channel extraction analysis on the target citrus orchard using the L growth trend feature identifiers to obtain L first extraction step lengths and L second extraction step lengths;

[0078] a fourth obtaining unit 16, configured to configure L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel according to the L first extraction step lengths and the L second extraction step lengths, and monitor and analyze the L fruit growth monitoring data sets using the configured L sunburn monitoring and analysis sub-channels to obtain a target sunburn monitoring result;

[0079] The first sending unit 17 is used to determine whether the target sunburn monitoring result meets a preset threshold value. If so, a warning instruction is obtained and the warning instruction is sent to the user terminal for sunburn warning.

[0080] Furthermore, the system also includes:

[0081] Taking each fruit tree in the target citrus orchard as an index, searching the K periodic growth record sets to obtain Q periodic growth record sets of Q fruit trees, wherein the Q fruit trees have position coordinate identifiers;

[0082] Performing sunburn granularity analysis on the Q period growth record sets, and calculating the mean of the analysis results to obtain Q sunburn granularities, wherein the sunburn granularity is the degree of sunburn caused by the fruit trees being subjected to high temperature and strong light stress;

[0083] The target citrus orchard is grid-divided and fused based on the Q sunburn granularities to obtain the L divided grid areas.

[0084] Furthermore, the system also includes:

[0085] According to the position coordinates of the Q fruit trees corresponding to the Q sunburn granularities, the landform of the simulated basin is constructed in combination with the Q sunburn granularities;

[0086] The Q sunburn granularities are sorted in order from small to large, the sunburn granularity at the first position is used as the starting water injection point, and the simulated water injection is performed from bottom to top on the simulated basin. When the water surface after water injection exceeds the starting water injection point and submerges the first sunburn granularity, it is used as the first sunburn granularity. It is determined whether the difference between the first sunburn granularity and the sunburn granularity at the starting water injection point meets a preset threshold. If not, a grid dividing ridge line is generated at the corresponding position according to the position coordinate identifier of the first sunburn granularity, and then water injection is continued;

[0087] If yes, continue to inject water. When the water surface after injection submerges the next sunburn particle size, take it as the second sunburn particle size. Determine whether the difference between the second sunburn particle size and the sunburn particle size at the starting water injection point meets the preset threshold. If not, generate a grid dividing ridge line at the corresponding position according to the position coordinate mark of the second sunburn particle size, and then continue to inject water.

[0088] If yes, continue to inject water until the water surface submerges the position of the maximum solar scorch particle size in the simulated basin, stop injecting water into the simulated basin, and obtain multiple grid division ridge lines;

[0089] Grid division and fusion are performed according to the multiple grid division ridge lines to obtain L divided grid areas.

[0090] Furthermore, the system also includes:

[0091] Dividing the target citrus orchard into a plurality of initial divided grid areas using the plurality of grid dividing ridge lines;

[0092] Traversing the plurality of initial divided grid regions to perform grid region density analysis, and taking a plurality of initial grid regions having a density higher than a preset grid region density threshold as a plurality of initial divided grid regions to be split;

[0093] The multiple initial grid areas to be split are split according to a preset splitting method, and the multiple initial divided grid areas are updated according to the splitting results, wherein the preset splitting method is to calculate the difference between the grid area density of the multiple initial grid areas to be split and the preset grid area density threshold, and the ratio of the multiple calculated differences to the preset grid area density threshold is used as the proportion of the split areas.

[0094] Furthermore, the system also includes:

[0095] Performing grid area density analysis on the updated multiple initial divided grid areas again, and taking multiple initial divided grid areas with density below a preset grid area density threshold as multiple initial divided grid areas to be fused;

[0096] The multiple to-be-fused initial divided grid regions are respectively merged into the initial divided grid region with the lowest grid region density in the adjacent multiple initial divided grid regions to obtain L divided grid regions.

