Description:Field of the invention
The present disclosure relates to a systems and method for agricultural soil and estimating soil nitrogen content. The present disclosure additionally relates to satellite remote sensing.
Background
Agricultural productivity depends on the quality of the farmland and soil tests can identify crop growth problems in the early stages. Nutrient content information is useful for accurate fertilization throughout precision agriculture implementations. All crops need nitrogen, which optimizes yield and plays an important role within the plant in ensuring nutrient is available when and where it is needed. This nutrient is even found in the roots, where proteins and enzymes regulate water and nutrient absorption.
In order to increase crop yield and agricultural fertilization, soil nutrient estimation is used as a key input. Due to the limited availability of ground-measured spectrum technology and the high cost of hyper spectral images, multispectral remote sensing is used to estimate soil nitrogen content.
If inappropriate amounts of nitrogen fertilizer are added to the soil over time, the balance in the total or available nitrogen is disturbed, resulting in lower crop yields and also increasing nitrate levels in groundwater beyond permissible limits, causing environmental pollution. The amount of nitrogen applied to fields every year by farmers is 115 million tons, but only 35% of it is used, resulting in 75 million tonnes of nitrogen entering waterways every year. It is estimated that nearly two-thirds of the nitrogen applied becomes an environmental pollutant as a result of this "nitrogen surplus." This not only causes environmental pollution but also increases agribusiness input costs by up to 50% and reduces crop yields by up to 40%.
In agriculture, farmers must balance giving their crops the right amount of fertilizer while minimizing costs and environmental damage. In Traditional method, physical soil sampling is done to check the nutrient content in the soil which is labor intensive and time consuming.

US11122734B1 describes a nitrogen analysis subsystem automatically analyses hyperspectral and/or multispectral remote sensing photographs to determine the value of nitrogen variables of crops or other plant life included in the images. For instance, the nitrogen analysis subsystem might have modules for data collection, function generation, and nitrogen estimator. The function generator module trains a mapping function using training data that was prepared by the data collector module. The nitrogen estimator module then use the mapping function to determine the values of nitrogen variables for a fresh remote sensing image that is not part of the training set. The results can then be reported and/or utilised to determine how much fertiliser should be added to a field of crops to promote plant growth.
US11468669B2 describes a computer-implemented method for predicting subfield soil properties for an agricultural field comprises, receiving satellite remote sensing data that contains several images of an agricultural field captured in various optical domains is one step in a computer-implemented technique for estimating subfield soil parameters for an agricultural field. Obtaining many environmental traits for the agriculture sector; using a variety of environmental factors and a variety of satellite remote sensing data to produce a variety of preprocessed images; using the variety of preprocessed images to identify a variety of agricultural field features; putting one or more machine learning models to work on the variety of features to produce a subfield soil property prediction for the agricultural field; sending the subfield soil property prediction to an agricultural computer system.
In the prior arts for monitoring the soil property or nitrogen content sample of the soil is required. Present scenario the challenges faced during physical sampling of the soil:
1. Problems in sampling due to soil variation across a field: Uniformity of soil sampling depth is one of the most critical parts of soil testing, yet it is one of the most common sources of error. If sampling is done without following the protocol, then the report may have the inappropriate results.
2. High cost per analysis: Physical soil testing is costlier where small and marginal farmers are unable to afford.
3. Suggest which crop to grow: Soil testing laboratories don't suggest which crop to grow according to the soil report.
4. To Replace good management: Generated soil reports from laboratories fail to provide nutrient management at the farm/region level.
5. Availability of soil testing laboratories: Farmers are unable to get soil tested for nutrient analysis due to a lack of soil testing laboratories.
6. Time taken to deliver soil report: It takes about 1-2 weeks for delivering the physical soil testing report leading to delay in knowing the nutrient content of the soil and planning for the future farm activities.
Implementing time-consuming and expensive soil sampling procedures are typically required for quantitative and absolute measurements of nitrogen content. One of the possible options is soil core sampling, which is typically difficult to scale. Furthermore, geographically dispersed maps created from point-based soil samples frequently use difficult and unreliable interpolation methods.
