Soil fertility evaluation system, method, device and storage medium

Abstract

An embodiment of the present specification provides a soil fertility evaluation system and method, which includes at least one sensor deployed inside or on the surface of the soil in a target area, a display module, and a processor; the at least one sensor is used to obtain sensor data of the soil in the target area, and the sensor data includes soil data and environmental data; the processor is used to: obtain sensor data within a preset time period from the at least one sensor at a preset interval, and based on the sensor data; determine a comprehensive index value and reliability of the soil fertility in the target area at at least one future time point; determine a soil fertility evaluation result based on the comprehensive index value and reliability; the display module is used to display the soil fertility evaluation result.

Description

Technical Field

[0001] This specification relates to the field of soil fertility evaluation, and in particular to a soil fertility evaluation system, method, device and storage medium. Background Art

[0002] Soil fertility refers to the ability of soil to regularly and timely supply and coordinate the air, temperature, nutrients and non-toxic substances required for plant growth. Soil fertility is affected by factors such as soil texture, structure, moisture conditions, temperature conditions, biological conditions, organic matter content, and pH. Appropriate soil fertility can help plants grow better. Therefore, planting personnel need to collect soil fertility information to adjust the soil fertility to an appropriate level to better meet plant growth. Most of the existing technologies analyze multiple evaluation indicators based on mathematical algorithms or mathematical models to determine soil fertility or soil quality. However, it does not involve how to determine the long-term changes in soil fertility. If the method of sampling, analysis and calculation (such as fitting) at multiple time points is adopted, it is time-consuming and labor-intensive.

[0003] Therefore, it is hoped to provide a soil fertility evaluation system that can reasonably and accurately predict the development trend of soil fertility, while reducing the amount of data processing and improving the efficiency of soil fertility monitoring. Summary of the Invention

[0004] One embodiment of this specification provides a soil fertility evaluation system. The system includes at least one sensor deployed inside or on the surface of soil in a target area, a display module, and a processor; the at least one sensor is configured to acquire sensory data of the soil in the target area, the sensory data including soil data and environmental data; the processor is configured to acquire the sensory data from the at least one sensor within a preset time period at preset intervals; based on the sensory data, a comprehensive index value and reliability of soil fertility in the target area at at least one future time point are determined; a soil fertility evaluation result is determined based on the comprehensive index value and the reliability; and the display module is configured to display the soil fertility evaluation result.

[0005] One embodiment of this specification provides a soil fertility evaluation method. The method is executed by a processor of a soil fertility evaluation system. The method includes: acquiring sensor data from at least one sensor within a preset time period at preset intervals, the sensor data including soil data and environmental data; determining a comprehensive index value and reliability of soil fertility in a target area at at least one future time point based on the soil data and the environmental data; determining a soil fertility evaluation result based on the comprehensive index value and the reliability; and transmitting the soil fertility evaluation result to the display module.

[0006] One or more embodiments of the present specification provide a soil fertility evaluation device, comprising at least one processor and at least one memory; the at least one memory is used to store computer instructions; the at least one processor is used to execute at least part of the computer instructions to implement the soil fertility evaluation method.

[0007] One or more embodiments of the present specification provide a computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes a soil fertility evaluation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0009] Figure 1 is an exemplary module diagram of a soil fertility evaluation system according to some embodiments of this specification;

[0010] Figure 2 is an exemplary flow chart of a soil fertility evaluation method according to some embodiments of this specification;

[0011] Figure 3 is an exemplary flow chart for determining a comprehensive indicator value of soil fertility in a target area at at least one future time point according to some embodiments of this specification;

[0012] Figure 4 is an exemplary flow chart for determining the reliability of soil fertility of a target area at at least one future time point according to some embodiments of the present specification;

[0013] Figure 5 is another exemplary flow chart of determining the reliability of soil fertility in a target area at at least one future time point according to some embodiments of the present specification. DETAILED DESCRIPTION

[0014] To more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly describes the drawings required for describing the embodiments. Obviously, the drawings described below are merely examples or embodiments of this specification. Those skilled in the art can apply this specification to other similar scenarios based on these drawings without inventive effort. Unless otherwise apparent from the context or otherwise noted, the same reference numerals in the figures represent the same structure or operation.

[0015] It should be understood that the terms "system," "device," "unit," and / or "module" used herein are a method for distinguishing different components, elements, parts, portions, or assemblies at different levels. However, if other terms can achieve the same purpose, the terms may be replaced by other expressions.

[0016] As used in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not refer to the singular but also include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0017] Flowcharts are used throughout this specification to illustrate the operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations do not necessarily need to be performed in exact order. Instead, the steps may be processed in reverse order or simultaneously. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0018] Figure 1 It is an exemplary module diagram of a soil fertility evaluation system according to some embodiments of this specification.

[0019] In some embodiments, the soil fertility evaluation system 100 may include a processor 110, a display module 120, at least one sensor deployed inside or on the surface of the soil in the target area (e.g., sensor 130-1, sensor 130-2...sensor 130-n) and at least one automatic detection unit (e.g., automatic detection unit 140-1, automatic detection unit 140-2...automatic detection unit 140-n).

