A corn staged fertilization control method based on internet of things

Abstract

The present application relates to the technical field of intelligent internet of things mechanical fertilization control, and particularly relates to a corn staged fertilization control method based on internet of things. First, the present application obtains various nutrient data, hyperspectral images and growth stages of corn plants; further, the present application fuses overall nutrient data and overall NDVI value in each corn region to obtain growth suitability; further, the present application obtains real-time growth influence coefficients of each plant pixel point according to time sequence fluctuation of growth suitability of the same corn region, in combination with differences between growth suitability of the corn region and adjacent corn regions, and area of the corn region; further, the present application analyzes fluctuation of the growth influence coefficients and spatial offset of the corn region, in combination with growth retardation of the corn plant, to obtain real-time growth interference accumulation coefficients; finally, the present application realizes precise control fertilization in a staged and regional manner according to all growth interference accumulation coefficients corresponding to each corn region in real time, in combination with low degree of each nutrient data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent Internet of Things mechanical fertilization control, and particularly relates to a corn staged fertilization control method based on an Internet of Things. BACKGROUND

[0002] Modern agriculture precisely manages corn fertilization through data driving, for example, partition monitoring of a corn field, real-time collection of soil humidity, nitrogen, phosphorus and potassium content to analyze soil fertility, identification of a growth stage of the corn plant in combination with the growth condition of the corn plant, and evaluation of nutrient requirements, partition fertilization of the corn field by a drip irrigation system or a drone based on the nutrient requirements.

[0003] However, the existing fertilization control method has poor control accuracy in Abiotic Stress Soils, for example, when the existing method is used to control fertilization in a drought area (for example, Gansu), the soil nutrients in the area are greatly affected by the water content in different regions, for example, when the water content is low, the dissolution and migration of nutrient ions are disturbed, which is not conducive to the absorption of nutrients by the corn plant roots, and the influence of topography and rainfall on water transport in the soil may cause certain migration in the drought area, further affecting the content of absorbable nutrients in the soil, so that it is difficult to accurately supplement fertilization according to the actual situation of the soil in different regions and the growth state of the crops. SUMMARY

[0004] In order to solve the technical problem of inaccurate fertilization control in Abiotic Stress Soils, the present application aims to provide a corn staged fertilization control method based on an Internet of Things, and the technical solution is as follows:

[0005] Nutrient data, hyperspectral images and NDVI values of the corn field are obtained at a preset collection frequency, and the growth stages of different corn plants are identified;

[0006] In the pixel points of the corn plants in each of the hyperspectral images, a corn region is obtained according to the similarity characteristics of the pixel points and adjacent pixel points; the growth suitability of each corn region is obtained by fusing the overall nutrient data and the overall NDVI value in each corn region;

[0007] The corn regions between adjacent frames of the hyperspectral images are matched; the real-time growth influence coefficient of each plant pixel point is obtained according to the time sequence fluctuation of the growth suitability of the same corn region, in combination with the difference in the growth suitability of the corn region and adjacent corn regions, and the area of the corn region;

[0008] According to the fluctuation of the growth influence coefficient of the single plant pixel point and the gravity change of the corn region to which the single plant pixel point belongs, and the growth delay degree and the growth stage of the corn plant to which the single plant pixel point belongs, the real-time growth disturbance accumulation coefficient of the corresponding plant pixel point is obtained; and according to all the growth disturbance accumulation coefficients of each corn region in real time, in combination with the low degree of each nutrient data, fertilization is controlled.

[0009] Further, the corn region acquisition method comprises:

[0010] In any of the hyperspectral images, the hyperspectral curve of each pixel point is converted into a hyperspectral vector; based on region growing, in all the pixel points of the corn plants, the cosine similarity of the hyperspectral vectors of the pixel point and the adjacent pixel points in the preset neighborhood of the pixel point is less than or equal to a preset similarity threshold as a growth termination condition, and the corn field is regionally divided.

[0011] Further, the growth suitability acquisition method comprises:

[0012] At the acquisition time corresponding to any of the hyperspectral images, the average of the nutrient data of all the monitoring points in each of the corn regions and the average of the NDVI of all the pixel points are taken as PCA input, the load vector of PC1 is taken as the weighted weight of each index and weighted summation is performed, and the growth suitability of each of the corn regions at the corresponding acquisition time is obtained.

[0013] Further, the growth influence coefficient acquisition method comprises:

[0014] Any of the corn regions is selected as a target region, and based on dividing a fluctuation period between the maximum value of the growth suitability and the maximum value and the first acquisition time;

[0015] According to the duration, the growth suitability and the maximum growth suitability of the target region at each of the fluctuation periods, an effective growth coefficient of each of the fluctuation periods is obtained; at the current time, according to the number and the average absolute value of the negative difference values in the first-order difference sequence of the effective growth coefficient, in combination with the first effective growth coefficient in the time sequence, a first influence factor is obtained;

[0016] According to the difference between the growth suitability of the target region and the adjacent corn regions thereof at all acquisition times and the area of the target region, a second influence factor is obtained;

[0017] The first influence factor and the second influence factor at the current time are fused to obtain the real-time growth influence coefficient of the target region as the real-time growth influence coefficient of each pixel point in the target region.

