Intelligent precision decision method and system for organic fertilizer field application

Abstract

The application discloses an organic fertilizer field returning intelligent precision decision method and system, and relates to the technical field of agricultural intelligent management. The organic fertilizer field returning intelligent precision decision method is used for acquiring soil data, crop varieties, target yield and other information, analyzing a nutrient supplement set and a crop absorption response set, and accurately calculating a fertilizer recommendation dose. The system uses historical fertilization records and crop growth image data, extracts crop growth characteristics in combination with an image recognition model, and dynamically optimizes a fertilization scheme. Finally, the system controls a fertilization machine to perform a fertilization operation according to the recommendation dose set. The application dynamically analyzes soil, crops and historical fertilization data, accurately calculates crop nutrient requirements, combines soil detection, crop varieties and target yield, and formulates a target supplement amount of nutrient elements. In combination with historical fertilization records and crop growth image data, the application deduces crop absorption responses, dynamically adjusts the fertilizer recommendation dose, avoids excessive or insufficient fertilizers, and thus improves fertilizer utilization efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent agricultural management technology, specifically to an intelligent and precise decision-making method and system for returning organic fertilizer to the field. Background Technology

[0002] Returning organic fertilizer to the field is an effective measure to improve soil fertility, enhance soil structure, and promote healthy crop growth. With the advancement of agricultural modernization, fertilization management is gradually shifting from traditional experience-based management to precision and intelligent management. Traditional fertilization methods are often based on fixed fertilization standards, neglecting the differences in soil types, crop growth stages, and climatic conditions, leading to low fertilizer efficiency, resource waste, and increasingly serious environmental pollution problems.

[0003] In recent years, with the development of precision agriculture and smart agriculture technologies, precision fertilization has become an important means to improve agricultural production efficiency and resource utilization efficiency. Precision fertilization methods typically rely on advanced soil testing technology, crop growth monitoring technology, and data analysis models. Based on real-time data collection and analysis of soil nutrient status, crop requirements, and climatic factors, they can accurately determine the amount and timing of fertilization.

[0004] The limitations of existing technologies include at least the following problems: In the process of applying organic fertilizers, existing technologies usually rely on traditional fixed fertilizer application rates and fail to take into account the dynamic changes in crop growth, such as the differences in nutrient requirements of crops at different growth stages and under different soil conditions. Due to the lack of a real-time adjustment mechanism for crop absorption response, existing methods often lead to inaccurate fertilizer application, which may result in nutrient over- or under-application, thereby affecting crop growth and yield, and may even lead to environmental pollution and fertilizer waste, increasing the difficulty and cost of farmland management. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent and precise decision-making method and system for returning organic fertilizer to the field, which solves the problem of inaccurate fertilization in existing technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent and precise decision-making method for returning organic fertilizer to the field, comprising the following steps: acquiring current soil testing data, current crop variety, and target crop yield for a designated plot; analyzing the nutrient replenishment set for the designated plot based on the current soil testing data and target crop yield; acquiring historical fertilization records and historical crop growth image data for the designated plot; inputting the historical crop growth image data into a pre-trained image recognition model to extract historical crop growth characteristics of the designated plot; combining the historical crop growth characteristics with historical fertilization records to analyze and obtain the nutrient crop absorption response set for the designated plot; determining the recommended fertilization dose set for the designated plot based on the nutrient replenishment set and the nutrient crop absorption response set; and controlling the designated machinery to perform organic fertilizer application operations on the designated plot according to the recommended fertilization dose set.

[0007] Furthermore, current soil testing data includes the actual content values ​​of several types of nutrient elements, while the nutrient supplement set includes the target supplementation amounts of several types of nutrient elements.

[0008] Further, the specific steps for analyzing the nutrient replenishment set of the designated plot are as follows: Obtain the fertilizer requirement parameters per unit yield of the current crop in the designated plot, and analyze the target total fertilizer requirement of various nutrient elements in combination with the target yield of the current crop. The fertilizer requirement parameters per unit yield of the current crop include the unit fertilizer requirement of various nutrient elements corresponding to the current crop yield; Obtain the soil effective supply coefficient of various nutrient elements in the designated plot, and analyze the absorbable supply of various nutrient elements in combination with the corresponding actual content values; Subtract the target total fertilizer requirement of various nutrient elements from the corresponding absorbable supply to obtain the target replenishment amount of various nutrient elements.

[0009] Furthermore, the historical fertilization record information includes the type, dosage, and time of each fertilization operation during the crop's historical growth cycle, and the historical crop growth image data consists of visible light images, red light images, and near-infrared images of the crop leaf area before and several days after each fertilization operation during the crop's historical growth cycle.

[0010] Furthermore, the visible light image is composed of several RGB pixels, each RGB pixel corresponding to a red channel pixel value, a green channel pixel value and a blue channel pixel value, respectively. The red light image is composed of several single-channel red light pixels, each red light pixel corresponding to a red light intensity value. The near-infrared image is composed of several single-channel infrared pixels, each infrared pixel corresponding to a near-infrared reflection intensity value.