[0020] In some embodiments, the processor 110 may be configured to obtain sensor data from at least one sensor within a preset time period at preset intervals, and determine, based on the sensor data, a comprehensive index value and reliability of soil fertility in the target area at at least one future time point. The processor 110 may also determine a soil fertility evaluation result based on the comprehensive index value and reliability. The processor 110 may also be configured to control at least one automatic detection unit to detect the soil in the target area and obtain a detection result in response to the comprehensive index value and reliability at at least one future time point satisfying a first preset condition. The processor 110 may also determine a soil fertility evaluation result corresponding to each future time point based on the comprehensive index value, reliability, and detection result at at least one future time point.

[0021] In some embodiments, the processor 110 may also determine a comprehensive index value of soil fertility in the target area at at least one future time point based on the soil data and environmental data using a soil index prediction model. The soil index prediction model may be a machine learning model.

[0022] In some embodiments, the soil index prediction model may include a comprehensive index prediction layer and a confidence prediction layer. The comprehensive index prediction layer may be used to process soil data and environmental data to determine a reference comprehensive index value for soil fertility in a target area at at least one future time point. The confidence prediction layer may be used to process the reference comprehensive index value, the amount of soil data, the amount of environmental data, and a preset future interval duration to determine the confidence level corresponding to each reference comprehensive index value.

[0023] In some embodiments, the processor 110 may further determine the reliability of the soil fertility of the target area at at least one future time point based on the soil data, the environmental data, and appearance characteristic information of plants grown on the soil of the target area. The appearance characteristic information is determined by the processor based on an image captured by an image sensor and including an image of the plants grown on the soil of the target area.

[0024] The display module 120 can be used to display the soil fertility evaluation results.

[0025] The sensor 130 may be used to obtain sensing data of the soil in the target area, where the sensing data includes soil data and environmental data.

[0026] The automatic detection unit 140 can be used to automatically detect the soil in the target area.

[0027] For more details about the processor, display module, sensor and automatic detection unit, please see below. Figure 2-Figure 5 and its contents.

[0028] It should be understood that Figure 1 The system and its modules shown can be implemented in various ways. It should be noted that the above description of the soil fertility evaluation system and its modules is only for convenience of description and does not limit this specification to the scope of the embodiments. It is understandable that for those skilled in the art, after understanding the principles of the system, it is possible to arbitrarily combine the modules or form a subsystem connected with other modules without deviating from this principle. In some embodiments, Figure 1The processor 110, display module 120, sensor 130, and automatic detection unit 140 disclosed herein may be different modules within a system, or a single module may implement the functions of two or more of the aforementioned modules. For example, the modules may share a storage module, or each module may have its own storage module. Such variations are within the scope of protection of this specification.

[0029] Figure 2 is an exemplary flow chart of a soil fertility evaluation method according to some embodiments of this specification. In some embodiments, process 200 may be executed by the processor 110 of the soil fertility evaluation system 100. Figure 2 As shown, the process 200 may include the following steps 210 to 260.

[0030] Step 210: Acquire sensing data within a preset time period from at least one sensor at preset intervals, where the sensing data includes soil data and environmental data.

[0031] The preset interval refers to a preset time interval for the sensor to collect sensing data of the soil in the target area.

[0032] The preset time period refers to a preset time period during which the sensor collects sensor data of the soil in the target area. For example, the preset time period may be one week, one month, or the like.

[0033] The sensor can be used to obtain sensory data of the soil in the target area. For example, the sensor can include a temperature sensor, a soil moisture sensor, a soil pH sensor, a soil temperature sensor, a soil organic matter content sensor, etc.

[0034] The sensor data refers to sensor data that can reflect the soil fertility of the target area. In some embodiments, the sensor data may include soil data and environmental data.

[0035] Soil data refers to data related to the soil in the target area. For example, soil data can include one or more of the following: soil region data, soil texture data (e.g., sand content, clay content), soil structure data (e.g., the proportion of granular soil, the proportion of blocky soil), soil moisture data, soil temperature data, soil organic matter content data, soil pH content, etc. Different types of soil data can be collected using corresponding sensors. For example, soil pH can be collected using a soil pH sensor.

[0036] Environmental data refers to data related to the environment of the target area. For example, environmental data may include the weather in the target area.

[0037] In some embodiments, the processor may obtain sensing data within a preset time period from at least one sensor at preset intervals.

[0038] Step 220 : Determine a comprehensive index value and reliability of soil fertility in the target area at at least one future time point based on the soil data and the environmental data.

[0039] The comprehensive index value refers to a comprehensive index value reflecting the soil fertility in the target area. For example, the comprehensive index value may include one or more physical, chemical, and biological indicators of soil fertility in the target area. In some embodiments, the comprehensive index value may be expressed as a numerical value within 100, with a larger value indicating better soil fertility.

[0040] In some embodiments, the processor may determine a comprehensive index value of soil fertility in the target area at at least one future time point based on the soil data and environmental data using a first preset comparison table. In some embodiments, the first preset comparison table includes a correspondence between reference soil data and reference environmental data and a reference comprehensive index value of soil fertility in the target area at at least one future time point. In some embodiments, the first preset comparison table may be constructed based on prior knowledge or historical data.

[0041] In some embodiments, the processor may also determine the comprehensive index value of soil fertility in the target area at at least one future time point based on the soil data and environmental data through a soil index prediction model. For more details on this part, please see below. Figure 3 and its contents.

[0042] Reliability refers to the reliability corresponding to the comprehensive index value.

