Soil fertility prediction system and method

Abstract

The invention discloses a soil fertility prediction system and method, and the system comprises a data collection module, a data prediction module, a decision execution module and a result display module, the data prediction module analyzes a collected image, and builds a fertility prediction model in combination with image data and fertility indexes, so as to achieve the prediction of a fertilization strategy of a detection region. The invention relates to the technical field of data analysis and prediction. According to the soil fertility prediction system and method, collected images are analyzed, planting area features are determined, planting parameters are extracted from a storage database to achieve image segmentation and determine sampling points, and fertility indexes of all collection points are calculated in combination with collected soil fertility data; image data and fertility indexes are combined to establish a fertility prediction model to realize fertilization strategy prediction of the detection area, so that soil fertility prediction is effectively and efficiently realized, the model is synchronously established to realize accurate prediction operation, and high precision of the fertilization strategy in the area is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and prediction, and in particular to a soil fertility prediction system and method. Background Art

[0002] Soil fertility is the ability of soil to provide and coordinate nutrients and environmental conditions for plant growth. It is a comprehensive reflection of various basic soil properties and has a crucial impact on crop yield and quality. Accurately predicting soil fertility status helps optimize fertilization, improve fertilizer utilization, reduce environmental pollution, and achieve sustainable agricultural development.

[0003] The reference patent is titled: A Soil Fertility Prediction Method and System, Electronic Device, and Storage Medium (Patent Publication Number: CN116562134A, Patent Publication Date: 2023-08-08). The method includes: determining the initial sampling point and central sampling point of the target plot; constructing a sampling point sequence for each initial sampling point; determining preliminary average soil fertility data and preliminary soil fertility change data for each sampling point sequence based on the fertility index of the soil sample corresponding to the top-ranked sampling point; performing fertility prediction on soil samples corresponding to sampling points not ranked top; updating the preliminary average soil fertility data and preliminary soil fertility change data based on the prediction results to obtain target average soil fertility data and target soil fertility change data; and obtaining soil fertility prediction data for the target plot based on the target average soil fertility data of all sampling point sequences.

[0004] Based on the description in the above-mentioned document, in the existing soil fertility prediction process, the areas involved often include planting areas and non-planting areas, and there is a large gap in the fertility data of the two. In addition, inaccurate positioning of the sampling points and the depth of soil collection will lead to deviations in the final fertility prediction. In addition, in the current process of fertility prediction, the fertility index in the region has a large deviation, and the planned fertilization amount is uneven. For this reason, the present invention provides a soil fertility prediction system and method. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a soil fertility prediction system and method, which solves the problem that in the existing soil fertility prediction process, the areas involved often include planting areas and non-planting areas, and there is a large gap in the fertility data of the two. In addition, inaccurate positioning of the sampling points and the depth of soil collection will lead to deviations in the final fertility prediction. In addition, in the current process of fertility prediction, the fertility index in the region has a large deviation, and the planned fertilization amount is uneven.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A soil fertility prediction system, comprising: The data acquisition module uses the image acquisition module carried by the drone equipment to locate the area to be detected. After positioning is completed, the image acquisition operation of the area to be detected is performed, and the soil fertility data in the image is collected by the sensor acquisition module carried by the drone equipment. The collected data is transmitted and stored in the storage database; The data prediction module analyzes the collected images and determines the characteristics of the planting area. It extracts planting parameters from the storage database to implement image segmentation and determine the sampling points. It calculates the fertility index of each collection point based on the collected soil fertility data. It then establishes a fertility prediction model based on the image data and fertility index to predict the fertilization strategy for the detection area. The decision-making execution module formulates execution strategies based on the prediction results and cooperates with the fertilization equipment to complete fertilization operations in areas with different fertility levels; The result display module displays the result data obtained from the analysis and processing in the form of text and charts.

[0007] Preferably, the operation of performing image acquisition in the data acquisition module is: Determine the boundaries of the area to be inspected, set the drone acquisition height, and the actual size of the image at the current height is A1 and A2. Plan the drone's flight path to achieve image acquisition. Set the acquisition starting point of the detection area, and the starting point is located on the horizontal boundary of the detection area parallel to the horizontal direction of the image, and the distance between the starting point and the left vertical boundary of the detection area is less than A2 / 2; Set the end point of the detection area, and the end point is located on the horizontal boundary of the detection area or the extension line of the horizontal boundary, and the distance between the end point and the right longitudinal boundary of the detection area is less than A2 / 2; The UAV device's movement path is as follows: first, it moves along the longitudinal boundary of the detection area, capturing images every time it moves a distance B1, and when it reaches the rear lateral boundary of the detection area shown in the image, it has moved a distance B2, thereby achieving an S-shaped path acquisition operation. After the acquisition is completed, the image data is spliced.

