A method and apparatus for determining crop growth status

Abstract

This application provides a method and apparatus for determining crop growth status. The aim is to acquire remote sensing images of key phenological periods and statistically analyze the distribution characteristics of crop growth to construct adaptive growth level intervals, thereby achieving timely and accurate judgment of crop growth status. The method for determining crop growth status includes: acquiring crop growth results during key phenological periods; calculating a skewness coefficient; determining the distribution state of the growth results based on the skewness coefficient; determining a confidence level based on the total number of pixels and spatial resolution of the remote sensing image; selecting relevant feature parameters according to the distribution state and confidence level of the growth results; constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with a preset number of growth levels and the selected relevant feature parameters; and matching each pixel value in the growth results with these intervals to obtain the crop growth level distribution results.

Description

Technical Field

[0001] This application relates to the field of agricultural remote sensing applications, and in particular to a method and apparatus for determining the growth status of crops. Background Technology

[0002] In the field of agricultural production, judging the growth status of crops is crucial for scientifically guiding agricultural production and improving yield and quality.

[0003] In existing technologies, the assessment of crop growth is based on data analysis after the harvest of grain products. This method of assessing crop growth based on grain products has a significant lag and cannot reflect changes in crop growth during the growth process, thus missing the best time for management and control.

[0004] Therefore, there is an urgent need for a new method to determine the growth status of crops, so as to accurately and timely assess the growth status of crops and provide strong support for agricultural production. Summary of the Invention

[0005] This application provides a method and apparatus for determining crop growth status. The purpose is to construct an adaptive growth level range by acquiring remote sensing images of key phenological periods and statistically analyzing the distribution characteristics of growth results, so as to achieve timely and accurate judgment of crop growth status.

[0006] To address the aforementioned technical problems, this application provides the following technical solutions: The first aspect of this application provides a method for determining the growth status of crops, including: Remote sensing images of crops in the target area during key phenological periods are acquired, and growth results are obtained based on the remote sensing images. The growth results are characterized by the pixel values ​​of the crops. Calculate the skewness coefficient of the growth results, and determine the distribution state of the growth results based on the skewness coefficient; The confidence level is determined based on the total number of pixels and spatial resolution of remote sensing images; Select relevant characteristic parameters based on the distribution and confidence level of the growth results; Based on the boundary values ​​of the confidence interval corresponding to the confidence level, and combined with the preset number of growth levels and the selected relevant feature parameters, crop growth intervals for different growth levels are constructed. Each pixel value in the growth results is matched with the crop growth interval to obtain the crop growth level distribution results for the target area.

[0007] A second aspect of this application provides an apparatus for determining the growth status of crops, comprising: The acquisition unit is used to acquire remote sensing images of crops in the target area during key phenological periods and obtain growth results based on the remote sensing images. The growth results are characterized by the pixel values ​​of the crops. A determining unit is used to calculate the skewness coefficient of the growth results in the acquiring unit, and to determine the distribution state of the growth results based on the skewness coefficient; The determining unit is used to determine the confidence level based on the total number of pixels and spatial resolution of the remote sensing image; The determining unit is used to select relevant characteristic parameters based on the distribution status and confidence level of the growth results; The construction unit is used to construct crop growth intervals for different growth levels based on the boundary values ​​of the confidence interval corresponding to the confidence level, combined with the preset number of growth levels and the relevant feature parameters selected in the unit. The result acquisition unit is used to match each pixel value in the growth results with the construction unit, which is used to match the crop growth interval in the target area, so as to obtain the crop growth level distribution results in the target area.

[0008] A third aspect of this application provides a storage medium including a stored program that, when the program is executed, controls the device where the storage medium is located to perform the aforementioned method for determining crop growth.

[0009] A fourth aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for determining crop growth conditions as described above.

[0010] Compared to existing technologies, this application provides a method for determining crop growth status. This technical solution enables scientific and adaptive classification of crop growth levels, significantly improving monitoring accuracy and regional applicability while reducing subjective interference and data costs. By calculating the skewness coefficient of growth results, it automatically identifies whether the data follows a normal or skewed distribution, overcoming the limitation of traditional methods that only assume a normal distribution, and making the classification more closely aligned with actual data characteristics. The confidence level is adaptively determined based on the total number of pixels and spatial resolution. The level interval is constructed by combining the distribution status with characteristic parameters such as mean, standard deviation, median, and confidence interval boundaries, without relying on empirical thresholds or fixed standards, greatly reducing subjective human influence and improving the adaptability of monitoring in different regions and at different critical periods. This solution only requires data from a single critical period to complete the rating, reducing data requirements and processing workload. It also supports setting the number of levels as needed to meet diverse monitoring scenarios. The overall method is simple, universally applicable, and can quickly output growth level distribution results, providing accurate and objective technical support for field management, yield prediction, disaster assessment, and agricultural insurance. Attached Figure Description

[0011] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of this application are illustrated by way of example and not limitation, with the same or corresponding reference numerals denoteing the same or corresponding parts, wherein: Figure 1 A flowchart illustrating a method for determining crop growth is shown schematically. Figure 2 A flowchart illustrating another method for determining crop growth is shown schematically. Figure 3 This diagram schematically illustrates a corn growth level display. Figure 4 A schematic diagram of a device for determining the growth status of crops is shown. Figure 5 A schematic diagram of another device for determining the growth status of crops is shown. Detailed Implementation

[0012] Exemplary embodiments of this application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0013] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0014] In agricultural production, assessing crop growth is crucial for scientifically guiding agricultural production and improving yield and quality. Current technologies rely on post-harvest data analysis of grain products to determine crop growth. This method, based on grain products, suffers from significant lag, failing to reflect changes in crop growth during the growing process and thus missing optimal management and control opportunities.

[0015] Based on this, the applicant of this application provides a method for determining the growth status of crops, as follows: Step 101: Obtain remote sensing images of crops in the target area during key phenological periods, and obtain growth results based on the remote sensing images.

[0016] In this embodiment, the crop growth results are characterized by pixel values. Key phenological periods are critical stages in crop growth that significantly impact growth (such as the jointing stage of wheat and the tasseling stage of maize), determined by both crop growth patterns and climatic characteristics. Remote sensing imagery, acquired via satellites (such as Sentinel and Landsat) or drones, contains spectral information about crop growth status and is the core data source for growth analysis. The growth results are generated based on vegetation indices extracted from remote sensing imagery, quantifying crop growth data using pixel values. A pixel value is the numerical value corresponding to the smallest unit pixel in a remote sensing image, reflecting growth-related information such as vegetation cover and growth vigor within the area covered by that pixel. Specifically, the remote sensing imagery for key phenological periods can be a specific time period within the key phenological period of the target year and a specific time period within the key phenological period of a comparison year. For example, in Northeast China, taking maize as an example, the tasseling stage is generally from mid-August to early September; remote sensing images of August 30th of the target year and August 30th of the comparison year (during the same period) can be selected. It can also be remote sensing imagery of a specific time during a key phenological period of a particular year. The specific selection can be determined based on the actual situation and is not limited here.