[0043] In some embodiments, the processor may determine the reliability of the comprehensive index value based on the soil data and the environmental data in the same manner as determining the comprehensive index value. For details, see above.

[0044] In some embodiments, the at least one sensor may include an image sensor. In some embodiments, the processor may determine the reliability of soil fertility in the target area at at least one future time point based on the soil data, the environmental data, and information about the appearance characteristics of plants planted on the soil in the target area.

[0045] An image sensor may be used to capture images of plants grown on soil encompassing a target area.

[0046] In some embodiments, appearance feature information may include plant height ranges, leaf features (e.g., leaf quantity ranges, leaf color ranges, etc.), flower and fruit features (e.g., presence of fruit, flower and fruit quantity ranges, flower and fruit size and color ranges, etc.). In some embodiments, plant appearance feature information may indirectly reflect a comprehensive index value. For example, the comprehensive index value is positively correlated with the lushness of plant growth derived from the appearance feature information.

[0047] In some embodiments, the appearance characteristic information may be determined by a processor based on an image of a plant grown on soil encompassing the target area, captured by an image sensor. For example, the appearance characteristic information may be determined by the processor based on multiple images of the plant grown on soil encompassing the target area, captured by an image sensor, at different growth stages, and then performing image recognition on the multiple images to obtain the appearance characteristic information of the plant.

[0048] In some embodiments, the processor can determine the reliability of the soil fertility of the target area at at least one future time point based on the appearance feature information of plants planted on the soil of the target area and the standard appearance feature information of the same plants planted under the same soil data, environmental data, planting time and other characteristics.

[0049] Standard appearance characteristic information refers to the appearance characteristic information that a plant should have under the same plant, soil data, environmental data, planting time and other conditions.

[0050] In some embodiments, the processor may determine the standard appearance feature information through a vector database based on the current comprehensive index value, weather conditions in the environmental data, plant species, and collection time.

[0051] For an explanation of the comprehensive index values, see Figure 2 Description of step 220. Weather conditions refer to the overall trend of weather conditions measured during a preset time period, from the time the plants are planted to the time point when the sensor data is collected, for example, (weather condition 1, weather condition 2, ...).

[0052] The collection time refers to the plant's growth period from the time the plant is planted to the time at least one sensor collects sensor data from the plant. For example, the collection time may include the plant's growth period from the time the plant is planted to the time at which at least one sensor collects sensor data from the plant during its seedling stage. The collection time is related to the plant's growth. The processor may subtract the time the plant was planted from the time the sensor data was collected to obtain the collection time.

[0053] Specifically, the processor can determine the first target feature vector based on the current comprehensive index value, weather conditions in the environmental data, plant species, and collection time, and determine the first associated feature vector through the vector database based on the first target feature vector; and determine the reference standard appearance feature information corresponding to the first associated feature vector as the standard appearance feature information at the current time point.

[0054] The vector database includes multiple first reference feature vectors, each of which has corresponding reference standard appearance feature information. The first reference feature vectors are constructed based on historical comprehensive indicator values, weather conditions in historical environmental data, historical plant species, and historical collection times.

[0055] In some embodiments, the processor may determine, based on the first target feature vector, a first reference feature vector that meets a preset condition in a vector database, and determine the first reference feature vector that meets the preset condition as the first associated feature vector. In some embodiments, the preset condition may include a minimum vector distance from the first target feature vector.

[0056] In some embodiments, the processor may determine the standard appearance feature information at the current time point based on the reference standard appearance feature information corresponding to the determined first associated feature vector.

[0057] In some embodiments, the processor may calculate the similarity between the standard appearance feature information and the actual appearance feature information of the plants planted on the soil of the target area at the current time point as the reliability of the soil fertility of the target area at the current time point.

[0058] In some embodiments, the processor may multiply the reliability of the current time point by the reliability attenuation coefficient to determine the reliability of the next future time point adjacent to the current time point, and so on, to determine the reliability of the soil fertility of the target area at at least one future time point.

[0059] In some embodiments, the processor may obtain the reliability attenuation coefficient using a statistical method based on historical data. For example, the processor may calculate the relationship between the current reliability and the reliability at various future time points in the historical data to obtain the reliability attenuation coefficient.

[0060] In some embodiments, the processor may also determine the reliability of soil fertility in the target area at at least one future time point based on other methods. For details, see Figure 4 and Figure 5 content.

[0061] Step 230: Determine the soil fertility evaluation result based on the comprehensive index value and reliability.

[0062] Soil fertility evaluation results refer to the evaluation results of the comprehensive level of soil fertility in the target area.

[0063] In some embodiments, the processor may directly use the comprehensive index value and the reliability as the soil fertility evaluation result. In some embodiments, the processor may also use the product of the comprehensive index value and the reliability as the soil fertility evaluation result.

[0064] In some embodiments, the system 100 may further include at least one automatic detection unit deployed within or on the surface of the soil in the target area. The automatic detection unit may be configured to automatically detect the soil in the target area. In some embodiments, the processor may control the at least one automatic detection unit to detect the soil in the target area in response to the comprehensive index value and reliability satisfying a first preset condition, obtain a detection result, and determine a soil fertility evaluation result based on the comprehensive index value, reliability, and detection result.

