A Mobile Application for Crop Optmisation using Satellite Imagery and Disease Detection using AI

Abstract

Despite being crucial to the world\'s food supply, small-scale farmers do not have access to resources or contemporary technologies. This makes it more difficult to make educated decisions, which lowers yields and causes financial hardship. To help farmers overcome this obstacle, our research uses satellite data to provide them with useful insights that will enable them to increase production and make wise decisions. We suggest a smartphone app that uses image-based disease diagnosis and artificial intelligence (AI) to transform precision agriculture. Using real-time data on plant disease, water use, and soil health, the app helps farmers reduce risks, maximize resource use, and eventually increase yields. Our approach helps small-scale farmers close the technology gap by converting their customary techniques into a more effective and sustainable model. We think there is a lot of potential for this technology to improve socioeconomic conditions and the standard of living for farmers throughout the world.

Introduction

I. INTRODUCTION

Significant changes in the global agricultural environment are being driven by developing diseases and pests, technology advancements, and climate change [1].

The population of the world is predicted to rise quickly, surpassing 9 billion people by the year 2050. [2] Small-scale farmers, who still make up a sizable share of the agricultural labour, frequently lack access to precision agriculture instruments, even as large-scale farms are increasingly using similar practices [3]. This discrepancy may result in less-than-ideal crop management techniques, lower yields, and unstable finances [4].

Mobile applications have the ability to close the technology divide in agriculture, according to recent research [5]. In order to provide small-scale farmers with practical insights, this article suggests creating an intuitive smartphone application that makes use of satellite data and artificial intelligence (AI). The app will provide functionalities such as:

A. Precision Agricultural Recommendations

Utilizing AI algorithms, the app will analyse satellite imagery and local weather data to provide farmers with site-specific recommendations for the crop, such as health of the crop and what action to take [1].

???????B. Image based Plant Disease Detection

Farmers will be able to recognize any plant illnesses by using real-time image capture and analysis with the app thanks to its integration of image recognition technologies [5]. Early discovery can direct suitable treatment and help reduce crop losses.

The developed smartphone application aims to enhance agricultural yields by giving small-scale farmers access to these technologies. Research has shown that precision agriculture methods can optimize resource allocation, leading to considerable increases in crop yields [3].

Enhance Decision-Making: Farmers will be better equipped to choose their crops thanks to real-time data access and AI-powered insights. Boost Socio-Economic Well-Being: Small-scale farmers may earn more money and have a more stable living if crop yields are raised and farm management techniques are strengthened [2].

II. LITERATURE REVIEW

This section discusses methodologies that have existed to handle the difficulties of detecting diseases and crop monitoring, with an emphasis on existing mobile apps.

???????A. Plant Disease Detection Methodologies

Experienced farmers, known for their expertise in agriculture, are often relied upon to diagnose pests and plant diseases on-site. However, this traditional method is arduous, time-consuming, and subjective, often resulting in limited success. Furthermore, inexperienced farmers may resort to indiscriminate pesticide use and erroneous judgments during the identification process, leading to compromised output, reduced quality, environmental damage, and financial losses. To address these challenges, leveraging images for disease identification has emerged as a promising solution [10].

The application of deep learning algorithms for plant disease identification represents a significant advancement in agricultural technology [7]. Initially, researchers focused on utilizing leaf vein patterns in plant photos for identification purposes. Dubey and Jalal, employed an enhanced support vector machine (SVM) combined with K-means clustering to successfully identify three types of apple diseases, achieving an impressive classification accuracy of 93% [6].

According to current research, Convolutional Neural Networks (CNNs) stand out as the most widely utilized classifiers for image recognition, demonstrating exceptional performance in image processing and classification [8]. CNNs leverage large datasets for training, enabling them to provide accurate results. A CNN-trained model using the Plant Village Dataset has been made accessible to mobile applications through platforms like Kaggle. This accessibility democratizes advanced disease identification technology, making it available to a wider audience [9].

???????B. Satellite image processing techniques

The advent of Earth observation satellites has opened up new avenues in agriculture, such as crop mapping using satellite imagery. Many applications for crop monitoring emphasize the use of machine learning classification algorithms with publicly available image sources like MODIS, Landsat, and Sentinel. With just a single photograph, it's now possible to obtain yearly crop maps or crop cycle maps, revolutionizing agricultural decision-making [11]. SVM and Random Forest classifications from the Sentinel satellite were used by R. Saini and S. K. Ghosh to classify crop kinds with an overall accuracy of 82% to 91% utilizing a single picture [12]Afterwards, to estimate and offer insights, a crop cycle's worth of photos is taken into consideration. The yearly crop maps used to build the multi-year cropping patterns were created by Martínez et al using a sequence of photos. A seven-year time-series (1993, 1994, 1996, 1997, 1998, 1999, and 2000) of crop maps made from Landsat 5 TM and Landsat 7 ETM+ images was used. [13]

To process crop indices with ease, the mobile application makes use of Agromonitoring API, a sentinel satellite agricultural insight provider. [14]

III. PROPOSED SYSTEM

The mobile application is the end-user experience , the app works with dual functionality mainly focusing to retrieve satellite data from crop monitoring API and local image processing for plant disease detection.

