Smart Agri Assist App

Smart Agri Assist App

Ms. Sakshi Singh Department of Computer Science &

Engineering Moradabad Institute of Technology Moradabad, India er.sakshi.singh911@gmail.com

Vanshika Agarwal Department of Computer Science &

Engineering Moradabad Institute of Technology Moradabad, India vanshika.agarwal1604@gmail.com

Sanchai Bhatnagar Department of Computer Science &

Engineering Moradabad Institute of Technology Moradabad, India sanchaibhatnagar@gmail.com

Shruti Chauhan Department of Computer Science &

Engineering Moradabad Institute of Technology Moradabad, India shrutichauhan0243@gmail.com

Yash Kumar Gupta Department of Computer Science &

Engineering Moradabad Institute of Technology Moradabad, India

yashgupta3074@gmail.com

Abstract –

This abstract presents Smart Agri Assist, a voice-based app that supports farmers with daily decisions regarding farming and is a companion to them, designed specifically for agricultural use. In many rural areas, farmers face difficulties in accessing agriculture experts, digital resources, or stable internet connection. Because of this, getting accurate farming guidance becomes difficult. The Smart Agri Assist application uses different technologies to help farmers perform their daily farming activities more efficiently. It uses Speech Recognition to listen to farmers and Text to Speech to speak to them in their preferred language. It also provides Smart Recommendations that helps farmers in better farming practices. Through this approach, farmers can receive useful information such as weather updates, fertilizer planning, and other crop related guidance by speaking to the application, without requiring any technical expertise. This application has been developed using the Python programming language along with lightweight open- source libraries such as SpeechRecognition, pyttsx3, datetime, OS and Web Browser. These libraries provide the means by which the application is accessible on lower-end devices within the bounds of multiple languages, and works partially offline, allowing users, in several rural environments, to use this innovative tool. Ultimately, this application addresses the difference between modern agricultural technologies and the needs of farmers to support their ability to make improvements to their agricultural output.

This app provides more than just general farming information; it provides specific advice based on where you live (region), the type of crops you grow (seasons), what kind of soil you use (soil type/quality), etc. It also has user-friendly features designed for farmers not very comfortable with computers or technology, like clear instructions, a step-by-step breakdown, and easy navigation through the app's features. Farmers will receive useful information on when to plant their crops (so they will not lose them), help improve the amount of produce each year (increasing yield), help them feel confident enough to plant and grow their crops, and help them to take action on their own

schedule (to maximize productivity). The Smart Agri Assist App is more than just a tool; it is a huge leap forward in improving the industry of agriculture and using AI-based tools within agriculture to assist rural farmers within the agricultural community.

Key WordsAgriculture, Artificial Intelligence, Mobile App, Crop Disease Detection, Smart Agriculture, Deep Learning, Market Forecasting.

1. INTRODUCTION

   Agriculture plays a vital role in the Indian economy. A large portion of the rural population depends on agriculture for their income, food, and overall livelihood. However, farmers face several challenges such as unpredictable weather conditions, pest attacks, crop disease, fluctuating market prices and difficulty in getting timely expert advice. Many farmers are also unaware of various government programs and schemes available to them. There are currently many different mobile applications and digital resources for farmers, but most of them come with only a few features. For example, there are apps that provide the weather, some that provide market prices, and some that provide information about crops. In order for a farmer to access all of the necessary information they would have to download and use multiple apps. The constant switching between different apps creates difficulties for farmers due to confusion in using many apps. They wasted time and energy in switching between apps. For farmers in rural areas, who are unaware with technology due to poor internet connectivity, having one place to access all required information would be a more appropriate option than using several other complex and hard to use applications. [1], [2],[4]