[0065] The first preset condition refers to the preset conditions that the comprehensive index value and the reliability need to meet respectively. In some embodiments, the first preset condition may refer to that the weighted mean of the comprehensive index value of at least one future time point is less than the preset comprehensive index value threshold and the reliability is greater than the preset first reliability threshold. In some embodiments, the first preset condition may also refer to that the weighted mean of the comprehensive index value of at least one future time point is less than the preset comprehensive index value threshold and the weighted mean of the reliability is greater than the first reliability threshold. The weight of the reliability can be that the closer to the current time point, the greater the weight of the reliability. The weighted mean of the comprehensive index value refers to the value obtained by multiplying the comprehensive index value corresponding to each future time point in at least one future time point by the corresponding weight and then adding them all up. In some embodiments, when calculating the weighted mean of the comprehensive index value, the closer the future time point is to the current time point, the greater the weight of the comprehensive index value corresponding to the future time point.

[0066] In some embodiments, the preset comprehensive indicator threshold in the first preset condition may be related to the reliability. For example, when the reliability is poor, the preset comprehensive indicator threshold may be increased.

[0067] In some embodiments of the present specification, by correlating the preset comprehensive index value threshold in the first preset condition with the reliability, when the reliability is poor, the preset comprehensive index threshold is adjusted to increase so as to trigger the automatic detection unit to perform automatic detection in a timely manner, thereby obtaining more accurate detection results, so as to judge the soil fertility evaluation results, thereby improving the accuracy of the soil fertility evaluation results.

[0068] The automatic detection unit refers to a unit that can automatically detect soil fertility. For example, the automatic detection unit can include a soil fertilizer nutrient detector.

[0069] The test result refers to the test result of soil fertility directly obtained after using the automatic detection unit to test the soil in the target area.

[0070] In some embodiments, when the difference between the weighted mean of the comprehensive index values ​​at at least one future time point and the detection result is less than a preset threshold and the reliability is greater than a first preset reliability threshold, the processor may use the predicted comprehensive index value at at least one future time point as the soil fertility evaluation result. In some embodiments, the preset reliability threshold is related to the absolute value of the difference between the weighted mean of the comprehensive index values ​​and the detection result of the automatic detection unit. For example, the smaller the absolute value of the difference, the smaller the preset reliability threshold.

[0071] In some embodiments, when the weighted mean of the comprehensive index value at at least one future time point is greater than or equal to the detection result, or the reliability is less than or equal to a preset second reliability threshold, the processor needs to re-predict the comprehensive index value and reliability at at least one future time point, for example, based on Figure 3 The first reliability threshold and the second reliability threshold may be different.

[0072] Step 240: Send the soil fertility evaluation result to the display module.

[0073] The display module refers to a device that can display the soil fertility evaluation results.

[0074] In some embodiments, the processor may send the soil fertility evaluation result to the display module via a network.

[0075] In some embodiments of the present specification, by analyzing soil data and environmental data related to soil fertility within a preset time period, the comprehensive index value and reliability of the soil fertility of the target area at at least one future time point are determined to determine the soil fertility evaluation result, and then based on the soil fertility evaluation result, the soil fertility development trend of the target area is reasonably and accurately predicted, while reducing the data processing volume to improve the efficiency of soil fertility monitoring.

[0076] In addition, since the detection structures (such as reagents, etc.) in the automatic detection unit are mostly one-time or limited in number, it is difficult to perform detection anytime and anywhere. By setting the comprehensive index value and reliability to meet the first preset condition (for example, the predicted comprehensive index value of soil fertility is poor), the automatic detection unit will be automatically detected to save resources. At the same time, the detection results of the automatic detection unit can also be used as supplementary data to jointly determine the soil fertility evaluation results to improve the accuracy of the soil fertility evaluation results, and then more accurately predict the soil fertility development trend of the target area, while reducing the amount of data processing to further improve the efficiency of soil fertility monitoring.

[0077] It should be noted that the above description of process 200 is for illustration and purpose only and does not limit the scope of application of this specification. Those skilled in the art may make various modifications and variations to process 200 under the guidance of this specification. However, such modifications and variations are still within the scope of this specification.

[0078] Figure 3 This is an exemplary flow chart for determining a comprehensive index of soil fertility of a target area at at least one future time point according to some embodiments of this specification. In some embodiments, process 300 may be executed by the processor 110 of the soil fertility evaluation system 100 .

[0079] In some embodiments, the processor can determine the comprehensive index value 360 ​​of soil fertility of the target area at at least one future time point based on the soil data 310-1 and the environmental data 310-2 using the soil index prediction model 300. For a description of soil data and environmental data, see Figure 2 Description of step 210. For description of comprehensive index values, please refer to Figure 2 Description of step 220.

[0080] In some embodiments, the soil index prediction model may be a machine learning model. In some embodiments, the soil index prediction model may be a machine learning model with a custom structure as described below. The soil index prediction model may also be a machine learning model with other structures, such as a time series model.

[0081] In some embodiments, the input of the soil indicator prediction model 300 may also include planting data 310 - 3 .

[0082] Planting data may include one or more of the following: frequency of tillage, frequency and amount of fertilization, frequency of residue cleaning (e.g., frequency of leaf removal), and planting density of the target area. Planting density refers to the number and type of plants planted per unit area of ​​soil, as well as the average distribution distance between adjacent plants planted on the soil. Average distribution distance refers to the average value of the distance between any two adjacent plants among all plants planted on the soil. The processor may determine the planting data of the plant by obtaining user input. The processor may also obtain the planting data of the plant from a known database. The known database stores the planting data of the plant.

