Plant Disease Detection

Abstract

Insects, viruses, and pests are typically the causes of plant diseases, which, if left unchecked, drastically reduce output. Numerous agricultural diseases are causing losses for farmers. When the cultivated area is large, measured in acres, it becomes tiresome for the growers to routinely check the crops. The suggested technology uses remote sensing photos to automatically detect diseases and offers a way to routinely check the agricultural area. The suggested approach alerts farmers about crop illnesses so they may take appropriate action.Early disease identification as soon as the illness begins to spread on the outer layer of the leaves is the aim of the suggested method. The initial step of the suggested system\'s operation deal with training data. Training both healthy and ill data sets is part of this. Phase two involves crop monitoring and disease findings using Canny\'s edge detection technology.

Introduction

1. Background

• Agriculture is the foundation of civilization and a key sector in India's economy.

• With a growing population, there is increasing pressure to boost crop yields.

• Crop diseases significantly reduce the quantity and quality of produce.

• Traditional disease identification methods (expert consultation, lab tests) are costly, slow, and require specialized knowledge.

2. Problem Statement

• There’s a lack of efficient, cost-effective systems for early and accurate crop disease detection.

• Although some advanced systems exist, they are not widely accessible, especially in rural or under-resourced regions.

3. Objectives

1. Detect whether a crop is diseased.

2. Recommend remedies for detected diseases.

3. Use machine learning (ML) and image processing techniques (e.g., RGB values, feature extraction) for automated analysis.

4. Perform tasks like pre-processing, feature extraction, classification, and model training.

4. Methodology

• Pre-processing: Resize images to a uniform size.

• Feature Extraction Techniques:

  • HOG (Histogram of Oriented Gradients): Captures contours and object appearance.

  • Hu Moments: Captures the leaf’s shape from grayscale images.

  • Haralick Texture Features: Captures surface texture differences between healthy and diseased leaves.

• Machine learning models are trained on image datasets to classify leaf health.

5. Existing Systems – Limitations

• Current detection methods like PCR, mass spectrometry, thermography, and hyperspectral imaging:

  • Are time-consuming.

  • Not cost-effective for large-scale or smallholder use.

  • Offer low accuracy in practical crop settings.

6. Proposed System

• Machine Learning-based image classification using:

  • Datasets of both healthy and diseased leaves.

  • RGB analysis, threshold detection, edge detection, and histogram comparison.

  • Remote sensing for regular monitoring.

• Aims to train the model with diverse crop images to recognize specific diseases and suggest treatment.

Advantages of Proposed System

? Improved accuracy in detecting diseases
? Faster and more robust than traditional methods
? Cost-effective and accessible for farmers

Conclusion

The architecture of an ML-based system that offers easily accessible, real-time local environmental data in rural agriculture fields is presented in this study. Researchers and agricultural field managers may obtain precise environmental data without having to visit the crop field to gather local data since the data is pushed in real-time to readily available cloud storage. A straightforward machine learning technique based on SVM regression was developed to forecast wind speed, average air temperature, and relative air humidity.This technology aids in the precise early detection of illness and assists farmers in forecasting the quantity of pesticides required for crops. This contributes to lower manufacturing costs and time consumption.

References

[1] Classification of Pomegranate Diseases Using Back Propagation Neural Networks, International Research Journal of Engineering and Technology (IRJET), Vol. 2, Issue 02 | May 2015 [1] S. S. Sannakki and V. S. Rajpurohit [2] \"Identifying Cotton Leaf Disease through Pattern Recognition Techniques,\" by P. R. Rothe and R. V. Kshirsagar,International Conference on Pervasive Computing (ICPC), 2015. [3] \"Leaf Disease Detection and Grading using Computer Vision Technology & Fuzzy Logic\" by Aakanksha Rastogi, Ritika Arora, and Shanu Sharma was presented at the 2nd International Conference on Signal Processing and Integrated Networks (SPIN) in 2015. [4] \"Automated Vision-Based Diagnosis of Banana Bacterial Wilt Disease and Black Sigatoka Disease,\" by GodliverOwomugisha, John A. Quinn, Ernest Mwebaze, and James Lwasa, was published prior to the first Africa\'s mobile ICT use international conference in 2014. [5] \"SVM-based Multiple Classifier System for Recognition of Wheat Leaf Diseases,\" by Yuan Tian, Chunjiang Zhao, Shenglian Lu, and Xinyu Guo, The 2010 Conference on Reliable Computing Proceedings (CDC\'2010), November 20–22, 2010.

Copyright

Copyright © 2025 Miss. Ayesha Firdose, Mrs. Sharvani V. 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.