   The Smart Agri Assist App was developed to support farmers in dealing with the difficulties that arise in modern agricultural practices. The goal of the Smart Agri Assist App is to provide a simple and user- friendly digital platform that farmers can use as a single solution to all their agricultural needs instead of using multiple applications for different types of information. Smart Agri Assist combines various technologies such as Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL), as well as other advanced API-based services to provide farmers with timely and accurate information as well as personalized support. The Smart Agri Assist App help farmers to search for information related to crop management, receive notifications of pest infestations, and send photos to healthcare specialists for reviewing potential diseases. It also provides real-time market price information and updates about available government schemes and programs. The Smart Agri Assist App also provides farmers with tools that will allow them to keep track of pest infestation occurrences and communicate with other users via an in-app message system. Thus, using a combination of all of the above-mentioned technologies and functionality will ultimately provide farmers with the resources needed for improving their production capabilities and ultimately increasing the profitability of their agricultural business.[3]

2. RELATED WORK

   Digital agricultural solutions have flooded the Market over the last Ten years and have provided different tools to assist farmers however; none have provided the complete solution to all farmer needs. Apps developed by the Government such as Kisan Suvidha offer weather and market price data to the farmer but do not offer personalized advice or provide crop or region-specific recommendations. IFFCO Kisan offers the ability for the farmer to receive voice-based assistance; however, it does not provide artificial intelligence solutions, crop image analysis, or medical diagnosis and assessments. AgroStar provides a solution for farmers to purchase agricultural inputs but does not provide the complete decision support system to the farmer. Studies have recently shown that Artificial intelligence technologies are growing and improving agriculture. Studies that have been done recently on CNN (Convolutional Neural Network) based plant disease classification systems demonstrate success in making timely and accurate diagnoses.[5], [6]

   and have shown that machine learning-powered advisory systems of plant disease diagnosis over the last few years.[2]

   Other agricultural technology market forecast studies as well as also environmental data modelling-based studies have reiterated the role of predictive analytics and the importance of predictive analytics models in agriculture.

   Despite all of the advancements in agricultural technology, all of the current technology operates in isolation of each otherand currently there are not any software platforms that integrate diagnosis, advisory services, market predictions, as well as multilingual farmer educational services together into one complete solution.

3. PROBLEM STATEMENT

   Today, there are many tools available for farmers, but most of them dont help address their individual needs. Most of these applications have only the basics (such as weather, market pricing, and simple information related to growing crops). They are not tailored to provide specific assistance. So, farmers need to switch between multiple applications to get relevant information from them all (e.g. confusion, too long, hard) and even more so for rural farmers who suck with using modern technology. Further, most existing tools for farming do not provide farmers with the ability to get any support based on their individual situations, nor will they provide them with anything real-time (i.e. immediate assistance) for a better decision-making process. Other than the weather, market price, and growth-related info, there are many other relevant factors that need to be accounted for when making a decision (e.g. seasons, bugs, and market fluctuation). Many of the existing tools do not consider specific conditions that affect farming decisions, that often use poor advice that may not be very effective. Important factors such as seasonal changes, pest attacks, and market fluctuations are sometimes ignored when providing recommendations. Because of this, the information provided by some tools may not always be useful for farmers in real situations. Another limitation is that many agricultural applications support only one or two languages, and some of them have limited voice features. Due to these restrictions, a large number of farmers find such applications difficult to use or understand.

   Due to the limitations of many existing farming tools, there is a clear need for a single mobile application that can bring multiple agricultural services together in one place. This platform makes it easier for farmers to access the information and provide support without depending on various different applications. The application should provide crop disease detection services, market price forecasting, farming recommendations, risk alerts, and community support. Bringing all these services together at one place can help farmers access useful information more easily. A mobile application that uses machine learning, deep learning, and reliable API technologies can link cutting-edge technology to real-life agricultural issues. The application will provide information to farmers at the right time in a format that is immediately usable in their day-to-day operations. The result will be improved decision-making by farmers, higher levels of productivity, and less risk.