[0083] In some embodiments of the present specification, since human operations may affect the comprehensive index value of the soil, the input of the soil index prediction model also considers the impact of planting data on the comprehensive index value of soil fertility in the target area at at least one future time point, which can further improve the accuracy of the comprehensive index value predicted by the soil index prediction model.

[0084] In some embodiments, the soil index prediction model 300 may include a comprehensive index prediction layer 320 and a confidence prediction layer 340. In some embodiments, the comprehensive index prediction layer and the confidence prediction layer may be time series models.

[0085] In some embodiments, the comprehensive index prediction layer 320 can be used to process the soil data 310 - 1 and the environmental data 310 - 2 to determine a reference comprehensive index value 330 of soil fertility in the target area at at least one future time point.

[0086] The reference comprehensive indicator value refers to the comprehensive indicator value to be selected.

[0087] In some embodiments, the input of the comprehensive indicator prediction layer 320 may also include planting data 310 - 3 .

[0088] In some embodiments, the confidence prediction layer 340 can be used to process the reference comprehensive index value 330, the data volume 330-1 of soil data, the data volume 330-2 of environmental data and the preset future interval duration 330-3 to determine the confidence 350 corresponding to each reference comprehensive index value.

[0089] The data volume of soil data refers to the total amount of soil data. For a description of soil data, see Figure 2 Step 210. For example, the amount of soil data may include the amount of soil data of the target area in the past five months or the amount of soil data of the past three months. The amount of environmental data refers to the total amount of environmental data. For an explanation of environmental data, see Figure 2 Step 210. The amount of environmental data is similar to that of soil data.

[0090] The preset future interval length refers to the length of the period between the future time point corresponding to the preset predicted next comprehensive indicator value and the previous future time point.

[0091] The amount of soil data and environmental data, and the length of the preset future interval can affect the accuracy of the confidence level corresponding to the predicted values ​​of each reference comprehensive indicator.

[0092] In some embodiments, the soil index prediction model 300 may output a comprehensive index value 360 ​​for at least one future time point based on the confidence level 350 corresponding to each of the at least one reference comprehensive index values ​​corresponding to each future time point. For example, the soil index prediction model 300 may select the reference comprehensive index value with the highest confidence level among the confidence levels 350 corresponding to each of the at least one reference comprehensive index values ​​corresponding to each future time point and output it as the comprehensive index value 360 ​​for the corresponding future time point.

[0093] In some embodiments, the soil indicator prediction model can be obtained by jointly training a comprehensive indicator prediction layer and a confidence prediction layer based on a large number of first training samples with first labels.

[0094] In some embodiments, the first training sample may include historical sample soil data, historical sample environmental soil data, the data volume of historical sample soil data, the data volume of historical sample environmental data, and the preset interval length of the historical sample. In some embodiments, the first training sample can be obtained based on historical data. The first label can be obtained based on the comprehensive index value actually measured of the historical sample corresponding to the input data in the historical data. For example, the first label can be: 1-(comprehensive index value predicted by the comprehensive index prediction layer-comprehensive index value actually measured by the historical sample) / comprehensive index value actually measured by the historical sample. The first label can be a real number between 0 and 1.

[0095] In some embodiments, the historical sample environmental data and historical sample soil data in the first training sample with the first label can be input into the initial comprehensive index prediction layer, and then the reference comprehensive index value of at least one historical sample time point output by the initial comprehensive index prediction layer, the data volume of the historical sample soil data, the data volume of the historical sample environmental data and the preset interval duration of the historical samples are input into the initial confidence prediction layer. A loss function is constructed through the prediction results of the first label and the initial soil index prediction model, and the parameters of the initial comprehensive index prediction layer and the initial confidence prediction layer are iteratively updated based on the loss function until the loss function converges, the number of iterations reaches a threshold, etc., the training is completed, and a trained soil index prediction model is obtained.

[0096] In some embodiments, each group of first training samples in the first training samples may further include historical sample planting data.

[0097] In some embodiments of the present specification, by processing soil data and environmental data using a trained soil index prediction model, the comprehensive index value and confidence level of the soil fertility of the target area at at least one future time point can be quickly and accurately determined, and the soil fertility evaluation results can be determined more quickly and accurately while reducing the amount of data processing. Then, based on the soil fertility evaluation results, the soil fertility development trend of the target area can be reasonably and accurately predicted, further improving the efficiency of soil fertility monitoring.

[0098] Figure 4 is an exemplary flow chart of a method for determining the reliability of soil fertility in a target area at at least one future time point according to some embodiments of this specification. In some embodiments, process 400 may be executed by the processor 110 of the soil fertility evaluation system 100. Figure 4 As shown, the process 400 may include the following steps 410 to 430.

[0099] Step 410 : Based on the soil data and the environmental data, a comprehensive index value of soil fertility in the target area at at least one future time point is determined by a soil index prediction model.

[0100] Step 420 , based on the comprehensive index value, weather conditions in the environmental data, plant species, and acquisition time series, determine a standard appearance feature sequence of soil fertility in the target area at at least one future time point.

[0101] For an explanation of comprehensive index values ​​and weather conditions, see Figure 2 Relevant instructions for step 220.