[0017] Specifically, in this step, firstly, by combining the growth patterns of crops in the target area (such as the growth cycle of corn from sowing to maturity) with local climate characteristic data (temperature, precipitation, sunshine duration, etc.), and through agricultural surveys and agricultural meteorological observation records, key phenological periods that significantly affect growth (such as the vigorous vegetative growth period and the critical reproductive growth period) are accurately selected. Remote sensing images with no cloud cover and meeting image quality standards are acquired during the key phenological periods, ensuring that the image time matching error does not exceed a preset number of days and that the target area is fully covered; if the target area requires multi-image stitching coverage, image stitching preprocessing must be completed first. Next, radiometric calibration (converting sensor digital quantization values ​​into surface reflectance), atmospheric correction (eliminating atmospheric scattering and absorption interference with the spectrum), and geometric correction (correcting image distortion caused by terrain and sensor attitude, ensuring that pixels accurately correspond to geographic coordinates) are sequentially performed on the acquired remote sensing images to improve data accuracy. Finally, using the year-on-year index method (calculating the difference between the vegetation index of the target year and the comparison year) or the anomaly index method (calculating the degree to which the vegetation index of the target year deviates from the average of the same period over many years), core vegetation indices such as the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) are calculated from the preprocessed images. Based on these indices, the growth-related parameters of each pixel are calculated using year-on-year and anomaly methods, forming a growth result that characterizes crop growth with pixel values.

[0018] Step 102: Calculate the skewness coefficient of the growth results and determine the distribution state of the growth results based on the skewness coefficient.

[0019] In this step, the skewness coefficient is a statistic used to measure the symmetry of the data distribution and to determine whether the dataset deviates from a normal distribution.

[0020]

[0021] Where sk represents the skewness coefficient. x i Let σ be the pixel value of the crop growth result, μ be the mean of the crop growth result, n be the number of pixels in the crop growth result, and σ be the standard deviation of the crop growth result. When sk = 0, the crop growth result follows a normal distribution; when sk ≠ 0, the crop growth result follows a skewed distribution. The distribution state refers to the numerical distribution characteristics of the growth result, which is mainly divided into two categories: normal distribution (symmetric data distribution, skewness coefficient is 0) and skewed distribution (asymmetric data distribution, skewness coefficient is not 0).

[0022] Step 103: Determine the confidence level of the growth results based on the total number of pixels and spatial resolution of the remote sensing image.

[0023] In this step, the total number of pixels is the total number of pixels in the remote sensing image of the target year. It is obtained by multiplying the number of rows and columns of the remote sensing image (i.e., total number of pixels = number of rows × number of columns). It is a core indicator reflecting the amount of data covering the image coverage area and must meet the preset condition of being greater than 10,000 to ensure the effectiveness of the analysis. The confidence level is a probabilistic indicator (such as 95% or 98%) that characterizes the reliability of the crop growth results. It is used to determine the confidence interval of the crop growth results. The higher the confidence level, the stronger the statistical reliability of the results. Its value is determined by both the total number of pixels and the spatial resolution.

[0024] Specifically, first, obtain the basic parameters of the remote sensing image for the target year: extract the number of rows and columns of the image from the image metadata, and determine the spatial resolution (r) of the image annotations. Second, n= Rows × Columns And n is greater than 10,000 n is the size of the remote sensing image used to assess the crop growth in the target area. Rows The number of rows in the remote sensing imagery used to assess crop growth. Columns The number of columns of remote sensing images used to assess crop growth; When n×r 2 ≥100,000 m 2 When determining the first confidence level, it indicates that the data volume is sufficient and the coverage is wide. The range of the first confidence level is P1≥98%, and P1 is the first confidence level. When 1 m 2 <n×r 2 <100,000 m2 At this point, a second confidence level is determined. This indicates that the data volume is relatively limited, and the range of the second confidence level is 95% ≤ P2 < 98%, with P2 being the second confidence level. This balances data reliability with analytical applicability.

[0025] Step 104: Select relevant characteristic parameters based on the distribution status and confidence level of the growth results.

[0026] In this step, the relevant feature parameters are the core statistical indicators used to construct crop growth intervals. These parameters vary depending on the distribution of the dataset. For a normal distribution, they include the standard score, mean, and standard deviation; for a skewed distribution, they include the confidence interval boundary values ​​and the median. The standard score is a statistic in a normal distribution that measures the degree to which data deviates from the mean. It is determined by consulting the normal distribution table based on the confidence level and reflects the data's relative position within the distribution. At a given confidence level, it includes the possible range of population parameters. In a normal distribution, this is calculated from the mean, standard deviation, and standard score; in a skewed distribution, the confidence interval is obtained using the Bootstrap method. The Bootstrap method is a resampling statistical method that determines the confidence interval for skewed data by repeatedly drawing samples and calculating statistics.

[0027] Specifically, first, the distribution of the growth results (normal or skewed distribution) is determined, which is obtained from the skewness coefficient in step 102. If it is a normal distribution: first, based on the confidence level determined in step 103, the corresponding standard score is obtained from the normal distribution table; then, the mean of the growth results is calculated using the arithmetic mean method, and the standard deviation of the dataset is calculated using the standard deviation formula; finally, the standard score, mean, and standard deviation are determined as relevant feature parameters. If it is a skewed distribution: based on the confidence level in step 103, the Bootstrap method is used to repeatedly sample (e.g., 1500 times), and after statistical analysis and sorting of the samples, the corresponding quantiles are used to determine the lower boundary value 'a' and the upper boundary value 'b' of the confidence interval; then, the median is obtained by taking the median value from the sorted dataset; finally, the boundary values ​​'a', 'b', and the median are determined as relevant feature parameters, providing data support for the subsequent construction of growth intervals.

[0028] Step 105: Based on the boundary values ​​of the confidence interval corresponding to the confidence level, and combining the preset number of growth levels with the selected relevant feature parameters, construct crop growth intervals for different growth levels.

[0029] In this step, the preset number of growth levels is set according to the needs of agricultural production monitoring. It can be flexibly selected as either odd or even, with levels from 1 to G corresponding to growth from best to worst, or vice versa. Crop growth intervals are numerical ranges divided based on feature parameters; each interval corresponds to a growth level, used to quantitatively determine the growth level of a single pixel. The confidence interval boundary values ​​are […]. -z×σ, The upper and lower limits of [+z×σ], under skewed distribution, are [a, b] obtained by the Bootstrap method, which are the basic ranges for constructing growth intervals. Based on the determined basic range, crop growth intervals for different growth levels are constructed. It is worth noting that when the distribution of growth results is normal, the confidence interval [...] is used... -z×σ, +z×σ] is the core range, combined with the preset number of growth levels G and relevant characteristic parameters (mean). The standard deviation (σ) and standard score (z) are used to construct a unified partitioning logic based on the parity of the number of grades.

[0030] When the distribution is normal and the number of preset growth levels is odd, the crop growth intervals for different growth levels are as follows: The crop growth range for the i=1th growth level is: ; The crop growth range for the growth level 1 < i < m is: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. The value of the j-th pixel representing the regional crop growth status. Let n be the mean of the crop growth results, and n be the number of pixels in the crop growth results. The standard deviation of the crop growth results is represented by z, where z is the standard score and m is the maximum value of the preset growth level number, and m is an odd number.

