Overview / Usage

This project aims to revolutionize the way we detect diseases in tomato plants.This project leverages the power of Convolutional Neural Networks (CNNs) to accurately detect diseases in tomato leaves from images.This model simply selects an uploaded image of a tomato leaf and outputs a decently accurate prediction of any disease present.The importance of staying up-to-date on the market trends of tomato production in today's world is crucial.Therefore, this project also includes a visualization of tomato markets in major cities in Karnataka. This feature provides valuable insights into the tomato industry and helps farmers make informed decisions about their crops.Due to the unavailability of a dataset that isolates Karnataka or tomato varieties in fertiliser recommendations the project provides a general fertiliser recommendation system.

IntelOneAPI was used in the project. The Intel oneAPI contributes to the acceleration of computation here by providing highly optimised algorithms for all phases of the project (preprocessing, transformation, analysis, modelling, validation, and decision making).It also increases the runtime accuracy immensely.In the future,Onednn and other intel optimised algorithms can be incorporated to increase accuracy of the project further.

I hope this project will contribute to the agricultural industry and empower farmers in Karnataka to make informed decisions about their crops.I welcome any feedback and contributions to make this project even better!

Methodology / Approach

Convolutional neural networks (CNNs) are used to analyse images of leaves, while other supervised algorithms are used to analyse other attributes. It is built on a deep learning architecture.

• IMPORT LIBRARIES
• IMPORT DATASET
• EDA AND LABEL IMAGE
• SPLIT THE DATA AND PREPROCESS IT
• BUILD A MODEL
• TRAIN THE MODEL
• ANALYSE THE OUTPUT
• VISUALISATION

Technologies Used

The tools used for this project mainly belong to the libraries:

• Tensorflow
• numpy
• matplotlib
• pathlib
• os

The Intel technology used :INTEL ONEAPI

Intel OneAPI provides a comprehensive set of tools, libraries, and frameworks that can help optimize code for specific hardware architectures, which can lead to significant improvements in execution times. This level of performance optimization can be particularly valuable for users working on computationally intensive tasks that require high levels of processing power.

In context to this project,CNN's prediction happens in seconds unlike in other platforms.Almost all epochs are processed for the whole of training dataset within seconds whereas in other platforms it takes significantly longer time.This was also done without any RAM crashes during the processing.Runtime accuracy was also increased whilst using OneAPI.And to whoever uses it,it is clear that Intel OneAPI offers a powerful and versatile set of tools that can help developers and researchers achieve their goals with maximum efficiency and performance.

It is also a notable mention that most libraries are intel optimised to perform efficiently.

Documents and Presentations

Repository

https://github.com/Keerthi-pk10/MLsolutionforagriculture