[0102] The acquisition time series refers to a sequence consisting of multiple acquisition times. For an explanation of the acquisition time, see Figure 2 Relevant description of step 220. For example, the acquisition time series may be (10 days, 11 days, ...).

[0103] The standard appearance feature sequence is a sequence consisting of the standard appearance feature information corresponding to each acquisition time in the acquisition time sequence. For a description of the standard appearance feature information, see Figure 2 Relevant description of step 220. The standard appearance feature sequence can reflect the continuous change of soil fertility in the target area.

[0104] In some embodiments, the processor may determine a standard appearance feature sequence through a vector database based on the current comprehensive index value, weather conditions in the environmental data, plant species, and acquisition time series.

[0105] For example, the processor can determine the second target feature vector based on the current comprehensive index value, weather conditions in the environmental data, plant species, and collection time series; based on the second target feature vector, determine the second associated feature vector through the vector database; and determine the reference standard appearance feature sequence corresponding to the second associated feature vector as the standard appearance feature sequence at the current time point.

[0106] The vector database contains multiple second reference feature vectors, where each second reference feature vector has a corresponding reference standard appearance feature sequence. The second reference feature vector is a feature vector constructed based on historical comprehensive index values, weather conditions in historical environmental data, historical plant species, and historical collection time series. For instructions on how to determine the second associated feature vector based on the second reference feature vector, please refer to Figure 2 In step 220 , a description of the first associated feature vector is determined.

[0107] Step 430 : Determine the reliability of the soil fertility of the target area at at least one future time point based on a comparison between the appearance feature information corresponding to each time point in the appearance feature sequence and the standard appearance feature information in the standard appearance feature sequence corresponding to each time point.

[0108] The appearance feature sequence is a sequence consisting of the actual appearance feature information of the plant corresponding to each collection time in the collection time sequence. For a description of appearance feature information, please refer to Figure 2 Relevant description of step 220. The appearance feature sequence can reflect the actual continuous change of soil fertility in the target area.

[0109] In some embodiments, the processor can collect images of plants grown on soil containing the target area based on the image sensor at each collection time, determine appearance feature information at each collection time, and then obtain an appearance feature sequence. For details about the processor collecting images of plants grown on soil containing the target area based on the image sensor, please refer to Figure 2 Relevant content in step 220.

[0110] In some embodiments, the processor can calculate the similarity between the appearance feature information corresponding to each time point in the appearance feature sequence and the standard appearance feature information corresponding to each time point in the standard appearance feature sequence, and then perform weighted averaging to obtain the result as the reliability of the soil fertility of the target area at the current time point.

[0111] In some embodiments, when calculating the weighted average of similarities, the weight is determined in such a manner that the closer the collection time point of the appearance feature information corresponding to the similarity is to the current time point, the greater the weight corresponding to the similarity.

[0112] In some embodiments, the processor may multiply the reliability of the current time point by the reliability attenuation coefficient to determine the reliability of a future time point adjacent to the current time point. Figure 2 Relevant instructions in step 220.

[0113] In some embodiments, the processor may also perform difference calculations between the appearance feature information at each time point in the appearance feature sequence and the standard appearance feature information in the standard appearance feature sequence corresponding to each time point to obtain multiple differences; and determine the reliability of the soil fertility of the target area at at least one future time point based on the weighted results of the multiple differences.

[0114] In some embodiments, the processor can determine the reliability of the soil fertility of the target area at the current time point based on a calculation formula. For example, the reliability at the current time point is a=k / [Σ(ki*bi)], where bi is the difference between the appearance feature information at the i-th acquisition time point in the appearance feature sequence and the standard appearance feature information in the standard appearance feature sequence corresponding to the time point, k is a preset coefficient, and ki is the weight of the difference between the appearance feature information at the i-th time point and the standard appearance feature information at the time point. The weight of the difference between the appearance feature information at each time point and the standard appearance feature information at the corresponding time point can be determined based on the planting data on the soil.

[0115] In some embodiments, for soils with different planting data, the weights of the differences between the appearance feature information at each time point and the standard appearance feature information at the corresponding time point may be different. For example, if the difference between the appearance feature information of a plant at a certain time point and the standard appearance feature information at that time point has a large randomness (e.g., if the appearance feature information of a plant at a certain time point is the appearance feature information of an unnatural change caused by human intervention in the planting of plants on the soil (e.g., crop rotation, fertilization, cleaning, etc.), then the difference between the appearance feature information and the standard appearance feature information at that time point has a large randomness), then the difference between the appearance feature information of the plant at that time point and the standard appearance feature information at that time point is set with a smaller weight.

[0116] In some embodiments, the processor may multiply the reliability of the current time point by the reliability attenuation coefficient to determine the reliability of a future time point adjacent to the current time point. Figure 2 Relevant instructions in step 220.

[0117] In some embodiments of the present specification, the reliability of the soil fertility of the target area at at least one future time point is determined by calculating multiple differences between the appearance feature information of the plants planted on the soil of the target area at each time point in the appearance feature sequence and the standard appearance feature information corresponding to each time point in the standard appearance feature sequence and performing a weighted summation. Since the long-term soil fertility changes of the soil in the target area are taken into consideration, the accuracy of the reliability of the assessment of the long-term fertility changes of the soil in the target area can be further improved.