[0097] Furthermore, the system also includes:

[0098] Retrieve according to the L growth trend characteristic identifiers respectively to obtain L citrus fruit sunburn first densities and L citrus fruit sunburn second densities, wherein the L citrus fruit sunburn first densities are the aggregation degrees of citrus fruits with mild sunburn in the sample citrus orchard with the L growth trend characteristic identifiers, and the L citrus fruit sunburn second densities are the aggregation degrees of citrus fruits with severe sunburn in the sample citrus orchard with the L growth trend characteristic identifiers;

[0099] Respectively calculating the ratio of the L citrus fruit sunburn first densities to the sum of the L citrus fruit sunburn first densities, and multiplying the ratio by a preset extraction step length, to obtain L first extraction step lengths, wherein the first extraction step length is smaller than the second extraction step length, and the first extraction step length is the minimum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard;

[0100] Calculate the ratio of the L citrus fruit sunburn second densities to the sum of the L citrus fruit sunburn second densities respectively, and multiply the ratio by the preset extraction step length to obtain L second extraction step lengths, wherein the second extraction step length is the maximum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard.

[0101] Furthermore, the system also includes:

[0102] Constructing L sunburn monitoring and analysis sub-channels, wherein the L sunburn monitoring and analysis sub-channels respectively have L first monitoring branches, L second monitoring branches and L monitoring data analysis network layers;

[0103] configuring the L first monitoring branches and the L second monitoring branches according to the L first extraction step lengths and the L second extraction step lengths respectively;

[0104] Acquire multiple sample fruit growth monitoring data sets, multiple sample first fruit monitoring data sets, and multiple sample second fruit monitoring data sets to perform supervised training on the L first monitoring branches and the L second monitoring branches, and use multiple sample first fruit monitoring data sets, multiple sample second fruit monitoring data sets, and multiple sample target sunburn monitoring results to perform supervised training on the L monitoring data analysis network layers, and update parameters according to the output results during the training process until the training is completed at convergence;

[0105] The L sunburn monitoring and analysis sub-channels are fully connected in parallel to generate the sunburn monitoring and analysis channel.

[0106] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0107] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0108] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An intelligent citrus fruit sunburn monitoring and early warning method, characterized in that: The method further comprises: Collect the periodic growth records of the target citrus orchard within the preset historical window and generate K periodic growth record sets; Based on the K periodic growth record sets, the target citrus orchard is grid-divided and merged to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers; Perform monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories; Monitor the fruit growth trend of the target citrus orchard according to the L first monitoring trajectories to obtain L fruit growth monitoring data sets; Using L growth trend feature identifiers, a multi-channel extraction analysis is performed on the target citrus orchard to obtain L first extraction steps and L second extraction steps; configuring L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel according to the L first extraction steps and the L second extraction steps, and using the configured L sunburn monitoring and analysis sub-channels to monitor and analyze the L fruit growth monitoring data sets to obtain target sunburn monitoring results; Determine whether the target sunburn monitoring result meets a preset threshold value, and if so, obtain an early warning instruction, and send the early warning instruction to the user end for sunburn early warning; Wherein, based on the K periodic growth record sets, the target citrus orchard is grid-divided and merged to obtain L divided grid areas, and the method further includes: Taking each fruit tree in the target citrus orchard as an index, searching the K periodic growth record sets to obtain Q periodic growth record sets of Q fruit trees, wherein the Q fruit trees have position coordinate identifiers; Performing sunburn granularity analysis on the Q period growth record sets, and calculating the mean of the analysis results to obtain Q sunburn granularities, wherein the sunburn granularity is the degree of sunburn caused by the fruit trees being subjected to high temperature and strong light stress; Based on the Q sunburn granularities, grid division and fusion are performed on the target citrus orchard to obtain the L divided grid areas; Wherein, L growth trend feature identifiers are used to perform multi-channel extraction analysis on the target citrus orchard to obtain L first extraction step lengths and L second extraction step lengths, and the method further includes: Retrieve according to the L growth trend characteristic identifiers respectively to obtain L citrus fruit sunburn first densities and L citrus fruit sunburn second densities, wherein the L citrus fruit sunburn first densities are the aggregation degrees of citrus fruits with mild sunburn in the sample citrus orchard with the L growth trend characteristic identifiers, and the L citrus fruit sunburn second densities are the aggregation degrees of citrus fruits with severe sunburn in the sample citrus orchard with the L growth trend characteristic identifiers; Respectively calculating the ratio of the L citrus fruit sunburn first densities to the sum of the L citrus fruit sunburn first densities, and multiplying the ratio by a preset extraction step length, to obtain L first extraction step lengths, wherein the first extraction step length is smaller than the second extraction step length, and the first extraction step length is the minimum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard; Calculate the ratio of the L citrus fruit sunburn second densities to the sum of the L citrus fruit sunburn second densities respectively, and multiply the ratio by the preset extraction step length to obtain L second extraction step lengths, wherein the second extraction step length is the maximum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard.