[0008] Preferably, the operation of stitching the image data is: The image at the starting point is used as a reference, and the direction of the path is combined to perform a segmentation operation on the image at the starting point. The segmentation line is set at (A1-B1) from the rear lateral boundary of the image at the starting point. The segmentation operation is then performed, and the image with the larger area after segmentation is retained. The adjacent images in the path direction are spliced ​​with the retained image. And stitching is carried out in sequence until the image data has the horizontal boundary on the rear side of the detection area. The combined path segmentation line is set at (A2-B2) away from the vertical boundary on the right side of the current image, and then the segmentation operation is implemented, and the image with a larger area after segmentation is retained. The image stitching is continued in the combined path until the image stitching at the end point is completed to form a complete image data.

[0009] Preferably, the data prediction module performs the following operations to analyze the collected images and determine the characteristics of the planting area: By performing a grayscale operation on the complete image data, a plurality of comparison points are set equidistantly between the midpoints of the boundary lines of the relative boundaries of the image data; The direction of the planting area feature is determined according to the change in the grayscale value of the comparison point, and then the feature interval line is obtained according to the different grayscale values ​​of the planting area feature and the separating gully feature. The midpoints of the adjacent comparison points where the grayscale value changes occur are connected in sequence as the interval line, and then the planting area feature and the separating gully feature are distinguished.

[0010] Preferably, the operation of extracting the planting parameters from the storage database to implement image segmentation and determine the sampling points in the data prediction module is: The spacing distance between plants is determined to be M based on the planting parameters, and the planting area feature is divided into multiple grid areas at equal distances of M based on the spacing lines between the two sides of the planting area feature and the gully feature, and the position of the grid areas is marked in sequence; Randomly select any grid area as the basic sampling area, and connect the diagonal points of the current basic sampling area. The intersection generated by the intersection is the basic sampling point. Sampling points are set after being separated by a grid area on all four sides of the basic sampling point, and all sampling points are set according to the current method.

[0011] Preferably, the data prediction module calculates the fertility index of each collection point in combination with the collected soil fertility data as follows: Extract planting parameters to determine the pre-buried depth of plant seeds, that is, to achieve the pre-buried depth of the equipment's sensor directly below the collection point when collecting soil fertility; Set the fertility data value of a collection point as P i , P i It represents the detection value P of the i-th fertility data category, and performs a weighted operation on each fertility data category, and the weighted sum is the fertility index. The specific formula is: ; Q j is the fertility index Q at the jth sampling point, W i is the weighted value W of the i-th fertility data category, and n represents the total number of fertility data categories.

[0012] Preferably, the data prediction module combines the image data and the fertility index to establish a prediction model to predict the fertilization strategy for the detection area as follows: A fertility prediction model for the current detection area is constructed by combining the characteristics of the planting area and the fertility index of each sampling point. The fertility prediction model is used to calculate the deviation values ​​of the fertility index of the sampling points with respect to the horizontal and vertical planting area characteristics in the detection area. The fertilization direction strategy in the planting area is determined based on the deviation values. Then, the planting parameters are extracted to determine the fertility interval requirement of the plant seeds as [F1, F2], and the average fertility index of the planting area is compared with the fertility interval requirement to determine the fertilization strategy; The predicted fertilization strategy results are obtained by combining the fertilization direction strategy and the fertilization amount strategy for transmission.

[0013] Preferably, the calculation operation of the deviation value is: The numerical calculation formula for the deviation of the horizontal planting area characteristics is: ; And Q m is the fertility index at the mth sampling point located in the horizontal planting area, Q c is the average fertility index of all sampling points in the horizontal planting area, and Q c =(Q1+Q2+…+Q m ) / m; The calculation formula for the deviation of the vertical planting area characteristics is: ; and is the fertility index at the 2mk+1 sampling point of the longitudinal planting area characteristics, Q d is the average fertility index of all sampling points in the longitudinal planting area characteristics, and ; The deviation values ​​of the horizontal planting area characteristics and the vertical planting area characteristics are compared, and if S1 < S2, the direction of the horizontal planting area characteristics is used as the predicted fertilization direction, otherwise the direction of the vertical planting area characteristics is used as the predicted fertilization direction.