4. PROPOSED SYSTEM

   The new Smart Agri-Assistance app is a digital agricultural platform that is smart, scalable, simple to use, and allows for multiple uses by Farmers. The Smart Agri-Assist App supports Farmers in a wide range of areas including: decision- making; disease diagnosis; market planning; and risk management. As such, the Smart Agri-Assist App is not only a single-purpose tool, but is also a Multi-Use Tool for the Farmer. The Smart Agri-Assist App does not contain any hardware, such as sensors or IoT devices, and is entirely a software-based application that uses AI and machine learning algorithms, APIs, and a backend system to collect, store, and process data. By combining these elements together, Farmers will have access to real-time, accurate, and automated Agricultural support.[8]

   1. System Architecture

      The Agri-Assist Smart Agriculture Support System has a modular, layered architecture designed for performance, scalability, and ease of integration. It contains four distinct layers (levels) called components:

      1. Frontend Layer (User Interface)

         • Technology: Kotlin

         • Purpose: Provides a smooth, cross-platform mobile interface for farmers.

         • Features:

           • Multilingual UI (Hindi, English, regional languages).

           • Voice first interaction using speech-to-text and text-to-speech.

           • Offline caching for poor network regions.

           • Intuitive dashboards, crop advisory cards, market price tables, and weather alert screens.

      2. Backend Layer (Application Server)

         • Technology: Django or Node.js

         • Responsibilities:

           • Handles API endpoints for disease detection, advisory generation, market data retrieval, and alert notifications. o Manages user authentication, session handling, and request routing.

           • Implements business logic for recommendations, community discussions, and record management.

      3. Database Layer

         • Database: PostgreSQL or SQLite

         • Purpose: Stores structured data such as:

           • User profiles, Crop disease datasets, Farming diary entries, Community forum discussions.

           • Alerts, notifications, and historical weather data.

             The application uses a relational database such as PostgreSQL or SQLite to store and manage the data which is required for the system. By organizing the information in a structured way, the database helps to fetch the data quickly.

      4. AI/ML Layer (Model Pipeline)

         A combination of supervised and deep learning models powers the intelligent automation in the app:

         • CNN (Convolutional Neural Networks): For crop disease classification using leaf images.

         • Regression Models: To predict fertilizer needs, yield estimation, and price forecasting.

         • Decision Trees / Random Forest: For generating rule-based advisory on irrigation, nutrients, pesticides, etc.

         • LSTM/GRU Models: For time-series tasks such as rainfall prediction and crop growth forecasting.

           Fig. 1: System Architecture

   2. Functional Modules

      The application is divided into four highly cohesive and closely coupled modules, each focusing on a core agricultural problem.

      1. Crop Health & Advisory Module

         • Uses CNN to detect crop diseases from leaf images.

         • Provides: Fertilizer recommendation, Irrigation scheduling, Pest and disease prevention tips, Growth stage-based advisory.

         • Includes a knowledge base for major crops like wheat, rice, maize, sugarcane, etc.

      2. Market & Financial Support Module

         • Integrates an automated agricultural market price data pipeline that periodically collects mandi price data from official government agricultural datasets.

         • A scheduled data ingestion system retrieves market price records and stores them in a PostgreSQL database, ensuring structured and consistent storage of crop price information.

           The database is automatically updated through a weekly automated process, which fetches the latest mandi price data and inserts only new records while preventing duplicates using database constraints.

         • Displays: Top selling mandis, Price trends

           across nearby markets, and comparative crop prices, helping in better selling decisions.

         • Recommends suitable government schemes, subsidies, loans, and crop insurance options.

      3. Communication & Community Platform

         • Voice enabled interaction for low-literacy farmers.

         • Multilingual chat and discussion forums for farmer-to-farmer knowledge exchange.

         • Option for experts to answer queries, share tips, or conduct live Q&A sessions.

      4. Risk Management & Digital Farming Diary. Data Collection

         • Real-time weather alerts (rain, storms, humidity, temperature).

         • Pest outbreak mapping using ML trends and historical data.

         • A digital farming diary to record: Sowing dates, Fertilizer application, Irrigation details, Harvest records.