[0018] Further, the method for obtaining the second influence factor comprises:

[0019] At each acquisition time, a growth difference coefficient of each acquisition time is obtained according to the difference between the growth suitability of the target region and the corn region adjacent thereto; and a second influence factor is obtained by fusing the mean value of the growth difference coefficient of the target region of all acquisition times and the mean value of the area.

[0020] Further, the method for obtaining the growth interference accumulation coefficient comprises:

[0021] Any plant pixel point is selected as a target pixel point; a mean value of a first-order difference value of the growth influence coefficient of the target pixel point is obtained as an influence growth factor;

[0022] All the corn regions to which the target pixel point once belonged are obtained as to-be-analyzed regions; a displacement factor is obtained according to the distance between the barycenter of all the to-be-analyzed regions and the corn region to which the first acquisition time belongs;

[0023] According to the duration of all complete growth stages of the corn plant corresponding to the target pixel point, the overgrowth degree relative to the corresponding preset standard growth duration, and in combination with the influence growth factor and the displacement factor, a real-time growth interference accumulation coefficient of the target pixel point is obtained.

[0024] Further, the method for controlling fertilization comprises:

[0025] In the current hyperspectral image, a real-time accumulation coefficient of each corn region is obtained by fusing the growth interference accumulation coefficients of all pixel points in each corn region;

[0026] A nutrient substance is selected as a target nutrient substance; a fertilization demand degree is obtained by combining the real-time accumulation coefficient according to the difference between a preset minimum concentration and a measured concentration of the target nutrient substance required by the plant in any corn region;

[0027] When the fertilization demand degree is greater than a preset demand threshold, a preset minimum concentration of fertilization in the current growth stage of the corresponding corn region is corrected based on the fertilization demand degree, a corrected fertilization concentration of the target nutrient substance is obtained, and fertilization is controlled.

[0028] Further, when the fertilization is controlled, an unmanned aerial vehicle is used to fertilize the corn region in need of supplemental fertilization.

[0029] Further, the method for matching the corn regions between the adjacent frames of the hyperspectral images comprises:

[0030] The two corn regions with the most overlapping pixels in coordinates between adjacent frames of the hyperspectral image are matched.

[0031] Furthermore, the method for identifying the growth stages of different maize plants includes:

[0032] The corn plant regions are obtained based on NDVI values. The hyperspectral detection results of the corn plant regions are input into a CNN network to identify the growth stages of different corn plant regions.

[0033] The present invention has the following beneficial effects:

[0034] This invention first acquires various nutrient data, hyperspectral images, and growth stages of maize plants in a maize field, and divides the field into maize regions to provide a foundation for subsequent analysis. It then integrates the overall nutrient data and overall NDVI value within each maize region to obtain the growth suitability of each region, achieving multi-dimensional data fusion and accurately assessing the suitability of each region for maize growth. Furthermore, it reveals the impact of drought stress on growth suitability based on the temporal fluctuations of growth suitability within the same maize region. Combining the differences in growth suitability between a maize region and adjacent regions, as well as the area of ​​the maize region, it reflects the range of water content influence, obtaining the real-time growth influence coefficient for each plant pixel, providing a basis for subsequent precise fertilization control. Next, it analyzes the spatial diffusion of drought stress based on the fluctuations of the growth influence coefficient of individual plant pixels and the changes in the center of gravity of the maize region to which it belongs. Combined with the growth delay of maize plants, it accurately represents the cumulative impact of drought stress on maize plants, providing a basis for final precise fertilization control. Finally, it controls fertilization based on the real-time cumulative coefficients of all growth disturbances corresponding to each maize region, combined with the degree of low levels of each nutrient data. This solution integrates hyperspectral images, NDVI values, and nutrient data to dynamically assess the impact and accumulation of drought on plant pixels, enabling precise fertilization by region and stage. This effectively addresses the technical challenges of inaccurate fertilization control caused by the spatial spread of drought stress and growth retardation. Attached Figure Description

[0035] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0036] Figure 1 A flowchart illustrating an IoT-based method for controlling phased fertilization of corn, as provided in one embodiment of the present invention;

[0037] Figure 2A flow chart of a growth influence coefficient acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the following describes in detail the specific implementation, structure, features and effects of a corn staged fertilization control method based on the Internet of Things according to the present application, with reference to the accompanying drawings and preferred embodiments. Different "one embodiment" or "another embodiment" in the following description do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0040] The following describes in detail the specific scheme of the corn staged fertilization control method based on the Internet of Things provided by the present application.

[0041] Please refer to Figure 1 , which shows a flow chart of a corn staged fertilization control method based on the Internet of Things provided by an embodiment of the present application, specifically including:

[0042] Step S1: Obtain various nutrient data, hyperspectral images and NDVI values of the corn field at a preset acquisition frequency, and identify the growth stages of different corn plants.