[0014] Preferably, the operation of comparing the average fertility index of the planting area with the fertility interval requirement is: If Q c ∈[F1, F2] or Q d ∈[F1, F2], it is predicted that the fertility index of the current planting area meets the fertility range requirements; If Q c >[F1, F2] or Qd > [F1, F2], the amount of fertilizer needs to be reduced; If Q c <[F1, F2] or Q d <[F1, F2], it is predicted that the fertility index of the current planting area is insufficient and the amount of fertilizer needs to be increased.

[0015] The present invention also discloses a soil fertility prediction method, which specifically comprises the following steps: Step 1: Use drone equipment to collect images of the test area, determine the characteristics of the planting area to be sampled, and determine the sampling points to implement soil fertility testing; Step 2: Based on the image data and fertility detection data, a fertility prediction model is established to predict the fertilization strategy for the detection area; Step 3: Implement the execution operations of the planning direction and planning usage based on the prediction results.

[0016] The present invention provides a soil fertility prediction system and method. Compared with the existing technology, it has the following advantages: 1. The soil fertility prediction system and method analyzes the collected images and determines the characteristics of the planting area, extracts planting parameters from the storage database to achieve image segmentation and determine the sampling points, calculates the fertility index of each collection point based on the collected soil fertility data, and establishes a fertility prediction model based on the image data and fertility index to predict the fertilization strategy for the detection area. In this way, soil fertility can be predicted effectively and efficiently, and a model can be established simultaneously to achieve accurate prediction operations, ensuring the high accuracy of the fertilization strategy in the area.

[0017] 2. This soil fertility prediction system and method, by determining the boundaries of the area to be detected, setting the path for image acquisition in the detection area, and performing splicing processing on the image data after acquisition, can not only improve the efficiency of data acquisition, but also ensure the accuracy of the content features in the image, so as to better determine the characteristics of the planting area and then determine the sampling points, providing a basis for achieving the accuracy of subsequent prediction operations, and reducing the problem of area omissions, so as to better complete soil fertility detection and prediction.

[0018] 3. The soil fertility prediction system and method extracts planting parameters to determine the spacing distance that needs to be maintained between planted plants. Based on the spacing lines on both sides of the planting area features and the separating gully features, the planting area features are divided into multiple grid areas at equal distances, and the positions of the grid areas are marked in sequence. The intersection points of the diagonal points of the current basic sampling area are connected and the intersections are the basic sampling points, thereby ensuring that the sampling points are the subsequent planting points, which can ensure the accuracy of data positioning and improve the fault tolerance rate.

[0019] 4. The soil fertility prediction system and method construct a fertility prediction model for the current detection area by combining the characteristics of the planting area and the fertility index of each sampling point. The fertility prediction model is used to calculate the fertility index deviation values ​​of the sampling points of the horizontal planting area characteristics and the vertical planting area characteristics in the detection area. The fertilization direction strategy in the planting area is determined based on the deviation value, and the average fertility index in the planting area is compared with the fertility range requirement to determine the fertilization strategy. In this way, the fertility situation in the area is determined by combining the fertility influence of multiple categories of factors, reducing deviations while forming a fertilization strategy to complete efficient prediction and planning operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a principle block diagram of the soil fertility prediction system of the present invention; Figure 2 The present invention is a flow chart of the soil fertility prediction method. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] See also Figure 1-Figure 2 , the present invention provides three technical solutions: Embodiment 1: A soil fertility prediction system, comprising: The data acquisition module uses the image acquisition module carried by the drone equipment to locate the area to be detected. After positioning is completed, the image acquisition operation of the area to be detected is performed, and the soil fertility data in the image is collected by the sensor acquisition module carried by the drone equipment. The collected data is transmitted and stored in the storage database; The data prediction module analyzes the collected images and determines the characteristics of the planting area. It extracts planting parameters from the storage database to implement image segmentation and determine the sampling points. It calculates the fertility index of each collection point based on the collected soil fertility data. It then establishes a fertility prediction model based on the image data and fertility index to predict the fertilization strategy for the detection area. The decision-making execution module formulates execution strategies based on the prediction results and cooperates with the fertilization equipment to complete fertilization operations in areas with different fertility levels; The result display module displays the result data obtained from the analysis and processing in the form of text and charts.