[0031] When the distribution is normally distributed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of growth levels of m; select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; replace the any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0032] When the distribution of crop growth results is skewed, based on the confidence interval [a, b] obtained by the Bootstrap method, and combined with the preset number of growth levels G and relevant feature parameters (median Q, boundary values ​​a and b), an asymmetric but adaptive interval system is constructed. For example, when the distribution is skewed and the preset number of growth levels is odd (the preset number of growth levels is equal to the maximum value of the preset growth level index), the crop growth intervals for different growth levels are as follows: When the distribution is skewed and the preset number of growth levels is odd, the crop growth intervals for different growth levels are as follows: The crop growth range for the i=1th growth level is: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. Let be the value of the j-th pixel in the regional crop growth results, Q be the median in the growth results, a be the lower boundary value of the confidence interval of the skewed distribution, b be the upper boundary value of the confidence interval of the skewed distribution, and m be the maximum value of the preset growth level number, and m is an odd number. When the distribution is skewed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0033] Step 106: Match each pixel value in the growth results with the crop growth interval to obtain the crop growth level distribution results of the target area.

[0034] In this step, pixel value matching involves comparing the specific value of a single pixel in the crop growth data with the growth level intervals constructed in step 105 to determine the growth level to which the pixel belongs. The target area is the specific geographical area where crop growth monitoring needs to be carried out. It can be set to different scales such as plots, townships, and counties according to actual needs, and its boundaries are consistent with the coverage area of ​​the remote sensing image.

[0035] Specifically, this step involves first organizing the core input data: clarifying the growth results generated in step 101 (including the specific values ​​of all pixels and their corresponding geographic coordinates), and the growth level intervals constructed in step 105 (including the upper and lower limits of the intervals, the corresponding level names, and growth descriptions). Second, performing pixel value extraction and matching: using Geographic Information System (GIS) tools or programming algorithms, traversing each pixel in the growth results, extracting its pixel value and geographic coordinates; comparing the pixel value sequentially with the upper and lower limits of each growth level interval to determine which interval it falls into, and thus determining the corresponding growth level of the pixel (e.g., "best growth," "average growth," etc.). Finally, the distribution results of growth levels are generated: the growth levels of all pixels are associated with geographic coordinates, and thematic maps of growth levels are drawn in the GIS platform. Different colors are marked according to the level (e.g., dark colors represent poor growth and light colors represent good growth), and elements such as administrative divisions, scale, and legend are marked. At the same time, a data table is output to count the number of pixels, coverage area, and proportion of each growth level, forming a complete distribution result that combines spatial visualization and quantitative statistics, providing a precise basis for agricultural production regulation.

[0036] In summary, the crop growth assessment method of this application, through a six-step closed-loop process, achieves timeliness, accuracy, and universality in growth monitoring, demonstrating significant technical effectiveness. This technical solution enables scientific and adaptive classification of crop growth levels, significantly improving monitoring accuracy and regional applicability while reducing subjective interference and data costs. By calculating the skewness coefficient of growth results, it automatically identifies whether the data follows a normal or skewed distribution, overcoming the limitation of traditional methods that only assume a normal distribution, and making the classification more closely aligned with actual data characteristics. The confidence level is adaptively determined based on the total number of pixels and spatial resolution, and the level interval is constructed by combining the distribution status with characteristic parameters such as mean, standard deviation, median, and confidence interval boundaries. This eliminates the need to rely on empirical thresholds or fixed standards, greatly reducing subjective human influence and improving the adaptability of monitoring in different regions and at different critical periods. This solution requires only data from a single critical period to complete the rating, reducing data requirements and processing workload, while also supporting the setting of the number of levels as needed to meet diverse monitoring scenarios. The overall methodology is simple and universally applicable, and can quickly output the results of growth level distribution, providing accurate and objective technical support for field management, yield prediction, disaster assessment and agricultural insurance.

[0037] Furthermore, to provide a more detailed explanation of the above embodiments, this application also provides another method for determining crop growth, such as... Figure 2 As shown, the following specific steps are provided in this embodiment of the application: Step 201: After acquiring remote sensing images of crops in the target area during key phenological periods, obtain the growth results based on the remote sensing images and determine the distribution of the growth results.

[0038] In this step, acquiring remote sensing images of crops in the target area during key phenological periods and obtaining growth results based on the remote sensing images includes: acquiring crop growth patterns and climate characteristic data in the target area, and determining the key phenological periods of crops based on the crop growth patterns and climate characteristic data; acquiring remote sensing images within the key phenological periods; extracting vegetation indices from the remote sensing images using the year-on-year index method or the anomaly index method, and determining the growth results of the target area based on the vegetation indices to form growth results.

[0039] The remote sensing images within the key phenological period can be remote sensing images of the target year and the comparison year corresponding to the key phenological period. Alternatively, the remote sensing images within the key phenological period can be remote sensing images of a specific time within the key phenological period of a specific year.

[0040] Specifically, the process involves acquiring crop variety profiles for the target area, obtaining field crop observation data within a predetermined period (e.g., the last ten years or five years), determining the complete growth cycle of crops from sowing to maturity and the duration of each stage, and simultaneously collecting climate characteristic data such as temperature, precipitation, and accumulated temperature released by the meteorological department of the target area to construct a "growth stage-climate factor" correlation model. This model quantifies the sensitivity of different growth stages to climate conditions, identifying key phenological periods that play a decisive role in growth, such as the tasseling stage of maize and the jointing stage of wheat, ensuring that the determination of key periods aligns with the actual growth needs of the crops. Secondly, based on the determined key phenological period time windows, an intelligent multi-source remote sensing data platform is used to select images with a cloud-free coverage percentage ≥ a predetermined cloud coverage probability and a spatial resolution ≤ a predetermined spatial resolution in meters. Simultaneously, images from the same period of the target year and comparable years with similar climates are acquired. Finally, a vegetation index calculation algorithm is used to accurately extract core indices from the preprocessed images. These core indices can be the Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EDI), etc. Growth results are calculated using either the year-over-year index method (difference between the target year and the comparison year's index) or the anomaly index method (deviation between the target year and the multi-year average). The growth results can be either the anomaly index value or the year-over-year index value (e.g., ...). T1 is the anomaly index, or T2 = T2 is the year-on-year index value. It is the NDVI value of each pixel in the remote sensing image of the target area in the target year. This involves comparing the NDVI values ​​of each pixel in the target area against the remote sensing images of the same year to form a comprehensive crop growth report, providing high-quality data support for subsequent distribution status determination. Furthermore, the remote sensing images within the key phenological period are images taken at a specific time within that key phenological period of a given year. Based on ground-measured target point data characterizing crop growth in the target area (such as leaf chlorophyll a value and grain moisture content), and combined with relevant indices calculated from remote sensing data (such as normalized difference vegetation index and maximum leaf area index), a growth monitoring model is constructed, and the crop growth results for the target area are obtained based on this model.