   Agricultural Challenge	Solution Provided by Smart Agri Assist
   Farmers often find difficult to identify crop diseases at an early stage.	The application uses a CNN-based image analysis model that helps farmers detect crop diseases by simply uploading an image of the affected plant.
   Weather conditions mostly change and affect crop growth.	The system has a weather API that provides real- time weather updates to help farmers.
   Farmers are not sure about the right time and place to sell their crops.	The app provides market price information through the automated market price which helps farmers to make better decisions.
   Farmers find difficulty to access to agricultural experts.	The platform provides an advisory module and community support feature where farmers can take help and share experiences.
   Many farmers are unable to use complex digital tools.	The application provides simple, easy-to-use voice- based feature that helps farmers with limited technical knowledge.

   Table 1. Agricultural Challenges and Proposed Solutions in Smart Agri Assist

5. ER DIAGRAM AND DATABASE DESIGN

   The database schema for Smart Agri-Assist is designed to maintain clean relationships between entities while ensuring efficient data retrieval.

   Major Entities:

   • User (Farmer/Expert)

   • Crop

   • Disease

   • Advisory

   • Market Price

   • Weather Alert Diary Record

   • Forum Post / Comment

     Relationships:

   • One user can grow multiple crops.

   • One crop can have multiple disease records.

   • One user can create multiple diary entries.

   • Forum posts support many-to-one and one-to- many relations.

   Fig. 2: ER diagram

6. METHODOLOGY

   The conceptual framework of the Smart Agri Assist system focuses on combining multiple components to help farmers in their daily agricultural activities. The system collects various types of data such as crop images, weather information and market price data from different sources. The information is used in machine learning techniques to detect crop disease and predict weather conditions and provide useful information to farmers.

   1. Data Collection & Digital Farming

      The Smart Agri Assist App makes use of different types of publicly available datasets obtained from Kaggle to provide useful agricultural information to farmers. These datasets include images of crops, information about the type of soil,

      climate and weather conditions, and historical market prices of the agricultural products. The combined dataset used for training and testing has an approximate size of 12.4 GB. The application can access information from various publicly available agricultural databases and government resources to provide useful insights for users by combining these resources. The data goes through several preprocessing steps before used in system in order to improve the accuracy and efficiency of the model. Different preprocessing methods can be applied depending on the type of data being used. For example, for images of crops, pre-processing can include resizing, normalizing, and augmenting images to create better quality images with respect to other forms of data inputted into the model. In the case of numerical datasets, pre-processing includes cleaning and standardizing data, as well as formatting data in a consistent way, so that correct and comparable input is provided into the model. All of this will improve the quality of predictions made by the model.[9], [10]

   2. Model Training

      This application uses a CNN (convolutional neural network) model to classify diseases based on images. CNNs use many different types of feature detection (for example: patterns, shapes, colors, and textures) to determine what characteristics of the disease can be recognized in the images

      – for example, leaf color, leaf texture and leaf pattern of infection – and is accomplished mostly by using the convolutional and pooling layers of the model.

      The application also uses LSTMs (long short-term memory) to analyze past weather and market data to build predictions for future weather conditions and because of the LSTM's ability to remember past data, the user can make more informed decisions about the crops they are growing and the soil characteristics of the farm.

      The use of the above-mentioned algorithms and techniques will allow the application to generate highly personalized recommendations for each farmer, based on their specific growing conditions.

   3. Workflow

   Smart Agri-Assist App functions through an intelligent workflow system consisting of a series of interconnected processes where AI is used for image processing, weather, and market data, multiple language support, all used to produce an output. The system is designed with farmers in mind and gives farmers the most accurate, timely, and actionable advice with the least amount needed from the farmer. Each of the individual processes that comprise the intelligent workflow are described below and are represented

   in the overall schematic diagram (Fig. 3) showing all of the interconnectivity within the system as part of the intelligent workflow.[7]

   Image Capture and CNN-Based Processing

   The workflow commences when the farmer uploads a picture of a crops leaf or plant via the app. This is how the apps diagnosis capabilities start functioning to diagnose any problems with the crop. The uploaded photo will automatically be sent from the farmers device to a backend server. There is a Convolutional Neural Network (CNN) model on the backend that will be responsible for completing this task.

