Description:ML AND IOT IN AGRICULTURE: PRECISION CROP RECOMMENDATION BASED ON RAINFALL

Technical Field
[0001] The embodiments herein generally relate to a method for ML and IOT in agriculture: precision crop recommendation based on rainfall.
Description of the Related Art
[0002] A machine learning algorithm will be used to create a system that can choose crops, water them on its own, and suggest fertilizers. There will be a system established to assist farmers in reducing labor hours, using less energy, and increasing productivity. In order to create an affordable smart farming module, a system utilizing IoT and ML is developed and published. The system employs cutting-edge techniques to increase the accuracy of the findings. With the aid of, which examines machine learning in the context of food production and agriculture, the potential of machine learning in precision agriculture is also investigated. Data is gathered and measured in relation to meteorological conditions and soil quality.
[0003] The dataset includes variables for temperature, humidity, pH of the soil, nutrients, and potassium. For those who lack literacy, this method is a great aid in determining which crop has to be sown. "The application of contemporary information technologies to provide, process, and analyze multi-source data of high spatial and temporal resolution for decision making and operations in the management of crop production" is the definition of precision agriculture. Increased productivity, soil degradation, efficient water use, decreased use of chemicals for cultivation, and the use of contemporary farming techniques to raise crop quality, yield, and cost may all result from this precise agriculture. Additionally, the differences between several types of machine learning models are investigated. This paper covers the application of various machine learning algorithms in sensor data analytics within the agricultural ecosystem, explains why some machine learning models are more appropriate for use in agriculture than others, and provides a detailed assessment of the field. A case study of an IoT-based smart farm prototype that integrates food, energy, and water systems is also included. An emerging area of agriculture called "crop recommendation using IoT and machine learning" seeks to maximize crop output by utilizing automation and data analytics.
[0004] This method involves gathering information on soil moisture, temperature, humidity, and nutrient levels using sensors and Internet of Things devices. In order to make educated judgments regarding which crops to plant, when to plant them, and how much water and fertilizer to apply, this data is then evaluated using machine learning algorithms. Together, sensors provide a network that may be accessed or connected to a cloud or backend, allowing the cloud to connect sensor responses in various geographic locations. When it comes to smart items without connectivity, there are four stages. The network of things, or distributed control systems using programmable logic controllers, was the next development after local information exchange. For a smart irrigation system with temperature, humidity, and moisture as factors, a decision tree algorithm is suggested. Carries out a thorough examination of a number of machine learning ideas for an Internet of Things-based smart agriculture system. In order to select a crop that is appropriate for the available soil data and to enable improved crop growth through optimized farming practices, decision tree and support vector machine (SVM) algorithms are suggested.
SUMMARY
[0001] In view of the foregoing, an embodiment herein provides a method for ML and IOT in agriculture: precision crop recommendation based on rainfall. In some embodiments, wherein the second method, the automatic watering system, looks at the weather and the moisture level of the soil. The system then decides whether or not the plant needs to be watered based on its moisture requirements. In order to promote healthy growth, the third system, referred to as the fertilizer recommendation system, offers recommendations for the amount and timing of fertilizer applications to the crop. These three subsystems when combined into a single core system can enhance crop development. With 96.87% prediction accuracy, the system was able to determine the best time to plant maize as well as the quantity of water and fertilizer needed to get the highest yield.
[0002] An investigation into the application of remote sensing technology to gather information on crop health and soil moisture for a crop recommendation system based on machine learning algorithms was published in the Journal of Applied Remote Sensing. In addition to offering crucial information, online journals typically offer advice and fixes in case of an issue. Anticipating issues and deceptions that may result in severe penalties for failure is crucial.
[0003] In some embodiments, wherein the prior to examining the NPK and pH of the soil, this method measures the chosen location's temperature, humidity, and rainfall. While an NPK sensor is used to collect the NPK values, a pH sensor is used to collect the pH data. Then, the temperature, humidity, and precipitation data are collected via an open weather API provided by a cloud platform. The potential for enhanced crop production, lower input costs, and greater sustainability in agriculture through crop recommendation utilizing IoT and machine learning. More studies in this area may result in crop management techniques that are more effective and sustainable, which would be advantageous for both farmers and the environment. Tree category algorithm, CHAID, KNN, and function-based Naïve Bayes as machine learners are used in an ensemble model with voting technique to recommend crop. Several techniques utilizing contemporary techniques, such as crop recommender systems utilizing algorithms like ensemble-based models, KNN, neural networks, and similarity-based models, are employed.