[0043] In an embodiment of the present application, the base fertilizer is uniformly applied in the corn field and the planting time of each corn plant is consistent. Starting from the time when the base fertilizer is applied, the preset acquisition frequency is 1 time per day; the various nutrient data at least include the concentrations of nitrogen, phosphorus and potassium;

[0044] The soil moisture detection sensor is used to monitor the soil water content at different positions in the corn field (for example, frequency domain reflectometer, etc.), and the measurement is performed once every half hour;

[0045] Uniformly select monitoring points in the corn field, wherein the multi-channel ion meter (connected to multiple ion selective electrodes) is used to monitor the dissolved ions at each monitoring point (in-situ calibration at the monitoring point), so as to reflect the nutrient content in the soil at the corresponding monitoring point (measurement is performed once every half hour);

[0046] For example: The electrode measures the nitrate nitrogen in the soil, reflecting the nitrogen content in the soil. The phosphate electrode reflects the phosphorus content in the soil (for example ) k + The electrode reflects the potassium content in the soil.

[0047] The ion concentration is obtained according to the measured potential value by using the Nernst equation (existing method), and the nutrition data collected at different times in a day are averaged to obtain the nutrition data of the whole day.

[0048] In order to avoid light interference, the unmanned aerial vehicle is used to obtain the hyperspectral image of the field at a fixed time period every day, usually the period of direct sunlight, for example, 12:00-14:00; wherein, the hyperspectral images obtained at adjacent time points have an overlap rate of 30%, the hyperspectral images at different positions are spliced based on a feature matching algorithm, the complete hyperspectral image of the corn field is obtained, and the image is calibrated with the actual scene progress to obtain the position coordinates of each monitoring point in the hyperspectral image.

[0049] It should be noted that the size of the corn field, the corn variety and the setting of the monitoring points (number, distribution, collection frequency, etc.) can be set by the implementer, and are not limited herein.

[0050] Considering that the growth stage of the corn plant has different requirements for different nutrients, and in corn planting, the NDVI value changes with the growth stage of the corn and shows a clear time sequence pattern, it is necessary to extract the NDVI value of each point in the image and identify the growth stage of different corn plants.

[0051] Preferably, in an embodiment of the present application, the NDVI value is used to obtain the corn plant area, and the hyperspectral detection result of the corn plant area is input into the CNN network to identify the growth stage of different corn plant areas.

[0052] As an example, the NDVI threshold is set to 0.4, the binary image distinguishes between vegetation and non-vegetation areas, the open operation (denoising) and the close operation (filling holes) optimize the binary mask, each independent connected domain is marked as a corn plant area, and small areas (noise) are removed, such as removing connected domains with less than 10 pixel points, and the threshold of small areas can be adjusted.

[0053] Further, the average spectral curve of each plant area is extracted and normalized for pretreatment; then the spectral data is input into a pre-trained 1D CNN model, the CNN model includes convolution layers, pooling layers and fully connected layers, the nonlinear mapping relationship between the spectral features and the growth stage is analyzed, and the growth stage classification result of the corn plant in the corn plant area is output.

[0054] It should be noted that the normalization can adopt linear normalization, and the normalization under the corresponding data dimension can be adopted, and the normalization adopted in the embodiments of the present scheme can all adopt the method; the growth stages of corn are divided by using existing V stage and R stage; the working principles of various sensors, unmanned aerial vehicles, hyperspectral cameras, methods for extracting NDVI values, obtaining corn plant regions, and distinguishing growth stages by means of CNN network are all common technical means for those skilled in the art, and will not be described in detail.

[0055] Among them, the fertilization and leaf nutrition are both expressed by element content, and the nutrition state of the leaf is also reflected by means of the NDVI value.

[0056] Step S2: In the pixel points of the corn plant of each hyperspectral image, the corn region is obtained according to the similar features of the pixel points and adjacent pixel points; the growth suitability of each corn region is obtained by fusing the overall nutrition data and the overall NDVI value in each corn region.

[0057] In the same corn field, due to different soil conditions (such as different water content and different terrain) and different consumption of nutrients in the field by different corn plants in the growth process, the degree of soil suitable for the growth of corn plants at different positions is different, so it is necessary to obtain the corn region according to the similar features of the pixel points and adjacent pixel points in the pixel points of the corn plant of each hyperspectral image.

[0058] Among them, in order to exclude the interference of irrelevant pixel points (such as empty land), only the pixel points of the corn plant are analyzed and divided, and the pixel points of the corn plant are referred to as plant pixel points.

[0059] In addition, considering that the nutrient supply capacity of different corn regions is different, and the NDVI value reflects the corn growth and health condition, the growth suitability of each corn region is obtained by fusing the overall nutrition data and the overall NDVI value in each corn region, which represents the degree of the corresponding region in the corn field suitable for the growth of corn, and prepares for subsequent analysis of the influence of water content on the growth of corn and optimization of fertilization control.