[0041] Furthermore, in some embodiments, the key phenological periods of crops in the target area can also be determined based on the crop's growth patterns. For example, they can be defined according to the core growth nodes with clearly defined physiological characteristics of the crop, combined with the key stages for yield formation in agricultural production practice, without relying on complex model calculations. For maize, the jointing stage is an important stage for rapid plant height growth and leaf area expansion, affecting photosynthetic efficiency; the large trumpet stage is a critical period for fertilizer and water requirements, directly related to the quality of female ear differentiation; and the tasseling and silking stage is the core period for pollination and seed setting, determining the ear setting rate and the number of kernels. These three periods can be used as key phenological periods for monitoring maize growth, ensuring that the delineation of key phenological periods both conforms to the crop's growth patterns and takes into account the practical application needs of agricultural production.

[0042] Step 202: Determine the confidence level of the growth results based on the total number of pixels and spatial resolution of the remote sensing image, and determine the relevant feature parameters; In this embodiment, determining the confidence level based on the total number of pixels and spatial resolution of the remote sensing image includes: obtaining the number of rows and columns of the remote sensing image and the spatial resolution of the remote sensing image; obtaining the number of rows and columns of the remote sensing image and the spatial resolution of the remote sensing image; n= Rows × Columns And n is greater than 10,000 n is the size of the remote sensing image used to assess the crop growth in the target area. Rows The number of rows in the remote sensing imagery used to assess crop growth. Columns The number of columns of remote sensing images used to assess crop growth; When n×r 2 ≥100,000 m 2 At that time, determine the first confidence level; When 1 m 2 <n×r 2 <100,000 m 2 At that time, determine the second confidence level.

[0043] Specifically, the total number of pixels is the total quantity of pixels in a remote sensing image, obtained by multiplying the number of rows and columns of the image (n = Rows × Columns, where n is the total number of pixels, Rows is the number of rows, and Columns is the number of columns). It is a core indicator reflecting the amount of image data and is required to meet a preset condition of being greater than a preset number of pixels (such as 10,000) in this embodiment. The spatial resolution is the actual ground area corresponding to a single pixel in the remote sensing image (unit: meter), which determines the portrayal accuracy of surface details in the image. For example, a 10-meter resolution means that a single pixel represents a 10-meter × 10-meter area on the ground. The confidence level is a probability indicator (such as 95%, 98%) characterizing the reliability of the growth trend results, used to determine the confidence interval, and its value is jointly determined by the total number of pixels and the spatial resolution.

[0044] Specifically, first, automatically parse the number of rows (Rows) and columns (Columns) of the remote sensing image metadata, and read its nominal spatial resolution r (unit: meter). For example, Sentinel-2 images are usually 10 meters. Subsequently, calculate the total number of pixels n = Rows × Columns. The system has a built-in verification mechanism to ensure that n > 10,000. If not satisfied, it will prompt insufficient data coverage and require supplementing or replacing the image. Then, calculate the effective monitoring area A = n × r 2 , which reflects the actual surface range involved in the analysis. When A ≥ the first preset area threshold, such as 100,000 m 2 , it indicates that the sample size is sufficient and the statistical representativeness is strong. At this time, determine that the first confidence level (P1) ≥ 98%; when 1 m 2 < A < the first preset area threshold, such as 1 m 2 < A < 100,000 m 2 , determine that the range of the second confidence level is 95% ≤ P2 < 98%, where P2 is the second confidence level.

[0045] In some embodiments, the selection of relevant characteristic parameters according to the distribution state and confidence level of the growth trend results includes: determining whether the distribution state of the growth trend results is a normal distribution; if so, based on the confidence level, look up the standard score in the normal distribution table; obtain the mean of the crop growth trend results and the standard deviation of the crop growth trend results; determine the standard score, the mean of the crop growth trend results, and the standard deviation of the crop growth trend results as relevant characteristic parameters.

[0046] The distribution includes normal and skewed distributions. The standard score (z-value) in a normal distribution measures the degree to which data deviates from the mean, reflecting the data's relative position within the distribution. The mean is the arithmetic mean of all pixel values ​​in the growth results, reflecting the overall average growth level. The standard deviation reflects the dispersion of pixel values ​​from the mean, measuring the range of fluctuation in growth data. Relevant characteristic parameters are the core indicators for constructing growth grade intervals; under a normal distribution, these include the standard score, mean, and standard deviation.

[0047] Specifically, first use the formula (where x) i For each individual pixel value, μ is the mean, n is the number of pixels, and σ is the standard deviation. The skewness coefficient sk of the crop growth results is calculated. If sk = 0, the dataset is considered to follow a normal distribution. Next, standard scores are matched based on the confidence level determined in step 201 (e.g., the z-value can be determined by looking up the normal distribution table based on the confidence level; for example, a 95% confidence level corresponds to z = 1.96, and a 98% confidence level corresponds to z = 2.33). Then, the mean and standard deviation of the crop growth results are obtained. The mean is obtained by summing all pixel values ​​and dividing by the total number of pixels n. The standard deviation σ is obtained by summing the squares of the differences between each pixel value and the mean, dividing by n, and then taking the square root. Finally, the obtained standard score z, the calculated mean μ, and the standard deviation σ are used together to determine the relevant feature parameters under the normal distribution scenario, providing data support for subsequent growth interval construction. By objectively determining the distribution type using the skewness coefficient, the unreasonable assumption of a default normal distribution in traditional methods is avoided, ensuring that the extracted feature parameters closely reflect the true patterns of the data. The calculation of the mean and standard deviation covers all pixel data, comprehensively reflecting the central tendency and dispersion of the dataset, thus guaranteeing the representativeness of the feature parameters. The feature parameter system formed by these three factors provides scientific and comprehensive data support for subsequent adaptive thresholding, effectively reducing subjective human interference and improving the accuracy and standardization of crop growth monitoring.

[0048] In some embodiments, selecting relevant feature parameters based on the distribution state and confidence level of the growth results includes: when the distribution state of the growth results is skewed, determining the confidence interval using the Bootstrap method based on the confidence level, and obtaining the boundary value based on the confidence interval; obtaining the median of the crop growth results in the growth results; and determining the boundary value and the median of the crop growth results as relevant feature parameters.

[0049] The Bootstrap method is a resampling statistical technique that determines the confidence interval for skewed data by drawing samples multiple times and calculating statistics, without requiring assumptions about the data distribution type. Under a skewed distribution, the confidence interval [a, b] encompasses the core range of the data, where 'a' is the lower boundary value and 'b' is the upper boundary value, reflecting the main distribution range of crop growth results. The median is the value in the middle after sorting the crop growth results numerically; it is not affected by extreme values ​​and reflects the central tendency of the data.