[0004] In some embodiments, wherein a comparison would be made between the accuracy of five different machine learning models: RFC, GBC, DT, XGB, and KNN. The most suitable method will be used to communicate the values obtained from the sensor and cloud. This system would then recommend the best crop that could grow there. It is connected to a GPS module, push button, and LCD display; the request is received by the flask web framework from the IoT module; data is analyzed, and the crop recommendation is made using the KNN Machine Learning algorithm. In this project, the push button is used to activate the GPS module, which retrieves the specific field's geo location. The request is then routed to the linked Flask web framework server. Low-power sensors and a GPU that powers a NN-based AI system are integrated into the embedded system. The AI system's central component, the RNN cum LSTM, is guaranteed to function for 180 days when powered by a Li-ion battery. WSN, or integrated information technology in agriculture, is essential to the gathering, tracking, and analysis of data from the field.
[0005] These and other aspects of the embodiments herein will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions, while indicating preferred embodiments and numerous specific details thereof, are given by way of illustration and not of limitation. Many changes and modifications may be made within the scope of the embodiments herein without departing from the spirit thereof, and the embodiments herein include all such modifications.

BRIEF DESCRIPTION OF THE DRAWINGS
[0001] The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
[0002] FIG. 1 illustrates a method for ML and IOT in agriculture: precision crop recommendation based on rainfall according to an embodiment herein; and
[0003] FIG. 2 illustrates a method for client-server model suggested for the internet of things according to an embodiment herein.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
[0001] The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
[0002] FIG. 1 illustrates a method for ML and IOT in agriculture: precision crop recommendation based on rainfall according to an embodiment herein. In some embodiments, a moisture sensor will be used to measure the moisture content of the soil, and a cloud platform open weather API will be used to check the weather. Flow sensors will be used to determine how much water is added to the soil. Figure 3 provides a functional illustration of the system. This system is designed to track the temperature measurements every P hours. If there is no chance of rain and the soil moisture content is low, the mechanism checks the temperature.
[0003] At the first architectural level, facilities for data capture and communication are enabled. The base station and the gateway are connected to the sensor network. The algorithm module and classification specification are integrated at the second level. Putting the ML algorithms into practice to get the server's results is the next step. A trained module on the server retrieves the particular crop for irrigation. Analytical sensors are used to measure and store factors such as pH, phosphorus, potassium, and nitrogen in the module. Specific sensors are used to measure and record data such as temperature, humidity, and rainfall in a database. The link connects the base station and the gateway to the sensor network. The classification specification and the module for the classification method are both included in the second level of the system. The next step involves utilizing machine learning methods to retrieve the results from the server. The server is equipped with a trained module that can locate and retrieve the precise crop that requires further watering.
[0004] To monitor crop quality and yield evaluation, ML with computer vision is examined for the categorization of an alternative set of crop photos in. Moving on to the systems and solutions suggested by the study, a system for creating and evaluating a smart farming system based on an intelligent platform that enables prediction through the use of artificial intelligence (AI) techniques is presented. These technologies can assist farmers in making decisions by giving them precise and fast advice based on real-time data. All things considered, crop suggestion utilizing IoT and machine learning is an intriguing field of study that has the potential to completely change how we think about agriculture. Through the integration of data analytics, automation, and agricultural expertise, we can strive towards a more efficient and sustainable farming industry in the future.
[0005] In some embodiments, the fertility of the soil affects crop development. One major factor influencing soil fertility is the N, P, and K content of the soil. These numbers can be analyzed with an NPK sensor. Every crop would have a set of consistent NPK values. The land where crops are grown was analyzed using each of the values taken into consideration. The crop being grown and the information that has been collected are both analyzed. Crop irrigation is influenced by a variety of environmental variables as well as soil fertility, or the availability of nutrients like nitrogen, phosphorus, and humidity in the soil. A crop can be chosen for irrigation based on its seven features in order to maximize output. It is a collection of several devices learning facts-mining algorithms.
[0006] It's a percentage of a device that comprises various functions including preprocessing data, file preparation, clustering, regression, classification, Association rules, instance-based total classification, and picture taking. Building the body from the ground up the collection, processing, and application of data analytics are among the various processes that are part of the monitoring and control procedure. The main source for the datasheet that we used was the website Kaggle. In order to assess a crop for the purpose of irrigating it and obtaining optimal yield, advice regarding the crop is sought.