[0011] Furthermore, the image recognition model includes an image input processing layer, a chlorophyll analysis layer, a leaf area analysis layer, a vegetation index analysis layer, an image color difference extraction layer, and an output layer. Historical crop growth characteristics include crop chlorophyll index, crop leaf area index, crop normalized vegetation index, and crop image color difference value.

[0012] Further, the specific steps for extracting the historical crop growth characteristics of the designated plot are as follows: In the image input processing layer of the image recognition model, the visible light, red light, and near-infrared images of the crop leaf area of ​​the designated plot before and several days after each fertilization operation are uniformly processed; in the chlorophyll analysis layer of the image recognition model, the corresponding crop chlorophyll index is analyzed based on the processed visible light images of the crop leaf area; in the leaf area analysis layer of the image recognition model, the corresponding crop leaf area index is analyzed based on the processed visible light and red light images of the crop leaf area; in the vegetation index analysis layer of the image recognition model, the corresponding crop normalized vegetation index is analyzed based on the processed red light and near-infrared images of the crop leaf area; in the image color difference extraction layer of the image recognition model, the corresponding crop image color difference value is analyzed based on the processed visible light images of the crop leaf area; in the output layer of the image recognition model, the crop chlorophyll index, leaf area index, normalized vegetation index, and image color difference value are output, constituting the historical crop growth characteristics of the designated plot.

[0013] Furthermore, the nutrient crop absorption response set includes the crop absorption response rate and crop absorption lag length corresponding to several types of nutrient elements. The specific steps for analyzing and obtaining the nutrient crop absorption response set for a designated plot are as follows: Based on the time of each fertilization, multiple fixed time windows are set, and the changes in the values ​​of each image growth index within each time window compared to the day before fertilization are analyzed, and the cumulative increase value within the time window is calculated; based on the cumulative increase value of the image growth index corresponding to different fertilization types and dosages, the average amplitude of the image response of each nutrient element is calculated to obtain the crop absorption response rate of each nutrient element; and when the image growth index shows positive growth for two consecutive days after fertilization, the crop absorption lag length of each nutrient element is determined in conjunction with the corresponding fertilization time.

[0014] Further, the specific steps for determining the recommended fertilization dose set for the designated plot are as follows: Read the target supplementation amount, crop absorption response rate, and crop absorption lag length for each nutrient element in the designated plot, and perform a comprehensive analysis to obtain the fertilization response index for the corresponding nutrient element; match the fertilization response index of each nutrient element with several preset recommended fertilization dose intervals, with each recommended fertilization dose interval corresponding to a recommended fertilization dose value; combine the recommended fertilization dose values ​​of all nutrient elements to form the recommended fertilization dose set for the designated plot.

[0015] An intelligent and precise decision-making system for returning organic fertilizer to the field includes: a current data acquisition unit for acquiring current soil testing data, current crop variety, and target crop yield for a designated plot; a nutrient supplementation analysis unit for analyzing the nutrient supplementation set for the designated plot based on the current soil testing data and target crop yield; a historical data acquisition unit for acquiring historical fertilization records and historical crop growth image data for the designated plot; a growth feature extraction unit for inputting historical crop growth image data into a pre-trained image recognition model to extract historical crop growth features for the designated plot; a crop absorption response analysis unit for analyzing the historical crop growth features in conjunction with historical fertilization records to obtain the nutrient crop absorption response set for the designated plot; a fertilizer recommendation dosage determination unit for determining the fertilizer recommendation dosage set for the designated plot based on the nutrient supplementation set and the nutrient crop absorption response set; and a fertilizer application operation control unit for controlling designated machinery to perform organic fertilizer application operations on the designated plot according to the fertilizer recommendation dosage set.

[0016] The present invention has the following beneficial effects: (1) The intelligent and precise decision-making method for returning organic fertilizer to the field can accurately calculate the nutrient requirements of crops by dynamically analyzing soil, crop and historical fertilization data. By combining current soil testing data, crop varieties and target yields, specific target supplementation amounts can be formulated for each nutrient element. Combined with historical fertilization records and crop growth image data, the crop's absorption response to fertilizer at each stage can be deduced. This method is based on the comprehensive analysis of crop growth characteristics and historical data, and can dynamically adjust the recommended fertilization dosage rather than relying on a fixed fertilization plan, thereby effectively avoiding the problem of excessive or insufficient fertilizer. The calculation of the recommended fertilization dosage not only considers the immediate needs of crop growth, but also fully considers the effective supply of soil and the absorption lag effect after fertilization. This precise fertilization method greatly improves fertilizer utilization efficiency, avoids fertilizer waste in traditional fertilization methods, and also avoids the negative impact of excessive fertilization on soil and environment.