2. The method according to claim 1, characterized in that Based on the Q sunburn granularities, the target citrus orchard is grid-divided and merged to obtain the L divided grid areas, and the method further includes: According to the position coordinates of the Q fruit trees corresponding to the Q sunburn granularities, the landform of the simulated basin is constructed in combination with the Q sunburn granularities; The Q sunburn granularities are sorted in order from small to large, the sunburn granularity at the first position is used as the starting water injection point, and the simulated water injection is performed from bottom to top on the simulated basin. When the water surface after water injection exceeds the starting water injection point and submerges the first sunburn granularity, it is used as the first sunburn granularity. It is determined whether the difference between the first sunburn granularity and the sunburn granularity at the starting water injection point meets a preset threshold. If not, a grid dividing ridge line is generated at the corresponding position according to the position coordinate identifier of the first sunburn granularity, and then water injection is continued; If yes, continue to inject water. When the water surface after injection submerges the next sunburn particle size, take it as the second sunburn particle size. Determine whether the difference between the second sunburn particle size and the sunburn particle size at the starting water injection point meets the preset threshold. If not, generate a grid dividing ridge line at the corresponding position according to the position coordinate mark of the second sunburn particle size, and then continue to inject water. If yes, continue to inject water until the water surface submerges the position of the maximum solar scorch particle size in the simulated basin, stop injecting water into the simulated basin, and obtain multiple grid division ridge lines; Grid division and fusion are performed according to the multiple grid division ridge lines to obtain L divided grid areas.

3. The method according to claim 2, characterized in that Grid division and fusion are performed according to the multiple grid division ridge lines to obtain L divided grid areas, and the method further includes: Dividing the target citrus orchard into a plurality of initial divided grid areas using the plurality of grid dividing ridge lines; Traversing the plurality of initial divided grid regions to perform grid region density analysis, and taking a plurality of initial divided grid regions having a density higher than a preset grid region density threshold as a plurality of initial divided grid regions to be split; The multiple initial divided grid areas to be split are split according to a preset splitting method, and the multiple initial divided grid areas are updated according to the splitting results, wherein the preset splitting method is to calculate the difference between the grid area density of the multiple initial divided grid areas to be split and a preset grid area density threshold, and use the multiple ratios of the calculated multiple differences to the preset grid area density threshold as the ratio of the multiple areas split from the multiple initial divided grid areas to be split to the corresponding multiple initial divided grid areas to be split.

4. The method according to claim 3, characterized in that The method further comprises: Performing grid area density analysis on the updated multiple initial divided grid areas again, and taking multiple initial divided grid areas with density below a preset grid area density threshold as multiple initial divided grid areas to be fused; The multiple to-be-fused initial divided grid regions are respectively merged into the initial divided grid region with the lowest grid region density in the adjacent multiple initial divided grid regions to obtain L divided grid regions.