[0060] Preferably, in an embodiment of the present application, the region division method of each hyperspectral image is consistent, and one of them is selected for description;

[0061] In any one of the hyperspectral images, the hyperspectral curve of each pixel point is converted into a hyperspectral vector;

[0062] As an example, for each pixel point of the corn plant region in the figure, the reflectivity of a single pixel point at all wavebands is arranged in ascending order of waveband to obtain the hyperspectral vector of the pixel point.

[0063] The higher the cosine similarity between vectors, the higher the similarity. Therefore, based on region growing, the cosine similarity between the pixel points and the adjacent pixel points in the preset neighborhood is less than or equal to the preset similarity threshold, and the region is divided.

[0064] As an example, the preset neighborhood is an eight-neighborhood, the preset similarity threshold is 0.7, the seed point is set by using a fixed grid method, for example, the horizontal axis resolution is 1920, the vertical axis resolution is 1080, the grid parameter is 1 / 40, the horizontal axis is divided into 48 parts, the vertical axis is divided into 27 parts, the horizontal and vertical division lines are distributed in a grid, and the pixel points of the plants at the intersection of the division lines are taken as the seed points.

[0065] An independent growth region is created for each seed point. Starting from each seed point, the plant pixel points in the eight-neighborhood are checked. When the cosine similarity is greater than 0.7, the pixel is incorporated into the current growth region, and the expansion continues. If the similarity is less than or equal to 0.7, the growth in that direction stops. When all directions cannot continue to grow, the current corn region division is completed.

[0066] The cosine similarity quantifies the similarity between the hyperspectral vectors, indicating the similarity characteristics of the pixel points and the adjacent pixel points. The cosine similarity is a prior art and will not be described again.

[0067] Preferably, in an embodiment of the present application, there are multiple monitoring points in a corn region, and there are multiple data of the same nutrient. The average value represents the nutrient data of the entire corn region. Similarly, the average value of the NDVI of all pixel points is obtained.

[0068] Considering that PCA is a common dimension reduction and feature extraction method, the average values of the various nutrient data and the NDVI may be related to each other, and both reflect the growth of the corn crop to a certain extent. PC1 represents the direction that explains the maximum degree of data variation, usually containing the most important information. The loading vector can be understood as the contribution of each index to the principal component, and the weight distribution is more objective.

[0069] Based on this, at any acquisition time corresponding to a hyperspectral image, the average values of the various nutrient data of all monitoring points in each corn region and the average value of the NDVI of all pixel points are taken as the PCA input. The loading vector of PC1 is taken as the weighted weight of each index and is weighted and summed to obtain the growth suitability of each corn region at the corresponding acquisition time, realizing efficient fusion of multi-dimensional data and accurately evaluating the suitability of corn growth at the acquisition time corresponding to the hyperspectral image to which each corn region belongs.

[0070] It should be noted that each item of nutritional data and NDVI value is normalized in the respective data dimension before inputting PCA; the PCA algorithm is known to those skilled in the art and will not be described here.

[0071] Step S3: matching the corn regions between adjacent hyperspectral images; obtaining the real-time growth influence coefficient of each plant pixel point according to the time series fluctuation of the growth suitability of the same corn region, in combination with the difference in growth suitability between the corn region and the adjacent corn region, and the area of the corn region.

[0072] When the soil moisture content fluctuates (especially in dry soil), different regions in the soil may have different water contents, which reduces the dissolution and migration ability of the nutrient elements in the base fertilizer, and is not conducive to the absorption of the roots of the corn crop. Therefore, the growth range of the nutrient substances in the region with a larger degree of water loss is smaller than that in the normal region after applying the base fertilizer, and the nutrient absorption of the corn plant may be continuously stressed, resulting in a continuous low effective nutrient in the region, and slow growth of the plant in the corresponding region.

[0073] Unlike the growth stage of individual plants caused by genetic defects or pests and diseases, low soil moisture content (drought) inhibits the growth of plants in a larger range rather than a small random difference, and the influence range is relatively stable in space. Because the growth of corn plants is affected by multiple factors and changes over time, the division results of the corn regions in the hyperspectral images will also change, so it is necessary to match the corn regions between adjacent hyperspectral images to facilitate the analysis of the long-term impact on the plants.

[0074] Preferably, in an embodiment of the present application, the two corn regions with the most coordinate coincident pixel points between adjacent hyperspectral images are matched.

[0075] It should be noted that the monitoring device or some markers (such as landmarks) in the field remain unchanged, which facilitates the coordinate alignment of different hyperspectral images.

[0076] Considering that drought stress affects the growth suitability of the land, which is reflected in the time series fluctuation of the growth suitability, and the difference in growth suitability between the corn region and the adjacent corn region reflects the similar distribution of the growth suitability in space, which indirectly reflects the range of drought impact; and the area of the corn region reflects the range of drought impact.

[0077] Therefore, according to the time series fluctuation of the growth suitability of the same corn region, in combination with the difference in growth suitability between the corn region and the adjacent corn region, and the area of the corn region, the real-time growth influence coefficient of each plant pixel point is obtained, which represents the current drought stress impact degree of each plant pixel point, and provides a basis for subsequent precise control of fertilization.