[0017] (2) This intelligent and precise decision-making method for returning organic fertilizer to the field can provide personalized fertilization plans for each crop and each soil by comprehensively analyzing the nutrient replenishment set and crop absorption response set of the set plot. Traditional fertilization methods usually rely on general fertilization standards and ignore the differences in soil fertility, different crop growth stages and nutrient absorption response after fertilization. This may result in some areas having excessive nutrients and others having insufficient nutrient supply, thereby affecting crop growth and yield. Based on the target yield of crops, soil nutrient detection results and historical fertilization data, this invention dynamically assesses the nutritional needs of crops and proposes precise fertilization suggestions. By extracting the growth characteristics of crops through image recognition models, the nutrient absorption of crops at different growth stages can be obtained. Furthermore, the fertilization dosage can be customized according to different soil conditions and crop varieties. This personalized fertilization recommendation method effectively solves the problem of different fertilization needs under different crop and soil conditions, ensures that crops obtain the best nutrient supply, and improves the growth quality and production efficiency of crops.

[0018] (3) The intelligent and precise decision-making method for returning organic fertilizer to the field realizes the automation and precision of fertilization operation through intelligent control technology. Traditional fertilization methods usually require farmers to decide the amount and time of fertilization based on experience. This method is often limited by human resources and the precision of operation, resulting in unstable fertilization effect, or even over-fertilization or under-fertilization, thereby increasing agricultural management costs and labor intensity. This invention, through intelligent fertilization recommendation decision-making and automated control, can directly control the fertilization machinery to carry out fertilization operations on the target plot according to the fertilization recommendation dose set. The fertilization machinery can automatically perform operations according to the calculated fertilization recommendation dose and fertilization time window, reducing human operation errors and improving fertilization precision. Intelligent fertilization not only improves work efficiency, but also reduces the labor intensity of farmers and reduces labor costs in agricultural production. At the same time, the automated system can monitor the fertilization process in real time to ensure the precise control of fertilization dose, effectively avoiding over-fertilization and fertilizer waste, which helps to achieve sustainable agricultural development and improve the efficiency of farmland management.

[0019] (4) This intelligent and precise decision-making system for returning organic fertilizer to the field achieves intelligent, automated, and precise agricultural fertilization operations by highly integrating various functional modules. Traditional fertilization methods in the existing technology often rely on manual judgment of fertilization time, dosage, and method, which not only has a large human error, but also often faces problems such as low operation efficiency and difficulty in responding to crop growth needs in real time. Existing automatic fertilization equipment usually cannot dynamically adjust the fertilization plan based on soil, crop, and historical data, and it is also difficult to make flexible scheduling in the actual fertilization process, resulting in low fertilizer use efficiency and difficulty in achieving precision agriculture. This invention, through modular system design, integrates soil, crop, and historical data to achieve intelligent, automated, and precise agricultural fertilization operations. The system integrates functions such as soil data acquisition, crop growth image analysis, fertilization recommendation, and operation control into an intelligent system. Each unit is interconnected, and through data sharing and an automated control system, the system can adjust the fertilizer dosage, timing, and method in real time during the fertilization process to ensure precise fertilization for each plot. At the same time, the system can monitor the effects of fertilization in real time and automatically optimize the fertilization plan and operation scheduling based on feedback results to avoid over-fertilization or under-fertilization. Through this optimized scheduling, the efficiency of fertilization operations has been significantly improved, the fertilizer utilization rate in agricultural production has been greatly increased, and waste and errors in the fertilization process have been reduced.

[0020] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0021] Figure 1 This is a flowchart of an intelligent and precise decision-making method for returning organic fertilizer to the field according to the present invention.

[0022] Figure 2 This is a flowchart illustrating the specific steps involved in analyzing the nutrient replenishment set of a designated plot in an intelligent and precise decision-making method for returning organic fertilizer to the field, as described in this invention.

[0023] Figure 3 This is a block diagram of an intelligent and precise decision-making system for returning organic fertilizer to the field according to the present invention. Detailed Implementation

[0024] Please see Figure 1This invention provides a technical solution: an intelligent and precise decision-making method for returning organic fertilizer to the field, comprising the following steps: acquiring current soil testing data, current crop variety, and target crop yield of a designated plot; analyzing the nutrient replenishment set of the designated plot based on the current soil testing data and target crop yield; acquiring historical fertilization records and historical crop growth image data of the designated plot; inputting the historical crop growth image data into a pre-trained image recognition model to extract historical crop growth characteristics of the designated plot; combining the historical crop growth characteristics with historical fertilization records to analyze and obtain the nutrient crop absorption response set of the designated plot; determining the recommended fertilization dose set of the designated plot based on the nutrient replenishment set and the nutrient crop absorption response set; and controlling the designated machinery to perform organic fertilizer application operations on the designated plot according to the recommended fertilization dose set.

[0025] Specifically, current soil testing data includes the actual content values ​​of several types of nutrient elements, while the nutrient supplement set includes the target supplement amount of several types of nutrient elements.