[0023] Among them, by analyzing the collected images and determining the characteristics of the planting area, the planting parameters are extracted from the storage database to achieve image segmentation and determine the sampling points, and the fertility index of each collection point is calculated based on the collected soil fertility data. The fertility prediction model is established based on the image data and the fertility index to predict the fertilization strategy of the detection area. In this way, the soil fertility can be predicted effectively and efficiently, and the model is established simultaneously to achieve accurate prediction operations, ensuring the high accuracy of the fertilization strategy in the area.

[0024] In the embodiment of the present invention, the operations for image acquisition in the data acquisition module are: Determine the boundaries of the area to be inspected, set the drone acquisition height, and the actual size of the image at the current height is A1 and A2. Plan the drone's flight path to achieve image acquisition. Set the acquisition starting point of the detection area, and the starting point is located on the horizontal boundary of the detection area parallel to the horizontal direction of the image, and the distance between the starting point and the left vertical boundary of the detection area is less than A2 / 2; Set the end point of the detection area, and the end point is located on the horizontal boundary of the detection area or the extension line of the horizontal boundary, and the distance between the end point and the right longitudinal boundary of the detection area is less than A2 / 2; The UAV device's movement path is as follows: first, it moves along the longitudinal boundary of the detection area, capturing images every time it moves a distance B1, and when it reaches the rear lateral boundary of the detection area shown in the image, it has moved a distance B2, thereby achieving an S-shaped path acquisition operation. After the acquisition is completed, the image data is spliced.

[0025] In the embodiment of the present invention, the operations for splicing image data are as follows: The image at the starting point is used as a reference, and the direction of the path is combined to perform a segmentation operation on the image at the starting point. The segmentation line is set at (A1-B1) from the rear lateral boundary of the image at the starting point. The segmentation operation is then performed, and the image with the larger area after segmentation is retained. The adjacent images in the path direction are spliced ​​with the retained image. And stitching is carried out in sequence until the image data has the horizontal boundary on the rear side of the detection area. The combined path segmentation line is set at (A2-B2) away from the vertical boundary on the right side of the current image, and then the segmentation operation is implemented, and the image with a larger area after segmentation is retained. The image stitching is continued in the combined path until the image stitching at the end point is completed to form a complete image data.

[0026] Among them, by determining the boundaries of the area to be detected, setting the path for image acquisition in the detection area, and performing splicing processing on the image data after acquisition, not only the efficiency of data acquisition can be improved, but also the accuracy of the content features in the image can be guaranteed, so as to better determine the characteristics of the planting area and then realize the determination of the sampling points, provide a basis for achieving the accuracy of subsequent prediction operations, reduce the problem of regional omissions, and better complete soil fertility detection and prediction.

[0027] In the embodiment of the present invention, the data prediction module performs the following operations to analyze the collected images and determine the characteristics of the planting area: By performing a grayscale operation on the complete image data, a plurality of comparison points are set equidistantly between the midpoints of the boundary lines of the relative boundaries of the image data; The direction of the planting area feature is determined according to the change in the grayscale value of the comparison point, and then the feature interval line is obtained according to the different grayscale values ​​of the planting area feature and the separating gully feature. The midpoints of the adjacent comparison points where the grayscale value changes occur are connected in sequence as the interval line, and then the planting area feature and the separating gully feature are distinguished.

[0028] The grayscale value is obtained by graying the image, which is an existing processing technology. However, when dealing with different features, the values ​​are different. For example, the grayscale values ​​of the planting area feature and the separating gully feature are different due to inconsistencies in height and shape. If the grayscale values ​​at the set comparison points are almost similar, it means that the current planting direction is the horizontal direction connecting the midpoints of the current boundary lines.

[0029] In the embodiment of the present invention, the operation of extracting the planting parameters from the storage database to implement image segmentation and determine the sampling points in the data prediction module is as follows: The spacing distance between plants is determined to be M based on the planting parameters, and the planting area feature is divided into multiple grid areas at equal distances of M based on the spacing lines between the two sides of the planting area feature and the gully feature, and the position of the grid areas is marked in sequence; Randomly select any grid area as the basic sampling area, and connect the diagonal points of the current basic sampling area. The intersection generated by the intersection is the basic sampling point. Sampling points are set after being separated by a grid area on all four sides of the basic sampling point, and all sampling points are set according to the current method.