   The dataset used to train the CNN model contains a collection of images of the crop, including that contain all classifications of disease, pest damage, and nutrient deficiency. Once the image has been uploaded, the CNN model will perform its own image analysis by extracting visual features, determining if there is any abnormality from normal, and finally classifying that crop as having either a fungal infection, bacterial disease, insect infestation, or environmental stress.

   By completing the image analysis, the CNN model can quickly and efficiently provide farmers, particularly in remote locations, with the equivalent of an expert-level diagnosis without the need for them to complete their own visual inspection.

   1. Weather Data Retrieval via API Integration

      As the user analyzes their crop images, there will be a request generated to the Weather API to obtain time- sensitive weather condition information specific to their location and or region. Some of these weather condition(s) will most likely include the following types of atmospheric data:

      • Temperature levels

      • Humidity percentage

      • Rainfall forecasts

      • Wind speed and direction

      • Potential extreme weather alerts

        These elements have a major impact on the development of a disease, irrigation system design, and the effectiveness of fertilizer. For instance, high humidity raises the liklihood of fungal outbreaks, while the potential for future rain would alter a recommendation for how much irrigation or pesticide to use.[11]

        Live weather data are integrated to make the app's advisory specific to the current situation, not just a generic advisory. The app utilizes real-time weather

        data to ensure that the advisory provided to users is not only highly pertinent, but also reflects the current and future weather conditions.

   2. Advisory Engine: Decision Intelligence Layer

      The Smart Agri-Assist applications use the Advisory Engine as a central decision maker. It integrates multiple input sources to create a scientifically based plan that is customized to the farmer's needs.

      This engine processes:

      • The CNN model's diagnosis of crop health

      • Real-time weather parameters

      • Agronomic rules embedded in the system

      • Seasonal patterns, best practices, and ML-driven logic

        By combining these factors, the advisory engine generates tailored recommendations such as:

      • Optimal irrigation schedules adjusted for current soil moisture or predicted rainfall

      • Fertilizer dosage and timing which is based on crop type and detected deficiencies

      • Disease management strategies, including both organic and chemical treatment options

      • Preventive measures to reduce the risk of future outbreaks

        The advisory is carefully structured to balance scientific precision so that farmers can easily implement the recommendations effectively, irrespective of technical background.

   3. Market Insights via Agri-Market API

      Similar to weather retrieval and advisory computation, the system establishes a real-time connection with the Market API. This component retrieves:

      • Latest mandi prices for various commodities

      • Price trends across nearby markets

      • Supply-demand fluctuations

      • Recent government notifications affecting prices

        This information provides farmers with high-value market intelligence and helps them to make informed decisions on:

      • When to harvest

      • Where to sell for the best price

      • Whether to store produce temporarily based on expected trends

      • Which markets offer competitive rates

        Market data integration ensures that farmers are not solely reliant on middlemen or outdated pricing. It also helps in empowering them economically and helping them maximize income.

   4. Final Output Generation via Multilingual App Interface

      The images diagnosed, coupled with the weather data, market insight, and advisory recommendations will produce a single, combined set of data (a final result). The output contains:

      • Detected crop disease or health issue

      • Recommended treatment or preventive measures

      • Irrigation and fertilizer guidance

      • Real-time weather alerts and forecasts

      • Latest mandi and commodity prices

      • Actionable insights for the next few days

        The complete set of data will then be provided to farmers via a multilingual user interface, making it accessible to users from different language backgrounds.

        The app supports:

      • Text-based output for detailed reading

      • Voice-based output using Text-to-Speech for farmers with low literacy

      • Regional language support for better comprehension and ease of use

   This last step helps close the gap between the data inputted and the resulting actionable insights by providing comprehensive, user-friendly information.