[0026] like Figure 2 As shown, the specific steps for analyzing the nutrient replenishment set of a designated plot are as follows: Obtain the fertilizer requirement parameters per unit yield of the current crop in the designated plot, and analyze the target total fertilizer requirement for each nutrient element in conjunction with the current crop target yield. The fertilizer requirement parameters per unit yield of the current crop include the unit fertilizer requirement of each nutrient element corresponding to the current crop unit yield. Obtain the soil effective supply coefficients of each nutrient element in the designated plot, and analyze the absorbable supply of each nutrient element in conjunction with the corresponding actual content values. Subtract the target total fertilizer requirement of each nutrient element from the corresponding absorbable supply (i.e., target total fertilizer requirement minus absorbable supply) to obtain the target replenishment amount (non-negative) of each nutrient element.

[0027] The specific formulas for calculating the target total fertilizer requirement and absorbable supply of a certain type of nutrient element are as follows: ;in, , , , , , The values ​​are, in order, the target total fertilizer requirement, unit fertilizer requirement, current crop target yield, absorbable supply, soil effective supply coefficient, and actual content value for a certain type of nutrient element.

[0028] The system retrieves the unit fertilizer requirements of various nutrients corresponding to the current crop yield per unit area through a built-in crop fertilizer requirement parameter database. This database is indexed by crop type and stores standard unit fertilizer requirements for common crops such as corn, wheat, rice, and soybeans under different ecological zones and planting conditions. The database can be constructed based on authoritative data released by agricultural research institutions and experimental field measurement results. During the analysis process, the system first automatically matches the crop variety of the set plot with the current target yield, extracting the corresponding unit fertilizer requirement parameter set. If no corresponding crop variety is found in the database, the system can provide a manual input interface for users to manually set the unit fertilizer requirement parameters, ensuring adaptability to special or new crop scenarios. Furthermore, in application scenarios supported by crop growth monitoring data, the system can also dynamically estimate unit fertilizer requirements parameters through a historical fertilization and yield data backtracking analysis model, improving the model's accuracy and robustness.

[0029] The soil available supply coefficient for various nutrient elements represents the proportion of measured nutrient content in the soil that can be effectively absorbed and utilized by crops. Its acquisition methods include static matching and dynamic evaluation. The system has a built-in supply coefficient parameter table by default, automatically matching recommended default supply coefficient values ​​based on the soil type (e.g., sandy loam, loam, clay), organic matter content level, and historical regional experimental data of the set plot. In application scenarios with soil sensor deployment capabilities, the system can read real-time soil status data (including pH, moisture, electrical conductivity, and estimated organic matter values) of the set plot and dynamically correct it using the embedded supply coefficient evaluation model to generate personalized supply coefficient results for the current plot. Furthermore, the system also allows users to manually adjust the supply coefficient values ​​in the parameter interface for manual correction of specific empirical rules or field test results.

[0030] This implementation plan provides precise fertilization recommendations by dynamically acquiring current soil data and target crop yields, combined with historical fertilization records and soil effective supply coefficients. First, the system's built-in crop nutrient requirement parameter database ensures that the fertilization amount for different crops in different ecological zones and planting conditions is set scientifically and reasonably. At the same time, the system supports a manual input interface, ensuring flexibility even for special or new crops. Second, the dynamic evaluation of the soil effective supply coefficient allows the system to make personalized adjustments based on soil type, real-time status data, and historical data, improving the accuracy of fertilization and avoiding the one-size-fits-all approach of traditional methods.

[0031] Specifically, historical fertilization records include the type, dosage, and timing of each fertilization operation during the crop's historical growth cycle. Historical crop growth image data consists of visible light, red light, and near-infrared images of the crop leaf area before and several days after each fertilization operation during the crop's historical growth cycle.

[0032] Among them, the visible light image, red light image and near-infrared image were taken once a day before each fertilization operation and for several days after each fertilization operation.

[0033] A visible light image consists of several RGB pixels, each corresponding to a red channel pixel value, a green channel pixel value, and a blue channel pixel value. A red light image consists of several single-channel red light pixels, each corresponding to a red light intensity value. A near-infrared image consists of several single-channel infrared pixels, each corresponding to a near-infrared reflection intensity value.

[0034] Among them, the visible light image, red light image, and near-infrared image are imaging results of the same leaf area in different bands. Therefore, the number of RGB pixels, the number of single-channel red light pixels, and the number of single-channel infrared pixels are the same.

[0035] The image recognition model includes an image input processing layer, a chlorophyll analysis layer, a leaf area analysis layer, a vegetation index analysis layer, an image color difference extraction layer, and an output layer. Historical crop growth characteristics include crop chlorophyll index, crop leaf area index, crop normalized vegetation index, and crop image color difference value.