[0007] In some embodiments, the pH reading with the N, P, and K values. Next, the cloud is queried for the temperature, precipitation, and humidity readings for the city that was supplied as input. All of these values are then fed through the KNN machine learning model to identify the best crop that can be grown there. Oranges are the crop that is recommended for growing based on the parameters given. Rule-based classifier that classifies data using a straightforward decision table. This classifier is composed of a hierarchical table whose entries are subdivided into another table based on the values of two additional features. This is comparable to stacking dimensions. The growing phase and the pruning phase are the two fundamental stages. The growth phase is the rule by greedily adding highest information gain: p(log(p/t) - log(P/T)). Any remaining sequences are added to the growing phase during the pruning phase. The entire set of rules is fixed, together with the optimization stage and discretion length fixing. Many environmental factors as well as the fertility of the soil—which is gauged by the quantity of easily obtainable nutrients like nitrogen, phosphorus, and humidity—determine whether crops need to be irrigated. To select a crop for irrigation that will allow for the achievement of a maximum yield, apply the seven features.
[0008] FIG. 2 illustrates a method for client-server model suggested for the internet of things according to an embodiment herein. In some embodiments, the inputs for temperature and precipitation data are the city in which the crop is grown and the farmed crop itself. It is considered that the farmland is 1 m2 in area, and for every 1 s of water flow past the flow sensor, 1 mm of water deficiency is filled. The input data shows that in order to meet the water demands, 7.5 seconds of total pump time are required. Following data collection, the dataset is formatted into a comma-delimited excel file that can be used to train the dataset in the WEKA tool using supervised learning techniques. Once preprocessing is finished, the chosen classifiers are classified, and their performance metrics are recorded and tabulated. Following the collection of data, the dataset is converted into a CSV file comma-delimited excel file, which may be used to train the data set using supervised learning techniques in the WEKA tool. The dataset is then prepared for use following this phase. Following the completion of the preprocessing phase, the selected classifiers will undergo the classification process, following which the performance characteristics will be recorded and tallied.
[0009] In some embodiments, a system that can automatically water plants when needed and suggest the best fertilizers based on crop type has been developed. It can also recommend crops based on the soil and climate in the area. The accuracy performance percentage of the three classifiers—function-based, rules-based, and multilayer perceptron—that we used to apply the machine learning technique was 88.5909% for the JRip classifier and 98.2273% for the multilayer perceptron. The three classifiers—the multilayer perceptron, the decision table, and the JRip—have demonstrated extremely low error and RMSE error inside the 0.1384–0.058 range. The JRip, the decision table, and the multilayer perceptron are the three different classifiers whose performances have shown remarkably minimal mistakes and RMSE errors in the 0.1384 to 0.058 range.
[0010] The next set of characteristics that must be taken into account when building the model are the weighted receiver operator characteristics 1. In some embodiments, For optimal results, the approach can be used on the following crops: cotton, coffee, bananas, oranges, rice, corn, grapes, and rice. Farmers may find this technology to be a useful tool for making decisions that will decrease labor-intensive tasks, save energy, and boost productivity. By preprocessing the data set through normalization, the model construction process takes less time, and the ROC measure is improved to the point that nearly all classifiers roughly display the same measure. The model's performance yielded an accuracy percentage of 88.59% for the Lazy Category Decision Table Classifier and 98.2273% for the Multilayer Perceptron.
[0011] The second iteration of the model clearly shows that preparing the data set in the form of normalization can reduce the amount of time required to develop the model. In addition, there is nitrogen, phosphorus, and potassium. In not too distant a time, the agriculture sector will become smart agriculture and never again face a decrease in productivity, yield, or quality. Consequently, the agriculture sector will see a shift to precision farming based on artificial intelligence and the Internet of Things.

, Claims:I/We Claim:
1. A method for ML and IOT in agriculture: precision crop recommendation based on rainfall, wherein the ML and IOT in agriculture method comprises;
the technologies enable people to make informed decisions and assist businesses in modifying their operations by providing current information on variations in weather;
the potential to optimize resource usage, lower input costs, and raise crop yields while preserving sustainability, crop recommendation using IoT and machine learning has significant promise for revolutionizing the agricultural sector;
combining machine learning algorithms with data from Internet of Things sensors, like soil moisture and temperature sensors, crop recommendation systems can give farmers access to real-time information about crop health and environmental conditions, empowering them to make well-informed decisions about crop management techniques;
the Internet of Things, the trending machine learning algorithm has enabled us to develop a model for the agriculture sector. This model assists farmers in selecting the crop with the highest yield by measuring essential parameters such as temperature, pH, humidity, rainfall, phosphorus, and potassium;
obtaining remote sensing information at time t from one or more high spatial resolution remote sensors, where the imagery information consists of values related to multiple pixels, each of which is equal to or smaller than 30 m×30 m; and
utilizing a remote sensing derived index to calculate a high-resolution Kc (HRKc) value for each pixel, where the pixel is entirely included within the borders of at least one subplot; this process yields a multiplicity of HRKc values.