[0078] wherein the plant pixel points are pixel points of corn plants.

[0079] Preferably, in one embodiment of the present application, referring to Figure 2 which shows a flow chart of a growth influence coefficient acquisition method provided by one embodiment of the present application, and specifically comprises:

[0080] Step S301: selecting any corn area as a target area, and dividing a fluctuation period between the maximum value of the growth suitability degree and the first acquisition time point based on the maximum value of the growth suitability degree.

[0081] Considering that under the influence of drought stress, the growth suitability degree of corn plants will have a small amplitude and short time growth after fertilization (dissolution of base fertilizer after rainfall or irrigation), and then remain relatively stable and decrease, the time domain is first divided based on the maximum value point and the first acquisition time point to obtain the fluctuation period.

[0082] Step S302: acquiring an effective growth coefficient of each fluctuation period according to the duration of the target area in each fluctuation period, the growth suitability degree at the starting time, and the maximum growth suitability degree; and acquiring a first influence factor according to the number and mean absolute value of negative difference values in a first-order difference sequence of the effective growth coefficient, and combining the first effective growth coefficient in time sequence, at the current time.

[0083] For a single fluctuation period, the longer the duration and the higher the maximum growth suitability degree, the better the effective growth of corn plants, and the smaller the starting time of the fluctuation period, the shorter the influence time of drought stress on corn plants, and the more conducive to corn growth, so the effective growth coefficient of each fluctuation period is acquired in combination with these parameters.

[0084] As an example, for any fluctuation period, the starting time of the period is taken as the denominator, the product of the duration of the period and the maximum growth suitability degree in the period is taken as the numerator, and the ratio of the fraction is taken as the corresponding effective growth coefficient, which shows the time sequence fluctuation characteristics of the growth suitability degree; wherein the first acquisition time point starts from 1 to avoid the denominator being zero.

[0085] Considering that the larger the number and mean absolute value of negative difference values in the first-order difference sequence of the effective growth coefficient, the stronger the decreasing trend of the effective growth coefficient with time, and the greater the influence degree of drought stress on the target area; the smaller the first effective growth coefficient, the lower the degree of initial suitability for corn growth, and the greater the influence of (low) soil water content, so the first influence factor is acquired in this way.

[0086] The product of the number of negative difference values and the mean absolute value in the first-order difference sequence of the valid growth coefficients up to the current time is taken as the numerator, the first valid growth coefficient in time sequence is taken as the denominator, and the ratio of the two is taken as the first influence factor of the target region in real time.

[0087] Step S303: According to the difference between the growth suitability of the target region and its adjacent corn region at each collection time and the area of the target region, a second influence factor is obtained.

[0088] It is considered that the larger the area of the target region up to the current time, the larger the area affected by the low soil moisture content. Meanwhile, the smaller the difference between the growth suitability of the target region and its adjacent corn region at each collection time, the more similar the influence of the low soil moisture content, which reflects the larger the range of drought influence.

[0089] Based on this, at each collection time, a growth difference coefficient of each collection time is obtained according to the difference between the growth suitability of the target region and its adjacent corn region. The mean of the growth difference coefficients of the target region at all collection times and the mean of the area are fused to obtain the second influence factor.

[0090] As an example, when the target region and other corn regions have adjacent edge pixel points, it is determined that the two corn regions are adjacent. At each collection time, the mean of the absolute value of the difference between the growth suitability of the target region and its adjacent corn region is taken as the growth difference coefficient of each collection time, which shows the difference between the growth suitability of the corn region and the adjacent corn region through the mean of the absolute value of the difference.

[0091] For the target region, the mean of the growth difference coefficients of the collection times up to the current time is taken as the denominator, the mean of the area is taken as the numerator, and the ratio of the two is taken as the second influence factor of the target region in real time.

[0092] The area of the region is the number of pixel points in the region.

[0093] Step S304: The first influence factor and the second influence factor are fused to obtain the growth influence coefficient of the target region in real time, which is taken as the growth influence coefficient of each pixel point in the target region in real time.

[0094] It is considered that the first influence factor reflects the degree of drought influence from the growth suitability, and the second influence factor reflects the degree of drought influence from the influence area. Therefore, the two are finally fused to obtain the influence coefficient of the target region in real time, which is directly taken as the growth influence coefficient of each pixel point in the target region in real time.

[0095] As an example, the product of the first influence factor and the second influence factor is taken as the growth influence coefficient.

[0096] In other embodiments of the present application, the implementer can also fuse the first influence factor and the second influence factor in a positive correlation manner such as addition or weighted summation, and the like, and will not be described again.

[0097] Step S4: According to the fluctuation of the growth influence coefficient of the single plant pixel point and the change of the gravity center of the corn region to which the single plant pixel point belongs, as well as the growth delay degree and the growth stage of the corn plant to which the single plant pixel point belongs, the real-time growth disturbance accumulation coefficient of the corresponding plant pixel point is obtained; according to all the growth disturbance accumulation coefficients of each corn region in real time, combined with the low degree of each kind of nutrient data, the fertilization is controlled.