[0036] The specific steps for extracting the historical crop growth characteristics of a designated plot are as follows: In the image input processing layer of the image recognition model, the visible light, red light, and near-infrared images of the crop leaf area of ​​the designated plot before and after each fertilization operation are uniformly processed. Specifically, the red channel pixel value, green channel pixel value, and blue channel pixel value of each RGB pixel in the visible light image of each day, the red light intensity value of each single-channel red light pixel in the red light image, and the infrared reflection intensity value of each single-channel infrared pixel in the near-infrared image are read. The pixel values ​​of each channel are linearly normalized to the [0,1] interval, and pixel registration and size unification are performed under the same spatial coordinates to ensure that the number of pixels and coordinate correspondence of each channel image are consistent. Subsequently, the images of the days before and after each fertilization are processed day by day and the indicators are extracted. Then, the results within the corresponding time period are averaged. In the chlorophyll analysis layer of the image recognition model, based on the visible light image of the processed crop leaf area, the corresponding crop chlorophyll index is analyzed. Specifically, the chlorophyll index is calculated for each image taken several days before and after fertilization, and the average of the index values ​​for each day is used as the crop chlorophyll index for that period. The calculation method for the daily chlorophyll index is as follows: The crop chlorophyll index is obtained by iterating through all pixels in the visible light image, calculating the relative dominance of the green channel relative to the red and blue channels for each pixel, and averaging the values ​​across all pixels in the entire leaf region. The calculation formula is as follows: ;in, To determine the crop chlorophyll index for the plot, , , The following are the normalized visible light images of crop leaf regions: The green channel pixel value, red channel pixel value, and blue channel pixel value of each RGB pixel. The red-blue interference correction coefficients preset in the database take values ​​in the range of [0.3, 0.5] in this embodiment. =1, 2, 3, ... , The number of RGB pixels in the visible light image; In the leaf area analysis layer of the image recognition model, the corresponding crop leaf area index is analyzed based on the visible light image and red light image of the processed crop leaf area. Specifically, in the previous image input processing layer, the geometric registration and size unification of the visible light image and the red light image have been completed, so that each pixel point of the two in the same spatial coordinates corresponds one-to-one, representing the light reflection intensity of the same leaf position in different wavelengths.

[0037] Based on this, the leaf area analysis layer performs leaf region extraction and area index calculation for each image taken before and several days after fertilization, and averages the results for each day. The calculation method for each day is as follows: Brightness features are compared for each pixel in the image. For each pixel, its brightness value in the visible light image and its reflectance value in the red light image are read. The ratio of these two values ​​can be used to determine whether the pixel belongs to the leaf region. Since leaves reflect red light weakly and visible light strongly, when the visible light brightness value of a pixel is significantly higher than its corresponding red light reflectance value, the pixel can be determined to belong to the effective area of ​​the leaf.

[0038] The system uses a leaf recognition threshold to distinguish leaf pixels from non-leaf pixels. Pixels with a ratio higher than the threshold are identified as "leaf pixels," while those with a ratio lower than or equal to the threshold are identified as "background or shadow pixels." This recognition process is performed cyclically across all pixels in the entire image, generating a binary distribution of the leaf region.

[0039] Subsequently, the leaf area analysis layer counts the number of pixels identified as "leaf pixels" and calculates the ratio of this number to the total number of pixels in the entire image to obtain the proportion of the leaf area in the entire image. This proportion is the leaf area index. In the vegetation index analysis layer of the image recognition model, based on the processed red light and near-infrared images of crop leaf areas, the corresponding normalized vegetation index (NVI) of the crop is analyzed. Specifically, the NVI of the crop is calculated for each image several days before and several days after fertilization, and the NVI of the crop within this time period is averaged to obtain the final NVI of the crop. The calculation method for each day is as follows: In the previous image input processing layer, the geometric registration and size unification of the red light image and the near-infrared image have been completed, ensuring that each pixel in the two images corresponds one-to-one with the other in the same spatial coordinates. This represents the red light intensity value and near-infrared reflectance value of the same leaf position in different wavelengths. Therefore, it can be regarded as a single image containing several pixels, each corresponding to a red light intensity value and a near-infrared reflectance value. The normalized difference between the near-infrared reflectance and red light intensity is calculated for each pixel in the image to obtain the pixel-level NDVI value. The average value is then calculated over all pixels in the entire image to form the crop normalized vegetation index, which is calculated using the following formula: ;in, This represents the normalized difference corresponding to a certain pixel. , These are, in order, the near-infrared reflectance intensity value and the red light intensity value corresponding to a certain pixel. This is a normalization adjustment factor stored in the database to prevent the denominator from being zero. In the image color difference extraction layer of the image recognition model, based on the visible light image of the processed crop leaf area, the corresponding crop image color difference value is analyzed. Specifically, the color difference is calculated for images from several days before and after fertilization, and the average color difference value of all images within the same time period is calculated. The color difference value of a single image is calculated by statistically analyzing the RGB channel average of all pixels within the leaf area and calculating the color deviation of each pixel. The root mean square of the deviation is used as the overall color difference value. The calculation formula is as follows: ;in, To set the color difference values ​​for crop images of a plot, , , The following are the normalized visible light images of crop leaf regions: The red channel pixel value, green channel pixel value, and blue channel pixel value of each RGB pixel. , , The values ​​shown are, in order, the mean pixel values ​​of the red channel, green channel, and blue channel in the normalized visible light image of the crop leaf region; In the output layer of the image recognition model, crop chlorophyll index, leaf area index, normalized vegetation index, and image color difference value are output, which constitute the historical crop growth characteristics of the set plot.