[0050] Specifically, firstly, the distribution state is determined based on the calculated skewness coefficient sk. If sk ≠ 0, it is confirmed that the growth results follow a skewed distribution. Secondly, the confidence interval under the skewed distribution is determined: according to the confidence level P determined in step 202 (P is the first confidence level or the second confidence level), the Bootstrap method is used to randomly select samples of the same size as the original dataset from the growth results, and the sampling is repeated for a preset number of times (e.g., more than 1500 times); the statistics are calculated for each sample obtained from each sampling, and all statistics are sorted in ascending order. The corresponding quantiles are taken as boundary values ​​(e.g., when P=98%, the 1% quantile is taken as a, and the 99% quantile is taken as b), forming the confidence interval [a, b]. Then, the median Q is calculated: all cell values ​​in the growth results are sorted in ascending order. If the total number of cells n is odd, the value of the (n+1) / 2th cell is taken as Q; if n is even, the average of the values ​​of the (n / 2)th cell and the (n / 2+1)th cell is taken as Q. Finally, the boundary values ​​a and b of the confidence interval, along with the median Q, are jointly determined as relevant feature parameters in the skewed distribution scenario, providing core data support for the subsequent construction of growth level intervals. Thus, the above method addresses the shortcomings of existing technologies that default to a normal distribution and ignore the characteristics of skewed data. The Bootstrap method eliminates the need to pre-determine the data distribution type, directly determining the confidence interval based on the inherent characteristics of the skewed data. Combined with the median, it constructs a feature parameter system that fully adapts to the real-world patterns of skewed distributions, making feature parameter extraction more realistic and fundamentally improving the scientific rigor of subsequent level classification. The confidence intervals in this method are dynamically generated through Bootstrap resampling combined with confidence levels, and the median is also calculated based on the current dataset. Both adaptively adjust with changes in growth data in the target region and during key phenological periods, avoiding the limitations of fixed standards and making the feature parameters more regionally specific and adaptable to the current situation. The median itself is unaffected by extreme values. The Bootstrap method, through multiple resampling and quantile filtering, effectively filters out interference from a few outlier pixels, ensuring that the confidence interval boundary values ​​and the median accurately reflect the core distribution characteristics of the crop growth, thus improving the reliability and representativeness of the feature parameters. This method, based on the crop growth results of a single key period, can complete the confidence interval calculation through resampling technology, eliminating the need for additional historical data accumulation. While ensuring effectiveness, it simplifies the data collection and processing process, meeting the needs of efficient monitoring in practical applications.

[0051] Step 203: Based on the boundary values ​​of the confidence interval corresponding to the confidence level, and combining the preset number of growth levels with the selected relevant feature parameters, construct crop growth intervals for different growth levels.

[0052] In some embodiments, when the distribution is skewed and the number of preset growth levels is odd, the step of constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with the number of preset growth levels and selected relevant feature parameters, includes: The crop growth range for the i=1th growth level is: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. Let be the value of the j-th pixel in the regional crop growth results, Q be the median in the growth results, a be the lower boundary value of the confidence interval of the skewed distribution, b be the upper boundary value of the confidence interval of the skewed distribution, and m be the maximum value of the preset growth level number, and m is an odd number. When the distribution is skewed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0053] The aforementioned method for dividing crop growth intervals dynamically adapts to data distribution characteristics, employing standard scores, mean-standard deviation or Bootstrap confidence intervals, and medians to construct feature parameters for skewed distributions. This completely overcomes the limitations of traditional fixed thresholds, ensuring that interval divisions align with the true patterns of crop growth in the region. This significantly improves the accuracy of growth level judgment and provides a reliable basis for precisely identifying areas with superior or inferior growth. Simultaneously, it rapidly generates datasets and constructs intervals based on remote sensing images of key phenological periods, eliminating the need for long-term data accumulation and enabling real-time dynamic monitoring of growth. This solves the lag problem of traditional methods, helping producers promptly capture abnormal growth signals and accurately grasp the critical timing for water and fertilizer regulation and disaster intervention. Furthermore, it supports flexible conversion between odd and even levels, allowing for adjustments to the grading precision based on different scenarios such as large-scale surveys and precision field management. The results can directly connect with agricultural production practices such as yield forecasting and resource optimization, providing strong support for scientific decision-making and helping to improve agricultural production efficiency and output stability.

[0054] It is worth noting that when the preset number of growth levels is even, the step of constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with the preset number of growth levels and selected relevant feature parameters, includes: dividing the growth levels according to the level division method when m=G+1, wherein the preset number of growth levels is G, including: obtaining crop growth intervals with a preset number of growth levels of m; merging any two adjacent crop growth intervals to obtain a new crop growth interval; replacing the any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G. For example, to obtain a preset number of growth levels G=4, it is necessary to first find a preset growth level G+1, which is the crop growth interval of different growth levels with a preset number of growth levels of 5, such as level 1 to level 5. Each level has a corresponding division interval, namely division interval 1 to division interval 5. At this time, if it is determined that the crop growth interval with a preset number of growth levels of 4 can be obtained by selecting any two adjacent crop growth intervals and merging them to obtain a new crop growth interval; and replacing the any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0055] For example, to determine the crop growth intervals with a preset number of growth levels G=4, it is necessary to find all crop growth intervals for preset growth levels m=G+1. For instance, levels 1 to 5 correspond to intervals 1 to 5, with growth from best to worst being level 1 > level 2 > level 3 > level 4 > level 5. If we need to determine the crop growth intervals with a preset number of growth levels G=4, all possible merging possibilities are as follows: Merging Level 1 and Level 2: The new growth level is 4, namely "New Level 1 (Original Level 1 + Original Level 2, corresponding to the merging of Division 1 and Division 2)", "New Level 2 (Original Level 3, corresponding to Division 3)", "New Level 3 (Original Level 4, corresponding to Division 4)", and "New Level 4 (Original Level 5, corresponding to Division 5)". The growth from best to worst is New Level 1 > New Level 2 > New Level 3 > New Level 4.

[0056] Merging Level 2 and Level 3: The new growth level is 4, namely "New Level 1 (original Level 1, corresponding to Division 1)", "New Level 2 (original Level 2 + original Level 3, corresponding to Division 2 and Division 3 merged)", "New Level 3 (original Level 4, corresponding to Division 4)" and "New Level 4 (original Level 5, corresponding to Division 5)". The growth from best to worst is New Level 1 > New Level 2 > New Level 3 > New Level 4.

[0057] Merging Levels 3 and 4: The new growth levels are 4, namely "New Level 1 (Original Level 1, corresponding to Division 1)", "New Level 2 (Original Level 2, corresponding to Division 2)", "New Level 3 (Original Level 3 + Original Level 4, corresponding to the merger of Division 3 and Division 4)", and "New Level 4 (Original Level 5, corresponding to Division 5)". The growth from best to worst is New Level 1 > New Level 2 > New Level 3 > New Level 4.

[0058] Merging Levels 4 and 5: The new growth levels are 4, namely "New Level 1 (Original Level 1, corresponding to Division 1)", "New Level 2 (Original Level 2, corresponding to Division 2)", "New Level 3 (Original Level 3, corresponding to Division 3)", and "New Level 4 (Original Level 4 + Original Level 5, corresponding to the merger of Division 4 and Division 5)". The growth from best to worst is New Level 1 > New Level 2 > New Level 3 > New Level 4.