5. The method according to claim 1, characterized in that The method further comprises: Constructing L sunburn monitoring and analysis sub-channels, wherein the L sunburn monitoring and analysis sub-channels respectively have L first monitoring branches, L second monitoring branches and L monitoring data analysis network layers; configuring the L first monitoring branches and the L second monitoring branches according to the L first extraction step lengths and the L second extraction step lengths respectively; Acquire multiple sample fruit growth monitoring data sets, multiple sample first fruit monitoring data sets, and multiple sample second fruit monitoring data sets to perform supervised training on the L first monitoring branches and the L second monitoring branches, and use multiple sample first fruit monitoring data sets, multiple sample second fruit monitoring data sets, and multiple sample target sunburn monitoring results to perform supervised training on the L monitoring data analysis network layers, and update parameters according to the output results during the training process until the training is completed at convergence; wherein, the L first monitoring branches and the L second monitoring branches are configured according to the L first extraction step lengths and the L second extraction step lengths respectively; the obtained multiple sample fruit growth monitoring data sets are used as input constraints of the L first monitoring branches and the L second monitoring branches, the multiple sample first fruit monitoring data sets and the multiple sample second fruit monitoring data sets are used as output constraints of the L first monitoring branches and the L second monitoring branches respectively, and the multiple frameworks constructed based on the convolutional neural network are supervised and trained until the output reaches convergence, thereby obtaining the L first monitoring branches and the L second monitoring branches; Taking a plurality of sample first fruit monitoring data sets and a plurality of sample second fruit monitoring data sets as input constraints and a plurality of sample target sunburn monitoring results as output constraints, multiple frameworks based on the convolutional neural network architecture are supervised and trained, and parameters are updated according to the output results during the training process until the output converges, thereby obtaining L monitoring data analysis network layers that have been trained; The L sunburn monitoring and analysis sub-channels are fully connected in parallel to generate the sunburn monitoring and analysis channel.

6. An intelligent citrus fruit sunburn monitoring and early warning system, characterized in that: The system comprises: The first generating unit is used to collect the periodic growth records of the target citrus orchard within a preset historical window and generate K periodic growth record sets; A first obtaining unit is used to perform grid division and fusion on the target citrus orchard based on the K periodic growth record sets to obtain L divided grid areas, wherein the L divided grid areas have L growth trend feature identifiers; A second generating unit is used to perform monitoring trajectory analysis based on the L divided grid areas to generate L first monitoring trajectories; A second obtaining unit is used to monitor the fruit growth trend of the target citrus orchard according to the L first monitoring trajectories to obtain L fruit growth monitoring data sets; A third obtaining unit is used to perform multi-channel extraction analysis on the target citrus orchard using the L growth trend feature identifiers to obtain L first extraction step lengths and L second extraction step lengths; a fourth obtaining unit, configured to configure L sunburn monitoring and analysis sub-channels of the sunburn monitoring and analysis channel according to the L first extraction step lengths and the L second extraction step lengths, and to monitor and analyze the L fruit growth monitoring data sets using the configured L sunburn monitoring and analysis sub-channels to obtain a target sunburn monitoring result; A first sending unit is used to determine whether the target sunburn monitoring result meets a preset threshold value, and if so, obtain an early warning instruction, and send the early warning instruction to the user end for sunburn early warning; The first obtaining unit is further used to retrieve the K periodic growth record sets by taking each fruit tree in the target citrus orchard as an index to obtain Q periodic growth record sets of Q fruit trees, wherein the Q fruit trees have position coordinate identifiers; Performing sunburn granularity analysis on the Q period growth record sets, and calculating the mean of the analysis results to obtain Q sunburn granularities, wherein the sunburn granularity is the degree of sunburn caused by the fruit trees being subjected to high temperature and strong light stress; Based on the Q sunburn granularities, grid division and fusion are performed on the target citrus orchard to obtain the L divided grid areas; The third obtaining unit is further used to perform retrieval according to the L growth trend characteristic identifiers, respectively, to obtain L citrus fruit sunburn first densities and L citrus fruit sunburn second densities, wherein the L citrus fruit sunburn first densities are the aggregation degrees of citrus fruits with mild sunburn in the sample citrus orchard with the L growth trend characteristic identifiers, and the L citrus fruit sunburn second densities are the aggregation degrees of citrus fruits with severe sunburn in the sample citrus orchard with the L growth trend characteristic identifiers; Respectively calculating the ratio of the L citrus fruit sunburn first densities to the sum of the L citrus fruit sunburn first densities, and multiplying the ratio by a preset extraction step length, to obtain L first extraction step lengths, wherein the first extraction step length is smaller than the second extraction step length, and the first extraction step length is the minimum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard; Calculate the ratio of the L citrus fruit sunburn second densities to the sum of the L citrus fruit sunburn second densities respectively, and multiply the ratio by the preset extraction step length to obtain L second extraction step lengths, wherein the second extraction step length is the maximum time period between two adjacent monitoring data when extracting the monitoring data obtained by trend monitoring of the target citrus orchard.

Images