[0040] In this implementation plan, precise image recognition and data processing technologies are used to extract crop growth characteristics in real time, providing a scientific basis for precision fertilization. First, through multi-dimensional image data processing (including visible light images, red light images, and near-infrared images), the system can accurately extract key growth characteristics such as crop chlorophyll index, leaf area index, normalized vegetation index, and image color difference value. The extraction of these characteristics not only reflects the nutritional status of the crop but also dynamically adjusts fertilization recommendations to adapt to the nutrient needs of the crop at different growth stages. Second, through multi-time period analysis of historical crop growth image data (several days before and after fertilization), the system can dynamically assess the impact of fertilization on crop growth and provide more personalized and precise fertilization plans by comprehensively analyzing image indicators, avoiding the shortcomings of traditional fertilization based on a single data source. In addition, the use of image color difference extraction and other methods improves the accuracy of crop health monitoring, making fertilization decisions more scientific and flexible.

[0041] Specifically, the nutrient crop absorption response set includes the crop absorption response rate and crop absorption lag length corresponding to several types of nutrient elements. The specific steps for analyzing and obtaining the nutrient crop absorption response set of a designated plot are as follows: Based on the time of each fertilization, multiple fixed time windows are set (including the 3rd, 5th, and 7th days after fertilization). The changes in the values ​​of each image growth index (including crop chlorophyll index, leaf area index, normalized vegetation index, and image color difference value) within each time window compared to the day before fertilization are analyzed, and the cumulative improvement value within the time window is calculated. Specifically, for each fertilization operation, target crop images are acquired one day before fertilization and on the 3rd, 5th, and 7th days after fertilization. Based on a pre-trained image recognition model, the image growth index values ​​of the crop area in each image are extracted. The index value at each time point after fertilization is subtracted from the index value on the day before fertilization to obtain the change in the value of each index. The changes at all time points are then summed to obtain the cumulative improvement value within the time window. Based on the cumulative improvement values ​​of image growth indicators corresponding to different fertilization types and dosages, the average amplitude of image response for each nutrient element is calculated to obtain the crop absorption response rate of each nutrient element. Specifically, the fertilization type and corresponding dosage indicated in the historical fertilization records are correlated with the cumulative improvement value of image growth indicators after each fertilization. The cumulative improvement values ​​of multiple fertilization records involving the same nutrient element are averaged to obtain the average improvement amplitude of each image indicator under the action of that nutrient element. Then, the crop absorption response rate of each nutrient element is calculated. The response rate is defined as the ratio of the average improvement value of the image indicator to the fertilization dosage, which is used to measure the growth response intensity brought about by a unit fertilization dosage. When the image growth index shows positive growth for two consecutive days after fertilization, the lag time for crop absorption of each nutrient element is determined based on the corresponding fertilization time. Specifically, based on daily image acquisition, the daily changes in image indices are calculated continuously after fertilization. If the change value of a certain index is positive for two consecutive days after fertilization, the starting time of the change is recorded as the response start date. The time difference between the response start date and the fertilization date is calculated to obtain the crop absorption lag time corresponding to that nutrient element. If multiple nutrient elements are involved in the same fertilization, their respective lag times are calculated based on their dominant indices.

[0042] In this implementation scheme, by combining image recognition and fertilization data, dynamic monitoring and adjustment of fertilization effects are achieved. Specifically, the system analyzes image data before and after fertilization by setting multiple time windows, extracting growth indicators such as crop chlorophyll, leaf area, vegetation index, and color difference value, thereby accurately measuring crop growth changes after fertilization. By accumulating these changes, the system can evaluate the fertilization effect in real time and provide targeted absorption response rates for each nutrient element, avoiding the blindness and waste in traditional fertilization methods. In addition, this invention also provides a scientific basis for adjusting fertilization time and amount by analyzing the crop absorption lag time. The determination of the lag time can accurately reflect the crop's absorption response speed to fertilizer, further optimizing fertilization strategies and ensuring a high degree of matching between fertilization and crop needs. Through this method, agricultural production not only improves fertilizer utilization and reduces waste, but also avoids environmental pollution and promotes the development of precision agriculture.

[0043] Specifically, the steps for determining the recommended fertilization dose set for a designated plot are as follows: Read the target supplementation amount, crop absorption response rate, and crop absorption lag length for each nutrient element in the designated plot, and perform a comprehensive analysis to obtain the fertilization response index for the corresponding nutrient element; match the fertilization response index of each nutrient element with several preset recommended fertilization dose intervals, with each interval corresponding to a recommended fertilization dose value (each nutrient element corresponds to a type of organic fertilizer); combine the recommended fertilization dose values ​​of all nutrient elements to form the recommended fertilization dose set for the designated plot.