The mobile app segregated into individual modules to deliver scalable usage. Each module provides a use case to the farmer to get insights and suggestions.

???????A. Authentication - Firebase

To securely personalize data, farmers can use the authentication function offered by the application. We immediately register the user in Firebase (a cloud storage and database hosted by Google) using their Google Account.

???????B. Cloud Data Storage

For an improved user experience Personal information including the user's name, phone number, and crop map coordinates for location are kept in a different cloud cluster in Mongo DB. The application connects to the database via a node.js server. This allows scalability and reduces server load for authentication.

???????C. Agro Monitoring API

The geographic coordinates are the input to the Agro Monitoring API to obtain crop data. It creates a polygon with the selected co-ordinates. The interface's design presents the user with a Google Maps screen where they may choose four coordinates and form a polygon. This polygon is later analysed and produces fair results of various vegetation indexes, a custom algorithm is used to convert these indexes into actionable results for farmers.

???????

Conclusion

The mobile application discussed in this research paper serves as a pivotal bridge between traditional farming practices and modern technological solutions, particularly in the realm of precision agriculture. By seamlessly integrating features for precision farming and plant disease identification, coupled with actionable suggestions, it empowers farmers to make informed decisions to enhance their crop yield. This innovative tool not only addresses existing challenges in agriculture but also exemplifies the transformative potential of technology in facilitating sustainable and efficient farming practices.

References

[1] E. Omia et al., “Remote Sensing in Field Crop Monitoring: A Comprehensive Review of Sensor Systems, Data Analyses and Recent Advances,” Remote sensing (Basel), Jan. 06, 2023. https://www.mdpi.com/2072-4292/15/2/354 [2] “The State of Food and Agriculture 2017 | FAO.” https://www.fao.org/family-farming/detail/en/c/1043688/ [3] “The Precision Farming Revolution,” SpringerLink. https://link.springer.com/book/10.1007/978-981-13-9686-1 [4] A. Monteiro and S. R. D. Santos, “Sustainable Approach to Weed Management: The Role of Precision Weed Management,” Agronomy (Basel), Jan. 04, 2022. https://www.mdpi.com/2073-4395/12/1/118Mouazen, [5] H. Baummller, “Mobile Technology Trends and their Potential for Agricultural Development,” Social Science Research Network, Jan. 01, 2013. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2359465 Bandyopadhyay, [6] S. R. Dubey and A. S. Jalal, “Adapted Approach for Fruit Disease Identification using Images,” IGI Global eBooks, Jun. 17, 2013. https://www.igi-global.com/chapter/content/77603 [7] G. L. Grinblat, L. C. Uzal, M. G. Larese and P. M. Granitto, \"Deep learning for plant identification using vein morphological patterns\", Comput. Electron. Agricult., vol. 127, pp. 418-424, Sep. 2016. [8] A. Kamilaris and F. X. Prenafeta-Boldú, \"Deep learning in agriculture: A survey\", Comput. Electron. Agricult., vol. 147, pp. 70-90, Apr. 2018. [9] “PlantVillage Dataset,” Kaggle, Oct. 30, 2018. https://www.kaggle.com/datasets/emmarex/plantdisease [10] J. Li and X. Wang, “Plant diseases and pests detection based on deep learning: a review,” Plant methods, Feb. 24, 2021. https://plantmethods.biomedcentral.com/articles/10.1186/s13007-021-00722-9 [11] C. Yang, J. H. Everitt, and D. Murden, “Evaluating high resolution SPOT 5 satellite imagery for crop identification,” Computers and Electronics in Agriculture, Feb. 01, 2011. https://www.sciencedirect.com/science/article/abs/pii/S0168169910002632 [12] R. Saini and S. K. Ghosh, “CROP CLASSIFICATION ON SINGLE DATE SENTINEL-2 IMAGERY USING RANDOM FOREST AND SUPPOR VECTOR MACHINE,” The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Nov. 19, 2018. https://isprs-archives.copernicus.org/articles/XLII-5/683/2018/ [13] J. A. Martínez?Casasnovas, A. Martín-Montero, and M. A. Casterad, “Mapping multi-year cropping patterns in small irrigation districts from time-series analysis of Landsat TM images,” European journal of agronomy, Sep. 01, 2005. https://www.sciencedirect.com/science/article/abs/pii/S1161030104001212 [14] “Agromonitoring - Agro API Tech Documentation.” https://agromonitoring.com/api#satimg

Copyright

Copyright © 2024 Arun Adithya Zuturu, Mohammed Kaamil Mirza, Tanishq Agarwal. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.