[0118] In some embodiments of the present specification, the reliability of soil fertility in a target area at at least one future time point is determined by comparing appearance feature information at each time point determined based on images of plants grown on the soil in the target area with standard appearance feature information corresponding to each time point, thereby improving the accuracy of the determined reliability. Furthermore, when determining the standard appearance feature information at each time point, multiple factors influencing the plant's appearance feature information are taken into account, such as the comprehensive index value, weather conditions in the environmental data, plant species, and the time series of collection, further improving the accuracy of the determination of the standard appearance feature information at each time point.

[0119] Figure 5 is an exemplary flow chart for determining the reliability of soil fertility in a target area at at least one future time point according to some embodiments of this specification. In some embodiments, process 500 may be executed by the processor 110 of the soil fertility evaluation system 100. Figure 5 As shown, the process 500 may include the following steps 510 to 530.

[0120] Step 510 : Based on the soil data and the environmental data, a comprehensive index value of soil fertility in the target area at at least one future time point is determined by a soil index prediction model.

[0121] Step 520 , based at least on the appearance feature sequence and the comprehensive index value of the plants planted on the soil of the target area, predict appearance feature information of at least one future time point using an appearance feature prediction model.

[0122] In some embodiments, the appearance feature prediction model may be a machine learning model. In some embodiments, the appearance feature prediction model may include a neural network model (NN), a deep neural network (DNN), etc.

[0123] In some embodiments, the input of the appearance feature prediction model may include appearance feature sequence, comprehensive index value, collection time series, plant species and future weather data, and the output may include appearance feature information or appearance feature sequence of at least one future time point.

[0124] Future weather data refers to weather data related to at least one future time point. The processor can obtain future weather data from external weather bureaus. For a description of appearance feature sequences, see Figure 4 For details about the appearance characteristics, see Figure 2 For instructions on collecting time series, see Figure 4 Related description in step 420.

[0125] In some embodiments, the appearance feature prediction model is trained based on a large number of second training samples with second labels.

[0126] In some embodiments, the second training sample may include a historical sample appearance feature sequence, a historical sample collection time sequence, a historical sample plant species, weather data at the time of the second historical sample collection, and an actual comprehensive index value at the time of the second historical sample collection. The second historical sample collection time is after the first historical sample collection time, weather data at the time of the first historical sample collection exists before the weather data at the time of the second historical sample collection, and the actual comprehensive index value at the time of the second historical sample collection exists before the actual comprehensive index value at the time of the second historical sample collection. The second training sample can be obtained based on the historical data. The second label of the second training sample can be obtained based on the appearance feature sequence corresponding to the second historical sample collection time corresponding to the input data in the historical data.

[0127] Step 530 : Determine the reliability of the comprehensive index value of soil fertility in the target area at at least one future time point based on the appearance feature information and the actual appearance feature information corresponding to the arrival of the plant at each future time point.

[0128] Actual appearance feature information refers to appearance feature information under actual circumstances. Actual appearance feature information can be obtained by collecting images of plants on the soil of the target area using an image sensor and processing the images using image recognition. Image recognition methods may include, but are not limited to, one or more of structural image recognition methods and fuzzy image recognition methods. For a description of appearance feature information, please refer to Figure 2 Relevant instructions in step 220.

[0129] In some embodiments, the processor can construct a historical database based on the appearance feature information and the actual appearance feature information corresponding to the plant's arrival at each future time point. A specific method for constructing the historical database is as follows: the processor can calculate the vector distance between the predicted appearance feature information of the plant at a future time point and the actual appearance feature information of the plant at the future time point. If the vector distance between the predicted appearance feature information at the future time point and the actual appearance feature information of the plant growing to the future time point is less than a preset vector distance threshold, the predicted comprehensive index value for this future time point can be considered accurate, and the comprehensive index value for the future time point is saved and recorded in the database. At the same time, the processor can also determine the reliability of the accurate comprehensive index value based on a preset algorithm and record it in the historical database in association with the comprehensive index value for the future time point. The method for determining the reliability of the comprehensive index value by the preset algorithm is: (preset vector distance - vector distance) / preset vector distance obtained as the percentage as the reliability. The preset vector distance can be preset by a person skilled in the art based on experience. The processor can also determine the reliability of the comprehensive index value of the soil fertility of the target area at at least one future time point based on the historical database. For example, the processor finds based on historical data in the historical database that 9 out of 10 predicted soil fertility comprehensive index values ​​one month later are more accurate, and the historical database records the reliability corresponding to each predicted soil fertility comprehensive index value one month later. Then, the reliability corresponding to the 10 predicted soil fertility comprehensive index values ​​one month later can be averaged as the reliability corresponding to the predicted soil fertility comprehensive index value one month later.

[0130] In some embodiments, the processor may further calculate the vector distances between the appearance feature information predicted at each future time point in the appearance feature sequence and the actual appearance feature information of the plant when it grows to the corresponding future time point, and then perform weighted averaging of the vector distances to obtain a value to construct a historical database, and determine the reliability of the comprehensive index value of the soil fertility of the target area at at least one future time point based on the historical database. The closer to the current time point, the higher the weight corresponding to the vector distance between the appearance feature information predicted at the future time point and the actual appearance feature information of the plant when it grows to the future time point. The historical database is constructed in the same manner as based on the vector distance based on the value obtained by weighted averaging of the vector distances. For details, see above.