[0044] The specific formula for calculating the fertilization response index of a certain type of nutrient element is as follows: ;in, , , , The parameters are, in order, the fertilization response index, target supplementation amount, crop absorption response rate, and crop absorption lag length for a certain type of nutrient element. This refers to the hysteresis correction coefficient stored in the database.

[0045] The range of values ​​for the lag correction coefficient is as follows: Low hysteresis coefficient [0.1, 0.2]: It is suitable for nutrients that crops respond to quickly, or for soil nutrients that are released quickly.

[0046] In this situation, the crop's absorption response after fertilization is relatively rapid, and the lag time has little impact on the fertilizer dosage, thus resulting in a smaller coefficient.

[0047] Medium lag coefficient [0.2, 0.4): It is applicable to most conventional crops and nutrient elements.

[0048] This coefficient is appropriate, taking into account both the lag effect of crop absorption and maintaining the rationality of the recommended fertilizer dosage.

[0049] Higher lag coefficients [0.4, 0.5]: It is suitable for situations where crops respond slowly or soil fertility is low.

[0050] In this case, a relatively long lag period is required to see the effects of fertilization. Therefore, a larger correction factor is needed to increase the recommended fertilization dose to compensate for the effects of absorption lag.

[0051] In this implementation scheme, by calculating the fertilization response index, the fertilizer dosage can be precisely adjusted according to the characteristics of each nutrient element. First, by reading the target replenishment amount of the crop, the crop absorption response rate, and the lag time, combined with the lag correction coefficient, the fertilization response index of each nutrient element can be accurately calculated, thereby obtaining a scientific recommendation for the fertilizer dosage. The introduction of the lag correction coefficient allows the fertilization plan to be dynamically adjusted according to factors such as soil fertility and crop response speed, ensuring that fertilization is more in line with the actual needs of the crop. In addition, this invention makes the fertilization plan more targeted and adaptable to different crops and different soil conditions by setting the lag coefficient in three zones (low, medium, and high). The low lag coefficient is suitable for crops that respond quickly, avoiding over-fertilization; the medium lag coefficient is suitable for conventional crops, ensuring a balanced amount of fertilizer; and the high lag coefficient is suitable for crops that respond slowly or infertile soil, avoiding insufficient fertilization due to the lag effect. In this way, the system not only improves the efficiency of fertilizer use and reduces waste, but also optimizes crop yield through precise fertilization.

[0052] Please see Figure 3 This invention provides a technical solution: an intelligent and precise decision-making system for returning organic fertilizer to the field, comprising: a current data acquisition unit for acquiring current soil testing data, current crop variety, and target crop yield of a designated plot; a nutrient supplementation analysis unit for analyzing the nutrient supplementation set of the designated plot based on the current soil testing data and target crop yield; a historical data acquisition unit for acquiring historical fertilization records and historical crop growth image data of the designated plot; a growth feature extraction unit for inputting historical crop growth image data into a pre-trained image recognition model to extract historical crop growth features of the designated plot; a crop absorption response analysis unit for analyzing the historical crop growth features in conjunction with historical fertilization records to obtain the nutrient crop absorption response set of the designated plot; a fertilizer recommendation dosage determination unit for determining the fertilizer recommendation dosage set of the designated plot based on the nutrient supplementation set and the nutrient crop absorption response set; and a fertilizer application control unit for controlling designated machinery to perform organic fertilizer application operations on the designated plot according to the fertilizer recommendation dosage set.