[0030] Among them, the spacing distance that needs to be retained between the planted plants is determined by extracting the planting parameters. Based on the spacing lines on both sides of the planting area features and the separating gully features, the planting area features are divided into multiple grid areas at equal distances, and the position markings of the grid areas are realized in sequence. The intersection points generated by connecting the diagonal points of the current basic sampling area are the basic sampling points, thereby ensuring that the sampling points are the subsequent planting points, which can ensure the accuracy of data positioning and improve the fault tolerance rate.

[0031] In the embodiment of the present invention, the data prediction module calculates the fertility index of each collection point in combination with the collected soil fertility data as follows: Extract planting parameters to determine the pre-buried depth of plant seeds, that is, to achieve the pre-buried depth of the equipment's sensor directly below the collection point when collecting soil fertility; Set the fertility data value of a collection point as P i , P i It represents the detection value P of the i-th fertility data category, and performs a weighted operation on each fertility data category, and the weighted sum is the fertility index. The specific formula is: ; Q j is the fertility index Q at the jth sampling point, W i is the weighted value W of the i-th fertility data category, and n represents the total number of fertility data categories.

[0032] In the embodiment of the present invention, the data prediction module combines the image data and the fertility index to establish a prediction model to predict the fertilization strategy of the detection area. The operation is as follows: A fertility prediction model for the current detection area is constructed by combining the characteristics of the planting area and the fertility index of each sampling point. The fertility prediction model is used to calculate the deviation values ​​of the fertility index of the sampling points with respect to the horizontal and vertical planting area characteristics in the detection area. The fertilization direction strategy in the planting area is determined based on the deviation values. Then, the planting parameters are extracted to determine the fertility interval requirement of the plant seeds as [F1, F2], and the average fertility index of the planting area is compared with the fertility interval requirement to determine the fertilization strategy; The predicted fertilization strategy results are obtained by combining the fertilization direction strategy and the fertilization amount strategy for transmission.

[0033] Among them, a fertility prediction model of the current detection area is constructed by combining the characteristics of the planting area and the fertility index of each sampling point. The fertility prediction model is used to calculate the fertility index deviation values ​​of the sampling points of the horizontal planting area characteristics and the vertical planting area characteristics in the detection area. The fertilization direction strategy in the planting area is determined according to the deviation value, and the average fertility index in the planting area is compared with the fertility interval requirement to determine the fertilization strategy. In this way, the fertility situation in the area is determined by combining the fertility influence of multiple categories of factors, reducing the deviation while forming a fertilization strategy to complete efficient prediction and planning operations.

[0034] In the embodiment of the present invention, the calculation operation of the deviation value is: The numerical calculation formula for the deviation of the horizontal planting area characteristics is: ; And Q m is the fertility index at the mth sampling point located in the horizontal planting area, Q c is the average fertility index of all sampling points in the horizontal planting area, and Q c =(Q1+Q2+…+Q m ) / m; The calculation formula for the deviation of the vertical planting area characteristics is: ; and is the fertility index at the 2mk+1 sampling point of the longitudinal planting area characteristics, Q d is the average fertility index of all sampling points in the longitudinal planting area characteristics, and ; The deviation values ​​of the horizontal planting area characteristics and the vertical planting area characteristics are compared, and if S1 < S2, the direction of the horizontal planting area characteristics is used as the predicted fertilization direction, otherwise the direction of the vertical planting area characteristics is used as the predicted fertilization direction.

[0035] In the embodiment of the present invention, the operation of comparing the average fertility index of the planting area with the fertility interval requirement is: If Q c ∈[F1, F2] or Q d ∈[F1, F2], it is predicted that the fertility index of the current planting area meets the fertility interval requirements, and the adaptive fertilization operation is implemented; If Q c >[F1, F2] or Q d > [F1, F2], it is necessary to reduce the amount of fertilizer and achieve uniform fertility through ditches; If Q c <[F1, F2] or Q d<[F1, F2], it is predicted that the fertility index of the current planting area is insufficient and the amount of fertilizer needs to be increased.