[0098] In the same corn field, the degree of disturbance of nutrient absorption may be migrated to a certain extent due to the following factors: the difference in terrain between different regions in the corn field causes water to migrate from high terrain to low terrain, causing the accumulation of nutrient disturbance in the high terrain region to intensify; the scouring of the soil by climatic factors such as rainfall also has a certain influence on the degree of disturbance of nutrient absorption in different regions.

[0099] Because corn has a standard growth time at different growth stages, for example, VE corresponds to 0-5 days after emergence, V1 corresponds to 5-7 days after emergence, the generation delay degree of the corn plant can be analyzed, and the accumulation stress degree of the low soil water content is shown, so according to the fluctuation of the growth influence coefficient of the single plant pixel point and the change of the gravity center of the corn region to which the single plant pixel point belongs, as well as the growth delay degree and the growth stage of the corn plant to which the single plant pixel point belongs, the real-time growth disturbance accumulation coefficient of the corresponding plant pixel point is obtained, the cumulative influence degree of drought stress on the corn plant is accurately represented, and the basis for the final accurate control of fertilization is provided.

[0100] Preferably, in an embodiment of the present application, any plant pixel point is selected as a target pixel point for analysis of the plant pixel points one by one.

[0101] Considering that the first-order difference reflects the change speed, the greater the mean value of the first-order difference value, the greater the degree of continuous aggravation of the low water content, so the mean value of the first-order difference value of the growth influence coefficient of the target pixel point is obtained as the influence growth factor, which represents the fluctuation characteristics of the growth influence coefficient.

[0102] Considering that the target pixel point may have been divided into different corn regions, all the corn regions to which the target pixel point has belonged until the current time are obtained as the regions to be analyzed. Considering that changes in water content will cause changes in growth conditions, and spatial migration of soil moisture will cause spatial migration of corn regions, a displacement factor is obtained according to the distance between the gravity center of the corn region to which the target pixel point belongs at the first collection time and the gravity center of all the regions to be analyzed, which represents the change of the gravity center of the corn region.

[0103] As an example, the average of the distance between the center of gravity of the region to be analyzed in the current hyperspectral image and the center of gravity of the corn region to which the target pixel point belongs in the first collected hyperspectral image is taken as the real-time displacement factor of the target pixel point.

[0104] When the target pixel point does not belong to any corn region in the first collected hyperspectral image, the coordinate point of the target pixel point itself is taken as the center of gravity point.

[0105] Considering that there is a standard growth duration for the complete growth stage of a corn plant, the longer the actual growth duration, the greater the length deviation, the higher the growth delay, and the greater the cumulative impact of growth interference, the complete growth stage can be obtained from the current growth stage.

[0106] Therefore, according to the duration of all complete growth stages of the corn plant corresponding to the target pixel point, the length deviation relative to the corresponding preset standard growth duration, and the influence growth factor and displacement factor, the real-time growth interference accumulation coefficient of the target pixel point is obtained.

[0107] As an example, the average of the ratio of the duration of the complete growth stage of the corn plant corresponding to the target pixel point to the corresponding preset standard growth duration is taken as the delay degree coefficient, the product of the influence growth factor and the displacement factor is taken as the numerator, the sum of the delay degree coefficient and the preset positive zero parameter 1 is taken as the denominator, the fractional ratio is taken as the water influence factor, and the water influence factor is taken as the independent variable. After normalization by the sigmoid function, the real-time growth interference accumulation coefficient of the target pixel point is obtained.

[0108] Among them, the delay degree coefficient shows the growth delay of the corn plant based on the preset standard growth duration.

[0109] It should be noted that when the corn plant has not experienced a complete growth stage, the delay degree coefficient is 0, and the standard growth duration of each growth stage of the corn plant is affected by the selected corn variety of the implementer, for example, Zhengdan 958 has a shorter V stage, and Xianyu 335 has a longer R stage. The specific value can be obtained by referring to the experimental production data or production manual of the selected corn variety, which will not be described here.

[0110] In another embodiment of the application, it is also considered that the more growth stages the plant has experienced so far, the longer the duration of the drought impact, and the greater the cumulative impact of growth interference. Therefore, the number of growth stages experienced so far is also multiplied by the water influence factor to obtain the real-time growth interference accumulation coefficient of the target pixel point.

[0111] When it is considered that the drought accumulation stress degree of some corn regions in the current corn field is high and the measured nutrient data shows that the content of the nutrient element is low, it is indicated that the corresponding region needs to be supplemented with fertilization in time, so the fertilization is controlled according to all the growth interference accumulation coefficients of each corn region in real time, combined with the low degree of each nutrient data, to ensure the growth of corn plants.

[0112] Preferably, in an embodiment of the present application: the greater the growth interference accumulation coefficient of all pixel points in the current corn region, the greater the real-time accumulation degree of the corn region, so in the current hyperspectral image, the real-time accumulation coefficient of each corn region is obtained by fusing the growth interference accumulation coefficients of all pixel points in each corn region.