Present invention estimate the nitrogen content without hyperspectial image data. This invention does not require the physical sampling and hardware/model / apparatus for estimation the soil property.
Present invention offers satellite-based soil nutrient analysis at every 10x10m area. In contrast to the conventional system, which only offered one value per farm, our results are given for the entire farm area. Our results provide better coverage and are accurate for calculating nitrogen value for the entire farm because nutrient value might vary even within the same farm. This aids the farmer in getting the accurate level of nitrogen in the whole geotagged field. Additional benefits of our inventions are: i) there is no need for any physical model or hardware on the farm, and ii) farmers/users get the soil nutrient report instantly.
Summary
The present invention overcomes the limitations of the prior art by estimating the nitrogen concentration in soil. With the use of our technology, a farmer can geotag (locate and mark the farm on a map) the farm and get the farm's soil nitrogen status instantly. Every 10 metres, the soil nitrogen status is available, and recommendations for the necessary nitrogen dosage are also given. Using satellite-based data acquired over a farm, calibrated soil nitrogen model, and filtering and processing of the satellite data's spectral bands, information on the status of the soil nitrogen is delivered. The invention doesn't call for any actual soil sampling, doesn't call for taking soil samples any place, and gives the report instantly.
Present invention is related to a method and system. In the system, farmer has to geotag his farming area. A farm's geolocation is added to the cloud server, where the algorithm automatically runs, once it has been geotagged. Over the farm boundary, the algorithm downloads the time series of satellite data, and then it chooses the band that is appropriate for estimating nutrition. The data is then pre-processed, including cloud removal, atmospheric adjustment, radiometric correction, and speckle filtering, for the time when the farm is barren. A calibrated model is used to predict the soil nitrogen content from the chosen and pre-processed data. The farm is then divided into various zones depending on the estimated soil nitrogen content, with dosage recommendations made for each zone based on the available nutrients.
Brief description of the drawings
Figure:1 illustrates an example model that is configured to perform the functions described herein,
Detailed Description
The preferred embodiments of the present invention are solely used as examples in the pictures and the subsequent description.
A farmer can geotag (locate the farm on a map) the farm and obtain the soil nitrogen status of the farm using our technology in less than two minutes. The soil nitrogen status is supplied every 10 metres, and advice for the required nitrogen dosage are also provided. Information on the status of the soil nitrogen is supplied using satellite-based data obtained over a farm, a calibrated soil nitrogen model, and filtering and processing of the satellite data's spectral bands. The invention provides the report immediately and doesn't require any actual soil sampling or soil collection anywhere.
In one embodiment of the invention, farmer (100) first Geotagging of the farm through model and web application(101), Downloading and processing satellite data (102), Calibration of soil nitrogen model (103), Estimation of variable rate zones (104), Dosage recommendations(105) to farmer.
Geotagging of the farm through model and web application (101),First step in using our invention is to geolocate the farm through mobile application. Our innovation just requires the user/farmer to provide us the coordinates by entering the longitude and latitude or by simply walking around the farm and our application automatically picks up the coordinates of the walked boundary. The application is available to download from Google Play Store. Once a farm/s geotagged, users need to provide minimum necessary information on the crop for different services.
Downloading and processing satellite data(102)- One aspect of the invention starts downloading all the spectral bands of satellite data as soon as a farm or region has been geotagged. After the spectral bands have been downloaded, the bands are preprocessed (110) and filtered (111) to select the time when the land was barren. The data preprocessing such (112) as cloud removal, atmospheric correction are performed for the spectral bands along with soil nitrogen content is evaluated for the geotagged farm/region, and dosage recommendations are given for each zone and pixel.
Calibration of soil nitrogen model (103)-The machine learning model is continuously calibrated and validated by collecting the field data across the selected farms. More than 20,000 samples were collected globally to calibrate and validate the machine learning model. The entire field data is split into 70:30, with 70% for testing the model and 30% for calibration of the model. Once a model is calibrated, it is used to convert the satellite spectral band information into the available soil nitrogen (113). The field data is continuously collected to check the model for any discrepancy and if any is found, the model is updated for that particular condition.