[0036] The difference between the second embodiment and the first embodiment is that the present invention further discloses a soil fertility prediction method, which specifically includes the following steps: Step 1: Use drone equipment to collect images of the test area, determine the characteristics of the planting area to be sampled, and determine the sampling points to implement soil fertility testing; Step 2: Based on the image data and fertility detection data, a fertility prediction model is established to predict the fertilization strategy for the detection area; Step 3: Implement the execution operations of the planning direction and planning usage based on the prediction results.

[0037] The third embodiment is different from the first and second embodiments in that a practical case is also disclosed, in which a detection operation is performed on an area where the soil fertility is known, and the original collected data parameters are saved. The operation is performed through the application method of the present invention, and the data results are recorded for comparison; The soil fertility standards are set as shown in Table 1: Table 1 Classification fertility table

[0038] The corresponding collected data are shown in Table 2: Table 2 Numerical table

[0039] The soil fertility index was calculated to be 0.757, which was found to be medium to high fertility after comparison with the fertility classification table. Synchronous detection is achieved by using the existing soil fertility prediction system and the soil fertility prediction system of the present invention, and the calculated results are compared with the initial known parameters to determine the required time and result accuracy of the two, and the records are shown in Table 3: Table 3 Comparison results

[0040] In summary, by using the soil fertility prediction system and method of the present invention in actual application operations, the time required to complete the detection operation is shorter and the accuracy of the results obtained is higher, thus better realizing the application of actual operations.

[0041] At the same time, the contents not described in detail in this specification belong to the existing technology well known to those skilled in the art.

[0042] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0043] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A soil fertility prediction system, characterized by: include: The data acquisition module uses the image acquisition module carried by the drone equipment to locate the area to be detected. After positioning is completed, the image acquisition operation of the area to be detected is performed, and the soil fertility data in the image is collected by the sensor acquisition module carried by the drone equipment. The collected data is transmitted and stored in the storage database; The data prediction module analyzes the collected images and determines the characteristics of the planting area. It extracts planting parameters from the storage database to implement image segmentation and determine the sampling points. It calculates the fertility index of each collection point based on the collected soil fertility data. It then establishes a fertility prediction model based on the image data and fertility index to predict the fertilization strategy for the detection area. The decision-making execution module formulates execution strategies based on the prediction results and cooperates with the fertilization equipment to complete fertilization operations in areas with different fertility levels; The result display module displays the result data obtained from the analysis and processing in the form of text and charts.

2. A soil fertility prediction system according to claim 1, characterized in that: The operations for image acquisition in the data acquisition module are as follows: Determine the boundaries of the area to be inspected, set the drone acquisition height, and the actual size of the image at the current height is A1 and A2. Plan the drone's flight path to achieve image acquisition. Set the acquisition starting point of the detection area, and the starting point is located on the horizontal boundary of the detection area parallel to the horizontal direction of the image, and the distance between the starting point and the left vertical boundary of the detection area is less than A2 / 2; Set the end point of the detection area, and the end point is located on the horizontal boundary of the detection area or the extension line of the horizontal boundary, and the distance between the end point and the right longitudinal boundary of the detection area is less than A2 / 2; The UAV device's movement path is as follows: first, it moves along the longitudinal boundary of the detection area, capturing images every time it moves a distance B1, and when it reaches the rear lateral boundary of the detection area shown in the image, it has moved a distance B2, thereby achieving an S-shaped path acquisition operation. After the acquisition is completed, the image data is spliced.

3. A soil fertility prediction system according to claim 2, characterized in that: The operations for splicing the image data are as follows: The image at the starting point is used as a reference, and the direction of the path is combined to perform a segmentation operation on the image at the starting point. The segmentation line is set at (A1-B1) from the rear lateral boundary of the image at the starting point. The segmentation operation is then performed, and the image with the larger area after segmentation is retained. The adjacent images in the path direction are spliced ​​with the retained image. And stitching is carried out in sequence until the image data has the horizontal boundary on the rear side of the detection area. The combined path segmentation line is set at (A2-B2) away from the vertical boundary on the right side of the current image, and then the segmentation operation is implemented, and the image with a larger area after segmentation is retained. The image stitching is continued in the combined path until the image stitching at the end point is completed to form a complete image data.

4. A soil fertility prediction system according to claim 3, characterized in that: The data prediction module analyzes the collected images and determines the characteristics of the planting area as follows: By performing a grayscale operation on the complete image data, a plurality of comparison points are set equidistantly between the midpoints of the boundary lines of the relative boundaries of the image data; The direction of the planting area feature is determined according to the change in the grayscale value of the comparison point, and then the feature interval line is obtained according to the different grayscale values ​​of the planting area feature and the separating gully feature. The midpoints of the adjacent comparison points where the grayscale value changes occur are connected in sequence as the interval line, and then the planting area feature and the separating gully feature are distinguished.