[0059] Based on the significant technical effects of the above-mentioned even-numbered growth level merging scheme, it fully adapts to actual monitoring needs and optimizes the grading logic: First, it flexibly meets the needs of diverse monitoring scenarios. Different scenarios in agricultural production have different requirements for the precision of growth grading. For example, large-area agricultural surveys require coarser grading (4 levels), while precision field management may require finer grading (5 levels). This scheme, through the flexible conversion of "odd-numbered grading + adjacent merging," can quickly adapt to the monitoring needs of different number of grading levels without redesigning the grading algorithm, without changing the core feature parameters and confidence interval logic, thus improving the practicality and scenario adaptability of the method. Second, it ensures the continuity and consistency of the grading logic. The merging process strictly follows the principle of "adjacent level merging," without breaking the original odd-numbered level growth gradient relationship (from good to bad or vice versa), ensuring that the growth ranking of the new level is consistent with the original grading logic. At the same time, the merged intervals are natural connections to the original intervals, without numerical breaks or overlaps, ensuring the scientific nature and data continuity of the growth level division, and avoiding result deviations caused by adjustments to the number of grading levels. Third, it reduces data processing costs and operational complexity. Compared to redesigning independent grading algorithms for even-numbered levels, this scheme directly reuses the division results for odd-numbered levels, eliminating the need for additional calculation of feature parameters and reconstruction of grading inequalities, thus reducing data processing steps and computational load. Simultaneously, the merging rules are simple and clear, and can be automatically executed programmatically, reducing the difficulty of manual operation and the probability of errors, thereby improving the efficiency of crop growth monitoring. Fourth, it balances regional adaptability and result comparability. The merged grade intervals are still generated based on the confidence intervals and feature parameters of the original region, preserving the core characteristics of regional crop growth and avoiding the problem of regional inapplicability of fixed grading standards. Furthermore, crop growth results from different periods in the same region, and from the same period in different regions, can be merged and converted using a unified odd-numbered grading basis, ensuring horizontal and vertical comparability of results and providing convenience for cross-regional and cross-time period agricultural analysis.

[0060] In some embodiments, when the distribution is normally distributed and the number of preset growth levels is odd, the step of constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with the preset number of growth levels and selected relevant feature parameters, includes: The crop growth range for the i=1th growth level is: ; The crop growth range for the growth level 1 < i < m is: ; The crop growth range for the i=m growth level is: ; in, The value of the j-th pixel representing the regional crop growth status. Let n be the mean of the crop growth results, and n be the number of pixels in the crop growth results. The standard deviation of the crop growth results is represented by z, where z is the standard score and m is the maximum value of the preset growth level number, and m is an odd number. When the distribution is normally distributed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0061] The above-mentioned normal distribution-based classification of crop growth intervals is based on characteristic parameters constructed from standard scores, mean, and standard deviation. This aligns with the normal distribution characteristics of some growth data, avoiding the subjective bias of traditional fixed thresholds and significantly improving the accuracy of growth assessment. It can precisely distinguish different growth levels. Utilizing remote sensing images of key phenological periods, core parameters can be rapidly calculated without long-term data accumulation, enabling timely monitoring of growth and helping producers seize opportune moments for regulation. The grading results closely reflect regional realities, providing a scientific basis for precision irrigation, variable-rate fertilization, and other agricultural activities, effectively supporting agricultural production decisions and improving resource utilization efficiency and crop yield stability.

[0062] In some embodiments, when the preset number of growth levels is even, the construction of crop growth intervals with different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with the preset number of growth levels and selected relevant feature parameters, can be as follows: If the number of crop growth levels is even, let m = G+1, where G is even. Then, the level division can be performed according to the level division method when m = G+1. This can be achieved by merging any two adjacent levels in the growth interval with a preset growth level of G+1 to obtain a new level, i.e., calculating the preset number of growth levels as 2. To determine the crop growth range, it is necessary to first obtain crop growth ranges with a preset growth level of 2+1. For example, if the preset growth level is 3, the growth levels are growth level 1, growth level 2, and growth level 3, and each growth level corresponds to a growth range. When the preset growth level is an even number (2), for example: a new growth level 1 can be the original growth level 1, and its growth range is the growth range corresponding to the original growth level 1; the growth range of the new growth level 2 can be a new growth range determined by merging the growth ranges corresponding to the original growth levels 2 and 3. Alternatively, the growth range of the new growth level 1 can be a newly determined growth range by merging the growth ranges corresponding to the original growth levels 1 and 2, and the new growth level 2 can be the growth range corresponding to growth level 3.

[0063] Step 204: Based on each pixel value and the crop growth range, determine the crop growth level distribution results for the target area.

[0064] First, the global crop growth results generated in step 201 are called to determine the distribution status (normal or skewed) of the crop growth results and the crop growth intervals. Then, each pixel value in the growth results is matched with the crop growth intervals one by one to determine the crop growth level distribution results of the target area.

[0065] After determining the crop growth level distribution results in the target area, the original geographic coordinate information of the remote sensing image is integrated to reconstruct a spatialized growth level distribution map of all pixels with level labels, and simultaneously output statistical results such as the area proportion and spatial clustering areas of each level. Specifically, after determining the crop growth level distribution results in the target area, the implementation method of "coordinate fine matching - multi-level reconstruction - intelligent statistics - visualization output" is adopted, as follows: First, the automatic registration process between remote sensing image and GIS data is initiated to extract the geographic coordinate parameters (latitude and longitude range, projected coordinate system) of the original image. Through a matching algorithm based on surface features, the pixels with growth level labels are accurately aligned with GIS vector data such as administrative boundaries and field parcels, eliminating geometric distortion errors and ensuring that the level label of each pixel corresponds one-to-one with the actual geographical location. Secondly, a three-tiered reconstruction system of "pixel-field-region" is constructed: First, using the field as the basic unit, the level labels of all pixels within the same field are integrated to determine the overall growth level of the field; then, combined with regional topography, water systems, and other geographical elements, the discrete level pixels are spatially smoothed to generate a continuous spatialized growth level distribution map that fits the actual planting pattern, supporting both raster and vector formats for output. Finally, multi-dimensional intelligent statistics and visualization are carried out simultaneously: On the one hand, the area, proportion, and spatial clustering index of each growth level are calculated, automatically identifying abnormal growth clusters and marking the core range and radius of influence; on the other hand, GIS visualization technology is used to design a graded color system (e.g., dark green for superior levels and dark red for poor levels), overlaying base map elements such as administrative boundaries and transportation networks to generate a dynamic interactive map. Statistical reports are also output, including information such as the area ranking of each level, the topological relationship of clustered areas, and the field-level growth compliance rate, achieving a deep integration of spatial distribution and quantitative statistics, providing intuitive and accurate decision support for agricultural regulation.

[0066] Furthermore, this application provides specific examples as follows: Taking the monitoring of rice growth in a certain area as an example, the present invention will be further described in detail with reference to charts and specific implementation methods, but this application is not limited.

[0067] Step 1: Determination of phenological periods and processing of remote sensing data Based on agricultural surveys and agricultural meteorological observation data, the corn growing season in a certain region is divided into the following table, Table 1.

[0068]

[0069] Growth monitoring was conducted during the critical tasseling stage of maize growth. Remote sensing images with no cloud cover from 2024 and 2025 for this time period were retrieved and downloaded. The final selected remote sensing images are shown in Table 2. The acquired Sentinel-2 remote sensing images underwent radiometric calibration, atmospheric correction, and geometric correction.