Estimation of variable Soil zones (104)- Estimated soil nitrogen is provided every 10m. However, it may not be practical to apply fertilizer for every 10 meter spacing. Hence, for a large farm, zones are estimated to apply fertilizers zone wise rather than for every 10m. The quantitative information on soil nitrogen and qualitative information like high, low and medium is provided zone-wise as well. The farm/region is subdivided into multiple zones based on the available Nitrogen content with the help of contiguous clustering. A normal clustering approach may make discontinuous zones which may not be feasible. In the invention, contiguous clustering is applied to make contiguous zones which helps to fine tune fertility for specific areas. The Invention provides this through estimation, where farmers will be benefited by knowing the difference in nutrient supplying capability in the soil. Farmers can better match the yield and fertility to each zone.
Dosage recommendations (105)- The soil nutrient report also provides Nitrogen dosage recommendations like how much nitrogen is required by each crop as it plays an important role in the crop health and crop productivity. We also provide them with a quantity of basal dose to be added along with top dressing.
Through mobile application farmers receive soil reports instantly and recommendation is also provided for the next growing crop. We also have a web application which helps farmers and agri-enterprises by providing access to multiple plots at farm/regional-level information to monitor and compare the growth of the crops and to make the decision.
The model may comprises android software, IOS application etc.
In an example 1, the model system is programmed to facility the geotaging the farm area. After one or more field are geotagged, the spectral bands are downloaded from satellite data.
In an example embodiment, the agricultural intelligence computer system 130 is programmed to generate and cause displaying a graphical user interface comprising a data manager for data input. After one or more fields have been identified using the methods described above, the data manager may provide one or more graphical user interface widgets which when selected can identify changes to the field, soil, crops, tillage, or nutrient practices. The data manager may include a timeline view, a spreadsheet view, and/or one or more editable programs.
Table 1: Different location soil testing data
S.no Location Lab Data satellite remote sensing Error Accuracy (%)
1 Manchar, Pune (Location1) 273.62 276.49 2.87 98.95
2 Manchar, Pune (Location2) 247.71 267.35 19.64 92.10
3 Chikballapur 267.53 268.08 0.55 99.78
4 Chikballapur 282.87 276.35 6.25 97.70
5 Maharashtra 332.4 298.88 33.52 89.92
6 Khadimoli, MP 276 310 34 87.68
7 Kalapani, MP 284 277 7 97.54

Advantage of using our invention are:
1. No physical sampling required
2. Affordable for the farmers
3. Entire geotagged farm is estimated for soil nutrient analysis
4. Reports are received instantly.
5. A farmer can make the timely decision for the right kind/type and amount of fertilizer application based on our recommendation
6. 90-99 % accuracy results obtained which are very good for farming industry.
, Claims:We claim
1. A method to estimate the soil nitrogen contains comprising:
• first Geotagging of the farm through mobile application and web application(101),
• Downloading and processing satellite data (102),
• Calibration of soil nitrogen model (103),
• Estimation of variable soil zones (104),
• Dosage recommendations (105) to farmer.
Wherein farmers receive soil reports instantly on Mobile/web application and recommendation is also provided for the next growing crop.
2. The method to estimate the soil nitrogen contains as claimed in claim1 wherein farmer to geo locate his land by mobile/web location.
3. The method to estimate the soil nitrogen contains as claimed in claim1 wherein system downloading all the spectral bands of satellite data as soon as a farm or region has been geotagged.
4. The method to estimate the soil nitrogen contains as claimed in claim1 wherein the entire field data is split into 70:30, with 70% for testing the model and 30% for calibration of the model.
5. The method to estimate the soil nitrogen contains as claimed in claim1 wherein estimated soil nitrogen is provided every 10m.
6. The method to estimate the soil nitrogen contains as claimed in claim1 wherein mobile application farmers receive soil reports instantly and recommendation is also provided for the next growing crop.
7. The method to estimate the soil nitrogen contains as claimed in claim1 wherein the model may comprises android software, IOS application etc.
8. The method to estimate the soil nitrogen contains as claimed in claim1 wherein soil nitrogen content estimated values are 90-95% accurate compared to lab results.