[0113] As an example, in the current hyperspectral image, the mean value of the growth interference accumulation coefficients of all pixel points in each corn region is taken as the corresponding real-time accumulation coefficient.

[0114] The nutrient substances are selected as target nutrient substances by category, which facilitates the analysis of each nutrient substance in the current, and in the present example, the nutrient substances only include nitrogen, phosphorus and potassium;

[0115] It is considered that the greater the difference between the preset minimum concentration and the measured concentration required by the corn region, the greater the low degree of the measured concentration; at the same time, the greater the real-time accumulation coefficient, the greater the drought influence accumulation degree, so the fertilization requirement degree is obtained according to the difference between the preset minimum concentration and the measured concentration required by the plant for the target nutrient substance in any corn region, combined with the real-time accumulation coefficient.

[0116] As an example, the product of the difference between the preset minimum concentration and the measured concentration and the real-time accumulation coefficient is taken as the independent variable, which is mapped through the premnmx(x) function, x is the independent variable, and the mapping result is taken as the fertilization requirement degree.

[0117] When the fertilization requirement degree is greater than the preset requirement threshold, the preset minimum concentration of fertilization of the corresponding corn region in the current growth stage is corrected based on the fertilization requirement degree, the corrected fertilization concentration of the target nutrient substance is obtained and the fertilization is controlled.

[0118] As an example, the preset requirement threshold is 0.5, the sum of the fertilization requirement degree and the constant 1 is taken as the correction coefficient, the product of the preset minimum concentration of fertilization of the current corn region and the correction coefficient is taken as the corrected fertilization concentration, and the unmanned aerial vehicle is used to fertilize the corn region that needs to be supplemented with fertilization.

[0119] It should be noted that when there are corn plants in different growth stages in the corn area, the preset minimum concentration required by the corn area for the target nutrient is obtained by weighting and summing the preset minimum concentrations required by each growth stage of the corn plant according to the proportion of each growth stage of the corn plant as a weighting weight; similarly, the preset minimum concentration of fertilization is weighted and summed according to the proportion of each growth stage of the corn plant as a weighting weight, to obtain the preset minimum concentration of the target nutrient applied in the corresponding corn area;

[0120] The required nutrient concentration and fertilization concentration range of the corn plant in different growth stages are related to the corn variety and the planting area, and can also be obtained by consulting the experimental production data or production manual of corn, so as to obtain the preset minimum concentration and the preset maximum concentration of the fertilization concentration range, and the corrected fertilization concentration should not exceed the preset maximum concentration, and when it exceeds, only the preset maximum concentration is taken to avoid seedling burning;

[0121] The implementer can also change the calculation method, for example, multiplying the range of the target nutrient fertilization concentration in the current growth stage by the fertilization requirement degree, and taking the sum of the product and the preset minimum concentration as the corrected fertilization concentration.

[0122] The analysis process of each corn area is the same, and only one example is described here.

[0123] In another embodiment of the present application, it is also considered that part of the growth stage may not have a fertilization plan in the pre-fertilization plan, that is, the preset minimum concentration of fertilization may be 0, and at this time, the higher the fertilization requirement degree, the stronger the demand of the corn area for fertilization, and at this time, a preset high demand threshold of 0.75 is set, the difference between the real-time fertilization requirement degree and 0.75 is taken as a supplementary correction coefficient, and the product of the supplementary correction coefficient and the corresponding fertilization concentration at the last fertilization is taken as a supplementary fertilization concentration, so as to perform supplementary fertilization to relieve the transition to the next fertilization; the minimum time interval of adjacent two supplementary fertilizations can also be set, such as 15 days, to avoid seedling burning.

[0124] To sum up, in view of the technical problem of inaccurate fertilization control for adversity agricultural soil, the application provides a corn staged fertilization control method based on the Internet of Things. The application first acquires various nutrient data, hyperspectral images, growth stages of corn plants and corn regions; further fuses overall nutrient data and overall NDVI values in each corn region to acquire growth suitability; further acquires real-time growth influence coefficients of each plant pixel point according to time sequence fluctuations of the growth suitability of the same corn region, in combination with differences between the growth suitability of the corn region and adjacent corn regions and the area of the corn region; further analyzes fluctuations of the growth influence coefficients and spatial offsets of the corn region, in combination with growth retardation of the corn plants to acquire real-time growth interference accumulation coefficients; and finally realizes precise control of fertilization in a staged and regional manner according to all growth interference accumulation coefficients corresponding to each corn region in real time, in combination with the low degree of each kind of nutrient data.

[0125] It should be noted that the above-mentioned embodiment sequence of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. 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, multi-task processing and parallel processing are also possible or can be advantageous.