[0070]

[0071] Step 2: Production of Regional Crop Growth Results Based on Remote Sensing Based on the preprocessed remote sensing images, the crop growth results are calculated using the year-on-year index method, as shown in the following formula: In the formula, T represents the anomaly index. The image shows the NDVI value of each pixel in the remote sensing image of the study area in 2025. This represents the NDVI value of each pixel in the remote sensing image of the study area in 2024.

[0072] Step 3: Setting the crop growth level Based on the needs of crop growth monitoring in practical applications, the growth level G is set into 5 levels: Level 1 is the best growth, Level 2 is relatively good growth, Level 3 is average growth, Level 4 is poor growth, and Level 5 is the worst growth.

[0073] Step 4: Calculation of skewness coefficient of crop growth results and determination of data distribution By statistically analyzing the crop growth results, the pixel value, the number of pixels (n=2,312,800), and the mean... =0.1190, standard deviation =0.0654. According to the skewness calculation formula, the skewness of the crop growth results is sk=(-0.07), and the crop growth results follow a skewed distribution.

[0074] Step 5: Determining the confidence interval and parameters for crop growth results The study area has a remote sensing image pixel size of n = 2,312,800, an image resolution of r = 10 meters, and nr² = 231,280,000 square meters, which is greater than 100,000. Therefore, the confidence level P is set at 98%. The Bootstrap method is used to repeatedly sample 1500 times. The median of the 1500 samples is calculated, sorted from smallest to largest, and the 1st percentile is taken as a, and the 99th percentile as b. Therefore, a = (-0.110) and b = 0.2428. Step Six: Adaptive Threshold Classification of Crop Growth Levels Based on Normal Distribution Based on the confidence interval [-0.1102, 0.2428] under the condition that the crop growth results follow a skewed distribution, the median of the crop growth results is Q=0.1256, the number of crop growth levels is m=5, the numerical range of crop growth results at different levels is automatically calculated, and the crop growth status is clarified according to the characteristics of the changes in the crop growth results. The results are shown in Table 3 below.

[0075]

[0076] Step 7: Mapping the results of crop growth grading Based on the grading range in step six, the maize growth results obtained in step two during the tasseling stage are graded, and a maize growth grade display chart is created, as shown in the following figure. Figure 3 As shown.

[0077] Furthermore, as a response to the above Figure 1 , Figure 2 To implement the method shown, this application provides a device for determining crop growth. This device embodiment corresponds to the aforementioned method embodiment. For ease of reading, this device embodiment will not repeat the details of the aforementioned method embodiment, but it should be clear that the device in this embodiment can implement all the contents of the aforementioned method embodiment. Specifically, as shown... Figure 4 As shown, the device includes: The acquisition unit 41 is used to acquire remote sensing images of crops in the target area during key phenological periods and obtain growth results based on the remote sensing images. The growth results are characterized by the pixel values ​​of the crops. The determination unit 42 is used to calculate the skewness coefficient of the growth results in the acquisition unit 41, and determine the distribution state of the growth results based on the skewness coefficient. The determining unit 42 is used to determine the confidence level of the growth results based on the total number of pixels and spatial resolution of the remote sensing image. The determining unit 42 is used to select relevant characteristic parameters based on the distribution status and confidence level of the growth results; The construction unit 43 is used to construct crop growth intervals for different growth levels based on the boundary values ​​of the confidence interval corresponding to the confidence level, combined with the preset number of growth levels and the relevant feature parameters selected in the determination unit 42. The result acquisition unit 44 is used to match each pixel value in the growth results of the acquisition unit 41 with the crop growth interval in the construction unit 43 to obtain the crop growth level distribution results of the target area.

[0078] Furthermore, such as Figure 5As shown, when the distribution is skewed and the number of preset growth levels is odd, the construction unit 43 includes: The crop growth range for the i=1th growth level is: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. Let be the value of the j-th pixel in the regional crop growth results, Q be the median in the growth results, a be the lower boundary value of the confidence interval of the skewed distribution, b be the upper boundary value of the confidence interval of the skewed distribution, and m be the maximum value of the preset growth level number, and m is an odd number. When the distribution is skewed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0079] Furthermore, such as Figure 5 As shown, when the distribution is a normal distribution and the preset number of growth levels is odd, the construction unit 43 includes: When the distribution is normal and the number of preset growth levels is odd, the crop growth intervals for different growth levels are as follows: The crop growth range for the i=1th growth level is: ; The crop growth range for the growth level 1 < i < m is: ; The crop growth range for the i=m growth level is: ; in, The value of the j-th pixel representing the regional crop growth status. Let n be the mean of the crop growth results, and n be the number of pixels in the crop growth results. The standard deviation of the crop growth results is represented by z, where z is the standard score and m is the maximum value of the preset growth level number, and m is an odd number. When the distribution is normally distributed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

[0080] Furthermore, such as Figure 5 As shown, the determining unit 42 includes: The acquisition module 421 is used to acquire the number of rows and columns of the remote sensing image and the spatial resolution of the remote sensing image; Module 422 is defined as follows: n = Rows × Columns, where n is greater than 10,000. n is the size of the remote sensing image used for the crop growth results in the target area, Rows is the number of rows in the remote sensing image used for the crop growth results, and Columns is the number of columns in the remote sensing image used for the crop growth results. When n×r 2 ≥100,000 m 2 At that time, determine the first confidence level; When 1 m 2 <n×r 2 <100,000 m 2 At that time, determine the second confidence level.

[0081] Furthermore, such as Figure 5 As shown, the building unit 43 further includes: The judgment module 431 is used to determine whether the distribution of the growth results is a normal distribution; If the judgment module 431 determines that it is yes, it looks up the standard score in the positive pressure distribution table based on the confidence level. The parameter module 432 is used to obtain the mean and standard deviation of the crop growth results. The parameter determination module 432 is used to determine the standard score, the mean of crop growth results, and the standard deviation of crop growth results as relevant characteristic parameters.

[0082] Furthermore, such as Figure 5 As shown, the parameter determination module 432 further includes: When the distribution of growth results is skewed, the confidence interval is determined using the Bootstrap method according to the confidence level, and the boundary value is obtained based on the confidence interval. Obtain the median of crop growth results from the growth results; The boundary values ​​and the median of crop growth results were determined as relevant characteristic parameters.

[0083] Furthermore, such as Figure 5 As shown, the acquisition unit 41 includes: The crop data acquisition module 411 is used to acquire crop growth patterns and climate characteristics data in the target area, and to determine the key phenological periods of crops based on the crop growth patterns and climate characteristics data. The crop data acquisition module 411 is used to acquire remote sensing images during the key phenological period; The dataset determination module 412 is used to extract vegetation indices from the remote sensing images of the crop data acquisition module 411 using the year-on-year index method or the anomaly index method, and to determine the growth results of the target area based on the vegetation indices, thereby forming growth results.

[0084] Furthermore, this application also provides a storage medium, the storage medium including a stored program, which, when the program is executed, controls the device where the storage medium is located to execute the above-described method for determining crop growth.

[0085] Furthermore, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for determining crop growth conditions as described above.