[0131] In some embodiments of the present specification, based on a method of comparing the appearance feature information of the plant at a predicted future time point with the actual appearance feature information of the plant when it arrives at the future time point, while reducing the amount of data processing, it is possible to quickly determine the reliability corresponding to the comprehensive index values ​​of the multiple future time points, which is used to predict the soil fertility development trend of the target area, so as to further improve the efficiency of soil fertility monitoring.

[0132] While the basic concepts have been described above, it will be apparent to those skilled in the art that the detailed disclosure is merely illustrative and does not limit this specification. Although not explicitly stated herein, various modifications, improvements, and revisions to this specification may be made by those skilled in the art. Such modifications, improvements, and revisions are suggested in this specification and remain within the spirit and scope of the exemplary embodiments of this specification.

[0133] This specification also uses specific terms to describe the embodiments of this specification. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "one embodiment," "an embodiment," or "an alternative embodiment" two or more times in different locations in this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of this specification may be appropriately combined.

[0134] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0135] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0136] In some embodiments, numbers are used to describe the quantity of components and attributes. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise stated, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the description and claims are approximate values, which may change according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of this specification are approximate values, in specific embodiments, the settings of such numerical values ​​are as accurate as possible within the feasible range.

[0137] Each patent, patent application, patent application publication, and other materials, such as articles, books, specifications, publications, and documents, cited in this specification is hereby incorporated by reference in its entirety. This includes application history documents that are inconsistent with or conflict with the content of this specification, as well as documents (currently or subsequently attached to this specification) that limit the broadest scope of the claims of this specification. It should be noted that if the descriptions, definitions, and / or terminology used in the accompanying materials are inconsistent or conflicting with the content of this specification, the descriptions, definitions, and / or terminology used in this specification will control.

[0138] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A soil fertility evaluation system, characterized in that: The system includes at least one sensor deployed in or on the soil of a target area, a display module, and a processor; The at least one sensor is used to obtain sensor data of the soil in the target area, wherein the sensor data includes soil data and environmental data; The processor is configured to: acquiring the sensing data within a preset time period from the at least one sensor at a preset interval; Determining a comprehensive index value and reliability of soil fertility in the target area at at least one future time point based on the sensor data; Determining a soil fertility evaluation result based on the comprehensive index value and the reliability; The display module is used to display the soil fertility evaluation result; The processor is further configured to: Determining a comprehensive index value of soil fertility in the target area at at least one future time point using a soil index prediction model based on the soil data and the environmental data, wherein the soil index prediction model is a machine learning model; The soil index prediction model includes a comprehensive index prediction layer and a confidence prediction layer; The comprehensive index prediction layer is used to process the soil data and the environmental data to determine a reference comprehensive index value of the soil fertility of the target area at at least one future time point; as well as The confidence prediction layer is used to process the reference comprehensive index value, the data volume of the soil data, the data volume of the environmental data, and the preset future interval time to determine the confidence corresponding to each reference comprehensive index value; The at least one sensor includes an image sensor, and the processor is further configured to: determining the reliability of soil fertility in the target area at at least one future time point based on the soil data, the environmental data, and appearance characteristic information of plants planted on the soil in the target area; The appearance feature information is determined by the processor based on an image of plants planted on soil including the target area captured by the image sensor.

2. The system according to claim 1, wherein: The system further includes at least one automatic detection unit deployed inside or on the surface of the soil in the target area; The automatic detection unit is used to automatically detect the soil in the target area; The processor is further configured to: In response to the comprehensive index value and the reliability satisfying a first preset condition, controlling the at least one automatic detection unit to detect the soil in the target area and obtain a detection result; The soil fertility evaluation result is determined based on the comprehensive index value, the reliability and the detection result.

3. A soil fertility evaluation method, characterized in that: The method is executed by a processor of the soil fertility evaluation system according to claim 1, and the method includes: acquiring sensing data within a preset time period from the at least one sensor at preset intervals, the sensing data including soil data and environmental data; Determining a comprehensive index value and reliability of soil fertility in the target area at at least one future time point based on the soil data and the environmental data; Determining a soil fertility evaluation result based on the comprehensive index value and the reliability; and The soil fertility evaluation result is sent to the display module.

4. The method according to claim 3, characterized in that Determining the comprehensive index value of soil fertility in the target area at at least one future time point based on the soil data and the environmental data includes: Based on the soil data and the environmental data, a comprehensive index value of the soil fertility of the target area at at least one future time point is determined by a soil index prediction model, and the soil index prediction model is a machine learning model.

5. The method according to claim 3, characterized in that The at least one sensor includes an image sensor, and determining the reliability of soil fertility in the target area at at least one future time point based on the soil data and the environmental data includes: determining the reliability of soil fertility in the target area at at least one future time point based on the soil data, the environmental data, and appearance characteristic information of plants planted on the soil in the target area; The appearance feature information is determined by the processor based on an image of plants planted on soil including the target area captured by the image sensor.

6. A soil fertility evaluation device, characterized in that: The apparatus comprises at least one processor and at least one memory; The at least one memory is for storing computer instructions; The at least one processor is configured to execute at least part of the computer instructions to implement the soil fertility evaluation method according to any one of claims 3 to 5.

7. A computer-readable storage medium, characterized in that The storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the soil fertility evaluation method according to any one of claims 3 to 5.

Images