A Benchmark Dataset for Detecting Disease in Plant Leaves: An Essential Resource for Deep Learning Models.
Description
This dataset has been developed for research on plant leaf disease detection, specifically targeting Gourd, Zucchini, Bitter melon, Bean, Aubergine and Yardlong bean leaves. It contains high-resolution images capturing various disease symptoms to support the development and evaluation of deep learning models in agriculture. The dataset is ideal for use in image classification, segmentation, and disease diagnosis applications. Dataset Overview: Contains images of plant leaves affected by various diseases (Bacterial, YMV, Fungal, Viral infections) commonly found in agricultural crops. Captured under controlled conditions, ensuring realistic and diverse real-world scenarios. Includes images from multiple plant species and varying disease stages. Types of Plant Leafs: 1. Gourd (537) 2. Zucchini (186) 3. Bitter melon (300) 4. Bean (517) 5. Aubergine (242) 6. Yardlong bean (500) Key Features: Number of Images: 2820 Number of Augmented Images: 2561(to increase dataset variability and model robustness) File Formats: JPEG Disease Types: Bacterial, YMV, Fungal, and Viral infections. Applications: Deep Learning: This dataset is specifically designed for the development and evaluation of deep learning models for plant disease detection. It can be used to train convolutional neural networks (CNNs) and other advanced deep learning architectures for tasks such as image classification, object detection, and segmentation, providing insights into plant health status and disease severity. Machine Learning: In addition to deep learning, this dataset can also be utilised in traditional machine learning methods for disease recognition, but its primary value lies in enabling the use of cutting-edge neural networks for automated, scalable disease detection. Agricultural Technology: The dataset supports the development of mobile applications and automated systems for real-time plant health monitoring. Deep learning models trained on this data can be integrated into mobile platforms or drone-based systems to provide instant, accurate disease diagnosis in agricultural settings, aiding farmers in timely interventions. Agricultural Research: Researchers can use this dataset to better understand the impact of diseases on various plant species and their visual symptoms. The rich diversity of images will also help in studying disease progression, improving the design of predictive models, and contributing to better crop protection practices. Dataset Collection: 1. Compiled from publicly available plant images and field-controlled environments. 2. Includes varying lighting conditions, leaf positions, and disease stages for robust model training. By providing detailed, high-quality images of diseased plant leaves from multiple species, this dataset plays a critical role in advancing deep learning applications in agriculture. It enables more accurate disease detection, which can lead to faster responses and more sustainable farming practices.
Files
Steps to reproduce
Data Collection Locations: 1. Patukhali (southern Bangladesh, known for diverse crop farming) 2. Kushtia (located in the west of Bangladesh, famous for a variety of vegetable crops) 3. Gazipur (near the capital, with both commercial farming and research centers) These areas were chosen due to their varied environmental conditions and agricultural practices, which are representative of different stages of plant disease progression and farming environments. Image Capturing Methodology: 1. Phone Camera Data Augmentation: 1. Python libraries