[0086] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0087] It is understood that the relevant features in the above methods and apparatus can be referenced interchangeably. Furthermore, the terms "first," "second," etc., in the above embodiments are used to distinguish between embodiments and do not represent the superiority or inferiority of any particular embodiment.

[0088] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0089] The acquisition and / or use of user data involved in the embodiments of this application strictly comply with the laws, regulations, and industry standards of relevant countries and regions. The collection and acquisition of data involved in the embodiments of this application are all done in advance through proactive prompts or prominent markings to obtain user authorization or full authorization from all parties. The processing, manipulation, forwarding, and use of data involved in the embodiments of this application are all carried out with the full knowledge and authorization of the user or relevant parties. In implementing the embodiments of this application, the types of data or information, scope of use, and usage scenarios that may be involved are communicated to the user or relevant parties and authorization is obtained through appropriate means. The specific methods of notification and authorization may vary according to actual circumstances, and this application is not limited in this regard. The processing of personal information involved in the embodiments of this application is only carried out under circumstances with a legal basis (such as obtaining the consent of the personal information subject or being necessary for the performance of a contract), and will be processed within the prescribed or agreed scope. Sensitive personal information such as biometric information, medical and health information, financial account information, and precise location information involved in the embodiments of this application are all processed under the premise of specific purpose and sufficient necessity, and with the separate authorization and consent of the user or relevant parties. In some embodiments of this application, if the user or related party refuses to process personal information other than the necessary information required for the basic functions, it will not affect the use of the basic functions of the embodiments of this application.

[0090] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0091] In addition, the memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0092] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0095] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0096] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0097] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0098] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0099] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0100] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The above descriptions are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of determining a crop growth condition, characterized by, include: Remote sensing images of crops in the target area during key phenological periods are acquired, and growth results are obtained based on the remote sensing images. The growth results are characterized by the pixel values ​​of the crops. Calculate the skewness coefficient of the growth results, and determine the distribution state of the growth results based on the skewness coefficient; The confidence level of the growth results is determined based on the total number of pixels and spatial resolution of the remote sensing image. Select relevant characteristic parameters based on the distribution and confidence level of the growth results; Based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, and combined with the preset number of growth levels and the selected relevant feature parameters, crop growth intervals for different growth levels are constructed. Each pixel value in the growth results is matched with the crop growth interval to obtain the crop growth level distribution results for the target area.

2. The method according to claim 1, characterized in that, The process involves constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with a preset number of growth levels and selected relevant feature parameters. When the distribution is skewed and the number of preset growth levels is odd, the crop growth ranges for different growth levels are as follows: The crop growth range for the i=1th growth level is: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; No. The growth range of crops in the growth grade is as follows: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. Let be the value of the j-th pixel in the regional crop growth results, Q be the median in the growth results, a be the lower boundary value of the confidence interval of the skewed distribution, b be the upper boundary value of the confidence interval of the skewed distribution, and m be the maximum value of the preset growth level number, and m is an odd number. When the distribution is skewed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

3. The method according to claim 1, characterized in that, The process involves constructing crop growth intervals for different growth levels based on the boundary values ​​of the confidence intervals corresponding to the confidence levels, combined with a preset number of growth levels and selected relevant feature parameters. When the distribution is normal and the number of preset growth levels is odd, the crop growth intervals for different growth levels are as follows: The crop growth range for the i=1th growth level is: ; The crop growth range for the growth level 1 < i < m is: ; The crop growth range for the i=m growth level is: ; Where i is a sequence parameter used to distinguish different growth levels. The value of the j-th pixel representing the regional crop growth status. The average of the growth results. The standard deviation of the crop growth results is represented by z, where z is the standard score and m is the maximum value of the preset growth level number, and m is an odd number. When the distribution is normally distributed and the preset number of growth levels is even, the growth levels are classified according to the classification method when m=G+1, where the preset number of growth levels is G, including: Obtain crop growth intervals with a preset number of m growth levels; Select any two adjacent crop growth intervals and merge them to obtain a new crop growth interval; Replace any two adjacent crop growth intervals with the new crop growth interval to determine the crop growth interval when the preset number of growth levels is G.

4. The method according to claim 1, characterized in that, The determination of confidence level based on the total number of pixels and spatial resolution of remote sensing images includes: Obtain the number of rows and columns of the remote sensing image, and the spatial resolution of the remote sensing image; n= Rows × Columns And n is greater than 10,000 n is the size of the remote sensing image used to assess the crop growth in the target area. Rows The number of rows in the remote sensing imagery used to assess crop growth. Columns The number of columns of remote sensing images used to assess crop growth; when n x r 2 ≥ 100,000 m 2 a first confidence level is determined; when 1 m 2 < n x r 2 < 100,000 m 2 a second confidence level is determined.

5. The method according to claim 4, characterized in that, The selection of relevant feature parameters based on the distribution status and confidence level of the growth results includes: Determine whether the distribution of the growth results follows a normal distribution; If so, based on the stated confidence level, find the standard score in the normal distribution table; The mean and standard deviation of crop growth results for which growth results were obtained; The standard score, the mean of crop growth results, and the standard deviation of crop growth results were determined as relevant characteristic parameters.

6. The method according to claim 5, characterized in that, The selection of relevant feature parameters based on the distribution status and confidence level of the growth results includes: When the distribution of growth results is skewed, the confidence interval is determined using the Bootstrap method according to the confidence level, and the boundary value is obtained based on the confidence interval. Obtain the median of crop growth results from the growth results; The boundary values ​​and the median of crop growth results were determined as relevant characteristic parameters.

7. The method according to any one of claims 1-6, characterized in that, The acquisition of remote sensing images of crops in the target area during key phenological stages, and the obtaining of growth results based on the remote sensing images, includes: Obtain data on crop growth patterns and climate characteristics in the target area, and determine the key phenological periods of crops based on the data on crop growth patterns and climate characteristics; Acquire remote sensing images during the key phenological periods; The vegetation index is extracted from remote sensing images using the same index method or the anomaly index method, and the growth status of the target area is determined based on the vegetation index to form the growth status results.

8. A device for determining the growth status of crops, characterized in that, include: The acquisition unit is used to acquire remote sensing images of crops in the target area during key phenological periods and obtain growth results based on the remote sensing images. The growth results are characterized by the pixel values ​​of the crops. A determining unit is used to calculate the skewness coefficient of the growth results in the acquiring unit, and to determine the distribution state of the growth results based on the skewness coefficient; The determining unit is used to determine the confidence level of the growth results based on the total number of pixels and spatial resolution of the remote sensing image. The determining unit is used to select relevant characteristic parameters based on the distribution status and confidence level of the growth results; The construction unit is used to construct crop growth intervals for different growth levels based on the boundary values ​​of the confidence interval corresponding to the confidence level, combined with the preset number of growth levels and the relevant feature parameters selected in the unit. The result acquisition unit is used to match each pixel value in the growth results with the construction unit, which is used to match the crop growth interval in the target area, so as to obtain the crop growth level distribution results in the target area.

9. A storage medium comprising a stored program, characterized in that, When the program is running, it controls the device containing the storage medium to execute the method for determining the crop growth status as described in any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for determining the crop growth status as described in any one of claims 1 to 7.

Images