[0053] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0054] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart and precise decision-making method for returning organic fertilizer to the field, characterized in that, Includes the following steps: Obtain the current soil testing data, current crop variety, and target crop yield for the designated plot. The current soil testing data includes the actual content values ​​of several types of nutrient elements, and the nutrient supplement set includes the target supplement amount of several types of nutrient elements. Based on current soil testing data and target crop yield, analyze the nutrient replenishment set for the designated plots; Acquire historical fertilization records and historical crop growth image data for the designated plot. The historical fertilization records include the fertilization type, dosage, and time of each fertilization operation during the crop's historical growth cycle. The historical crop growth image data consists of visible light, red light, and near-infrared images of the crop leaf area several days before and after each fertilization operation during the crop's historical growth cycle. Historical crop growth image data is input into a pre-trained image recognition model to extract historical crop growth characteristics of a designated plot of land. By combining historical crop growth characteristics with historical fertilization records, a nutrient crop absorption response set for a designated plot is obtained. Specifically, based on the time of each fertilization, multiple fixed time windows are set, and the changes in the values ​​of each image growth index within each time window compared to the day before fertilization are analyzed, and the cumulative increase value within the time window is calculated. Based on the cumulative increase value of the image growth index corresponding to different fertilization types and dosages, the average amplitude of the image response of each nutrient element is calculated to obtain the crop absorption response rate of each nutrient element. When the image growth index shows positive growth for two consecutive days after fertilization, the crop absorption lag time of each nutrient element is determined in combination with the corresponding fertilization time. The nutrient crop absorption response set is composed of the crop absorption response rate of each nutrient element and the crop absorption lag time. Based on the nutrient supplementation set and the nutrient crop absorption response set, a set of recommended fertilization doses for a designated plot is determined. Specifically, this involves reading the target supplementation amount, crop absorption response rate, and crop absorption lag time for each nutrient element in the designated plot, calculating the fertilization response index for the corresponding nutrient element, and performing matching analysis with several preset recommended fertilization dose intervals. Each recommended fertilization dose interval corresponds to a recommended fertilization dose value. The recommended fertilization dose values ​​for all nutrient elements are then combined to form the set of recommended fertilization doses for the designated plot. Based on the recommended fertilizer dosage set, control the set operating machinery to carry out organic fertilizer application operations on the set plots; The specific steps for analyzing the nutrient replenishment set of the designated plot are as follows: Obtain the fertilizer requirement parameters for the current crop unit yield of the set plot, and analyze the target total fertilizer requirement of various nutrient elements in combination with the current crop target yield. The fertilizer requirement parameters for the current crop unit yield include the unit fertilizer requirement of various nutrient elements corresponding to the current crop unit yield. Obtain the soil effective supply coefficients of various nutrient elements in the designated plot, and analyze the absorbable supply of various nutrient elements in combination with the corresponding actual content values. The target supplement amount of each nutrient element is obtained by subtracting the target total fertilizer requirement from the corresponding absorbable supply.

2. The intelligent and precise decision-making method for returning organic fertilizer to the field according to claim 1, characterized in that, A visible light image consists of several RGB pixels, each corresponding to a red channel pixel value, a green channel pixel value, and a blue channel pixel value. A red light image consists of several single-channel red light pixels, each corresponding to a red light intensity value. A near-infrared image consists of several single-channel infrared pixels, each corresponding to a near-infrared reflection intensity value.

3. The intelligent and precise decision-making method for returning organic fertilizer to the field according to claim 2, characterized in that, The image recognition model includes an image input processing layer, a chlorophyll analysis layer, a leaf area analysis layer, a vegetation index analysis layer, an image color difference extraction layer, and an output layer.

4. The intelligent and precise decision-making method for returning organic fertilizer to the field according to claim 3, characterized in that, Historical crop growth characteristics include crop chlorophyll index, crop leaf area index, crop normalized vegetation index, and crop image color difference value.

5. The intelligent and precise decision-making method for returning organic fertilizer to the field according to claim 4, characterized in that, The specific steps for extracting historical crop growth characteristics of a designated plot are as follows: In the image input processing layer of the image recognition model, the visible light image, red light image and near-infrared image of the crop leaf area of ​​the set plot are uniformly processed before and after each fertilization operation. In the chlorophyll analysis layer of the image recognition model, the corresponding crop chlorophyll index is analyzed based on the visible light image of the processed crop leaf region. In the leaf area analysis layer of the image recognition model, the corresponding crop leaf area index is analyzed based on the visible light image and red light image of the processed crop leaf region. In the vegetation index analysis layer of the image recognition model, the corresponding normalized vegetation index of crops is analyzed based on the red light image and near-infrared image of the processed crop leaf area. In the image color difference extraction layer of the image recognition model, the corresponding crop image color difference value is analyzed based on the visible light image of the processed crop leaf region; In the output layer of the image recognition model, crop chlorophyll index, leaf area index, normalized vegetation index, and image color difference value are output, which constitute the historical crop growth characteristics of the set plot.

6. An intelligent and precise decision-making system for returning organic fertilizer to the field, employing the intelligent and precise decision-making method for returning organic fertilizer to the field as described in any one of claims 1-5, characterized in that, include: The current data acquisition unit is used to acquire the current soil testing data, current crop variety, and target crop yield of the designated plot. The nutrient supplementation analysis unit is used to analyze the nutrient supplementation set of a designated plot based on current soil testing data and target crop yield. The historical data acquisition unit is used to acquire historical fertilization records and historical crop growth image data for a designated plot of land. The growth feature extraction unit is used to input historical crop growth image data into a pre-trained image recognition model to extract historical crop growth features of a designated plot. The crop absorption response analysis unit is used to analyze historical crop growth characteristics and combine them with historical fertilization records to obtain the nutrient crop absorption response set for a given plot. The fertilizer recommendation dosage determination unit is used to determine the fertilizer recommendation dosage set for a given plot based on the nutrient replenishment set and the nutrient crop absorption response set. The fertilization operation control unit is used to control the set operation machinery to perform organic fertilizer application operations on the set plots according to the recommended fertilizer dosage set.

Images