This study explores the early detection of leaf diseases in tulsi plants, which contribute significantly to agricultural losses in India. Using image processing techniques and segmentation methods like thresholding, we develop a computer vision-based approach for scalable disease identification. The TulsiDoc dataset, downloaded from Mendeley Data, contains over 1000 samples, consisting of 500 healthy leaf images and` 500 diseased leaf images. The paper demonstrates how segmentation improves disease prediction accuracy, reduces false positives, and enhances our understanding of plant health, which is critical for food safety and agricultural science. Furthermore, we employ state-of-the-art deep learning models, including Inception V3, ResNet50, and VGG16, to classify the plant images and compare their accuracy. The model with the highest accuracy will be selected for further testing and classification, ensuring optimal performance in disease detection
INTRODUCTION
The early detection of leaf diseases is vital for maintaining plant health and boosting agricultural productivity, especially in a country like India, where agriculture plays a pivotal role in the economy. Among the myriad of plants grown in the country, the tulsi plant (Ocimum sanctum), commonly referred to as holy basil, holds a unique position due to its medicinal, cultural, and economic significance. Tulsi leaves are extensively used in Ayurvedic medicine, culinary preparations, and religious rituals, making the plant a valuable asset. However, like many other crops, tulsi plants are prone to various diseases, which can severely impact both their yield and quality. These diseases, often caused by fungal infections, pests, or environmental stress, not only reduce the economic returns for farmers but can also compromise the medicinal properties of the leaves.
Hardware Requirement
1)Pc
Software Requirement
1)Python idle 3.8
Algorithm
1)CNN2D
2)VGG16
3)Resnet50
Library
1)Keras
2)Tensorflow
3)Scikit-learn
4)Numpy
5)Pillow
6)Opencv-python