[0126] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A corn split application control method based on the Internet of Things, characterized in that, The method comprises: Obtaining each nutrient data, hyperspectral image and NDVI value of the corn field at a preset acquisition frequency, and identifying the growth stage of different corn plants; In the pixel points of the corn plants in each of the hyperspectral images, a corn region is obtained according to the similarity characteristics of the pixel points and adjacent pixel points; the overall nutrient data and the overall NDVI value in each corn region are fused to obtain the growth suitability of each corn region; The corn regions between adjacent frames of the hyperspectral images are matched; according to the time sequence fluctuation of the growth suitability of the same corn region, in combination with the difference between the growth suitability of the corn region and adjacent corn regions, and the area of the corn region, the real-time growth influence coefficient of each plant pixel point is obtained; According to the fluctuation of the growth influence coefficient of the single plant pixel point, the change of the gravity center of the corn region to which the single plant pixel point belongs, and the growth retardation degree and the growth stage of the corn plant to which the single plant pixel point belongs, the real-time growth disturbance accumulation coefficient of the corresponding plant pixel point is obtained; and according to all the growth disturbance accumulation coefficients corresponding to each of the corn regions in real time, in combination with the low degree of each nutrient data, the fertilization is controlled; The method for obtaining the growth influence coefficient comprises: Any of the corn regions is selected as a target region, and a fluctuation period is divided between the maximum values of the growth suitability and between the maximum values and the first acquisition time; According to the duration of each of the fluctuation periods, the growth suitability and the maximum growth suitability of the target region at the starting time, the effective growth coefficient of each of the fluctuation periods is obtained; at the current time, according to the number and the mean absolute value of the negative difference values in the first-order difference sequence of the effective growth coefficient, in combination with the first effective growth coefficient in time sequence, the first influence factor is obtained; According to the difference between the growth suitability of the target region and adjacent corn regions at all acquisition times and the area of the target region, the second influence factor is obtained; The real-time growth influence coefficient of the target region is obtained by fusing the current first influence factor and the second influence factor, as the real-time growth influence coefficient of each pixel point in the target region.

2. The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the corn region comprises: In any of the hyperspectral images, the hyperspectral curve of each pixel point is converted into a hyperspectral vector; based on region growing, in all the pixel points of the corn plants, the cosine similarity of the hyperspectral vectors between a pixel point and adjacent pixel points in a preset neighborhood of the pixel point is less than or equal to a preset similarity threshold as a growth termination condition, and the corn field is regionally divided. 3.The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the growth suitability comprises: At the acquisition time corresponding to any of the hyperspectral images, the mean value of each nutrient data of all monitoring points in each of the corn regions and the mean value of the NDVI of all pixel points are taken as the input of the PCA algorithm, the load vector of PC1 is taken as the weighted weight of each index and weighted summation is performed to obtain the growth suitability of each of the corn regions at the corresponding acquisition time.

4. The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the second influence factor comprises: At each acquisition time, a growth difference coefficient of each acquisition time is obtained according to a difference between the growth suitability of the target region and the corn region adjacent to the target region; and a second influence factor is obtained by fusing a mean value of the growth difference coefficients of the target region of all acquisition times and a mean value of the area of the target region.

5. The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, The method for obtaining the growth interference accumulation coefficient comprises: selecting any plant pixel point as a target pixel point; obtaining a mean value of a first-order difference value of the growth influence coefficient of the target pixel point as an influence growth factor; obtaining all the corn regions to which the target pixel point once belongs as to-be-analyzed regions, and obtaining a displacement factor according to a distance between the barycenter of all the to-be-analyzed regions and the corn region to which the target pixel point belongs at the first acquisition time; obtaining a real-time growth interference accumulation coefficient of the target pixel point according to a lengthening degree of all complete growth stages of the corn plant corresponding to the target pixel point relative to a corresponding preset standard growth length, in combination with the influence growth factor and the displacement factor.

6. The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, The method for controlling fertilization comprises: in the current hyperspectral image, fusing the growth interference accumulation coefficients of all pixel points in each corn region to obtain a real-time accumulation coefficient of each corn region; selecting a nutrient substance as a target nutrient substance; obtaining a fertilization requirement degree in combination with the real-time accumulation coefficient according to a difference between a preset minimum concentration and a measured concentration of the target nutrient substance required by the plant in any corn region; when the fertilization requirement degree is greater than a preset requirement threshold, correcting a preset minimum concentration of fertilization in the current growth stage of the corresponding corn region based on the fertilization requirement degree to obtain a corrected fertilization concentration of the target nutrient substance and control fertilization.

7. The corn staged fertilization control method based on the Internet of Things according to claim 1, characterized in that, When the fertilization is controlled, an unmanned aerial vehicle is used to fertilize the corn region in need of supplemental fertilization. 8.The corn staged fertilization control method based on the Internet of Things according to claim 1, wherein, The method for matching the corn regions between the adjacent frames of the hyperspectral images comprises: the two corn regions with the most coordinate coincident pixel points between the adjacent frames of the hyperspectral images are matched. 9.The corn staged fertilization control method based on the Internet of Things according to claim 1, wherein, The method for identifying the growth stages of different corn plants comprises: a corn plant region is obtained based on an NDVI value, a hyperspectral detection result of the corn plant region is input into a CNN network, and the growth stages of different corn plant regions are identified.

Images