   Fig. 3: Workflow Diagram

7. RESULTS AND DISCUSSION

   The results from the experimental study confirmed that the developed Smart Agri-Assist system was functional across all key components. In particular, the performance of the CNN-based disease detection model was both accurate and dependable. This allows for accurate identification of many kinds of plant disease regardless of lighting and background variability; thereby, demonstrating both the dependability of the image processing pipeline and the appropriateness of the selected model architecture.

   The advisory engine created by Smart Agri-Assist was relevant and contextually appropriate and validated as consistent with practiced agronomic best practices (e.g. crop management; irrigation scheduling; nutrient management). In addition, the LSTM forecasting module provided stable forecasts on weather trends and market fluctuations to assist farmers with their near-term decision-making.

   Usability testing is used to confirm that users within target segments feel favorable towards and will use the product. It is believed that the multilingual interface, voice recognition, and simplified navigation increased usability for farmers with limited technology experience.

   It was noted also that offline usage and lightweight data processing will provide a stable operation in poor communication networks due to their ability to adapt to rural settings.

   So, the findings indicate that the Smart Agri Assist software suite has the potential to solve many important problems in digital farming and to provide an accurate, integrated, and easy-to-use platform to increase productivity and support decision making at the farm.[13]

8. FUTURE SCOPE

   1. The Smart Agri-Assist system acts as a solid foundation on which to build a more intelligent ag- service system; however, there are many additional components that may be added to grow and optimize the present/more future improved efficiencies through further development/additional appropriations. Enhancements of this nature are anticipated to substantially increase the total systems accuracy, usability, and beneficial outcomes.

   2. Integration of Soil and Nutrient Analysis:

      Adding soil test data from internal sensors or image- based analysis of the soil's quality into an enhanced

      future release of the application would allow for more accurate fertilization recommendation(s)/optimal nutrient balance for the farmers.

   3. On-Device AI Model Deployment:

      With a lightweight CNN model locally on a mobile device, off-line disease detection would help farmers in remote locations without service or very poor internet connections.[12]

   4. Advanced Conversational Voice Assistant:

      Using an AI-based Conversational Assistant would permit farmers to ask questions naturally through the system about normal-paid issues like what to do with irrigation, precautionary measures concerning incoming rainy weather, when to fertilize and how to control pests.

   5. Smart Government Scheme and Policy Integration: Farmers will receive automatic alerts about pertinent government programs, subsidies, crop insurance options, loan qualifications, and agricultural policies that are specific to their area and crop type through the extension of this system.

   6. Enhanced Market Intelligence and Predictive Analytics:

      Future versions of the system could include predictive modelling of commodity prices as well as forecasting demand for commodities, providing farmers with well-founded methods to optimize the timing of crop sales and make data-based financial decisions.

   7. Disease Severity Estimation and Stage-wise Treatment:

      In addition to assessing crops, the system identifies the severity of disease or insect infestations and provides detailed models for treating disease or insect problems thus allowing farmers to protect their crops more efficiently.

   8. Farm Yield Forecasting:

      By leveraging historical data on crop perormance, soil condition and multiple years of climate history, the system generates yield forecasts, allowing farmers to effectively plan for the use of resources and their finances.

   9. IoT Integration for Smart Farming:

      Integration into the agricultural system with IoT- based sensors and devices for example; ground moisture sensors and temperature monitoring can allow a real-time tracking and autonomous decision making in farming.

   10. Community Knowledge Graph and Expert Network:

       A community forum can also develop into a structured knowledge graph to capture all the farmers experiences, expert recommendations and

       historical resolution of issues to build an ever- growing body of agricultural knowledge.

   11. Multi-Crop, Multi-Season Recommendation Engine:

   In addition, future improvement may be made to long-term planning through improved recommendations for optimal crop rotations, seeding windows, fertilizer cycles, and allocation strategies across multiple years.