5. A soil fertility prediction system according to claim 4, characterized in that: The operations of extracting the planting parameters from the storage database to implement image segmentation and determine the sampling points in the data prediction module are as follows: The spacing distance between plants is determined to be M based on the planting parameters, and the planting area feature is divided into multiple grid areas at equal distances of M based on the spacing lines between the two sides of the planting area feature and the gully feature, and the position of the grid areas is marked in sequence; Randomly select any grid area as the basic sampling area, and connect the diagonal points of the current basic sampling area. The intersection generated by the intersection is the basic sampling point. Sampling points are set after being separated by a grid area on all four sides of the basic sampling point, and all sampling points are set according to the current method.

6. A soil fertility prediction system according to claim 1, characterized in that: The data prediction module calculates the fertility index of each collection point in combination with the collected soil fertility data as follows: Extract planting parameters to determine the pre-buried depth of plant seeds, that is, to achieve the pre-buried depth of the equipment's sensor directly below the collection point when collecting soil fertility; Set the fertility data value of a collection point as P i , P i It represents the detection value P of the i-th fertility data category, and performs a weighted operation on each fertility data category, and the weighted sum is the fertility index. The specific formula is: ； Q j is the fertility index Q at the jth sampling point, W i is the weighted value W of the i-th fertility data category, and n represents the total number of fertility data categories.

7. A soil fertility prediction system according to claim 5, characterized in that: The data prediction module combines image data and fertility index to establish a prediction model to achieve the fertilization strategy prediction operation for the detection area: A fertility prediction model for the current detection area is constructed by combining the characteristics of the planting area and the fertility index of each sampling point. The fertility prediction model is used to calculate the deviation values ​​of the fertility index of the sampling points with respect to the horizontal and vertical planting area characteristics in the detection area. The fertilization direction strategy in the planting area is determined based on the deviation values. Then, the planting parameters are extracted to determine the fertility interval requirement of the plant seeds as [F1, F2], and the average fertility index of the planting area is compared with the fertility interval requirement to determine the fertilization strategy; The predicted fertilization strategy results are obtained by combining the fertilization direction strategy and the fertilization amount strategy for transmission.

8. A soil fertility prediction system according to claim 7, characterized in that: The calculation operation of the deviation value is: The numerical calculation formula for the deviation of the horizontal planting area characteristics is: ； And Q m is the fertility index at the mth sampling point located in the horizontal planting area, Q c is the average fertility index of all sampling points in the horizontal planting area, and Q c =(Q1+Q2+…+Q m ) / m,; The calculation formula for the deviation of the vertical planting area characteristics is: ； and is the fertility index at the 2mk+1 sampling point of the longitudinal planting area characteristics, Q d is the average fertility index of all sampling points in the longitudinal planting area characteristics, and ; The deviation values ​​of the horizontal planting area characteristics and the vertical planting area characteristics are compared, and if S1 < S2, the direction of the horizontal planting area characteristics is used as the predicted fertilization direction, otherwise the direction of the vertical planting area characteristics is used as the predicted fertilization direction.

9. A soil fertility prediction system according to claim 8, characterized in that: The operation of comparing the average fertility index of the planting area with the fertility interval requirement is: If Q c ∈[F1, F2] or Q d ∈[F1, F2], it is predicted that the fertility index of the current planting area meets the fertility range requirements; If Q c >[F1, F2] or Q d > [F1, F2], the amount of fertilizer needs to be reduced; If Q c <[F1, F2] or Q d <[F1, F2], it is predicted that the fertility index of the current planting area is insufficient and the amount of fertilizer needs to be increased.

10. A soil fertility prediction method, using a soil fertility prediction system according to any one of claims 1 to 9, characterized in that: The specific steps include: Step 1: Use drone equipment to collect images of the test area, determine the characteristics of the planting area to be sampled, and determine the sampling points to implement soil fertility testing; Step 2: Based on the image data and fertility detection data, a fertility prediction model is established to predict the fertilization strategy for the detection area; Step 3: Implement the execution operations of planning direction and planning usage based on the prediction results.

Images