9. CONCLUSION

   The Smart Agriculture – Assist app is a comprehensive decision-support tool for farmers, built on the integration of artificial intelligence with mobile app frameworks and real- time agricultural application programming interfaces (APIs). This unique platform addresses the limitations of existing fragmented solutions within agriculture by creating one platform where farmers can manage the disease detection of plants, receive market intelligence, know the forecast of the weather, and interact with the system in multiple languages using voice/keywords to help them make informed decisions, provide flexibility in adapting to agricultural challenges, and give the farmer timely, customized, and easily accessible guidance. This proposed solution has strong potential to improve productivity, reduce information disparity and support digital empowerment within the rural farming sector.

10. FUTURE WORK

• The accuracy of model can be improved by expanding datasets to include region specific crops, different soil types, and less common plant diseases.

• Natural Language Processing (NLP) features can be enhanced so that the system can better understand rural dialects and common speech patterns used by farmers.

• Pest detection and crop yield prediction can be improved by applying advanced deep learning models.

• A community-based knowledge sharing feature can also be added for localized problem solving and allowing farmers to share experiences,

• Implementing real-time satellite or drone data integration to increase the analytical capabilities of the system.

ACKNOWLEDGMENTS

We would like to express our sincere appreciation to the Moradabad Institute of Technology (MIT), Moradabad, for giving us the invaluable opportunity to pursue this project. The academic atmosphere of the Institute, along with its technical facilities and continuous encouragement have all contributed greatly to improving our practical knowledge and skills while developing the Smart Agri-Assist – AI Based Personalized Farming Assistant project.

We extend our heartfelt gratitude to our Guide, Ms. Sakshi Singh, for her continuous support and valuable assistance as well as her regular feedback and supervision throughout the development process. In addition, we are grateful to Prof. (Dr.) Rohit Garg, Prof. (Dr.) Manish Gupta and the other members of the Computer Science and Engineering Faculty at MIT for providing guidance and motivation when developing the Smart Agri-Assist Project and the numerous suggestions that helped shape the overall direction and quality of the work.

Lastly, we would like to express our most sincere gratitude to our families and friends who provided us with continued support and encouragement (in addition to moral support) during the development and training periods, and whose belief in us has provided us with strength and motivation.

REFERENCES

1. R. Sharma et al., AI-based Crop Disease Detection Using CNN, IEEE Access, 2022.

2. A. Kumar, Voice-Enabled Smart Farming Solutions for Indian Agriculture, Elsevier, 2021.

3. National Informatics Centre (NIC), Agricultural Market Price Data Portal, Government of India, 2023.

4. S. Patel and V. Singh, Deep Learning Models for Agricultural Pest Identification, IEEE Transactions on Computational Agriculture, 2023.

5. M. Jadhav and P. Deshmukh, Mobile-Based Decision Support Systems for Small-Scale Farmers, International Journal of Smart Computing, 2021.

6. L. Gupta and R. Verma, LSTM-Based Weather Forecasting for Crop Yield Optimization, IEEE Sensors Journal, 2022.

7. T. Chakraborty et al., A Review on AI-Driven Smart Farming Technologies, Journal of Agricultural Informatics, 2020.

8. World Meteorological Organization (WMO), Weather Data Standards and API Services, 2023.

9. S. Bose and K. Rajan, Design of User-Centric Mobile Applications for Rural Communities, ACM Computing Surveys, 2022.

10. P. S. Reddy, Machine Learning Approaches for Market Price Prediction in Agriculture, Springer Nature, 2021.

11. A. Sharma and P. Gupta, Deep Learning Based Crop Disease Detection Using Convolutional Neural Networks, IEEE Access, vol. 11, pp. 115432-115445, 2023.

12. S. Patel, R. Kumar and V. Singh, IOT and AI Driven Smart Farming Framework for Precision Agriculture, Computers and Electronics in Agriculture, vol. 208, 2023.

13. M. Verma and P. Agarwal, Machine Learning Based Decision Support System for Smart Agriculture, International Journal of Agricultural Informatics, vol. 24, no. 3, pp. 2231-2240, 2024