Okra disease dataset for classification and segmentation: Dataset collection, analysis and applications.

Area of Science:

• Agricultural Science
• Computer Science
• Plant Pathology

Background:

• Early diagnosis of okra leaf diseases is vital for crop health and agricultural productivity.
• Automated disease detection using deep learning requires comprehensive datasets.
• Existing datasets may not capture real-world variations common in Indian agricultural fields.

Purpose of the Study:

• To introduce a novel, comprehensive dataset of Indian okra leaf images for deep learning model development.
• To provide a benchmark resource for early-stage plant disease classification, detection, and segmentation.
• To support research in smart agriculture, machine learning-based disease diagnosis, and precision agriculture.

Main Methods:

• Collected 2500 okra leaf images from real-time agricultural fields in India.
• Categorized images into six classes: healthy and five common diseases (Leaf Curly Virus, Alternaria Leaf Spot, Cercospora Leaf Spot, Phyllosticta Leaf Spot, Downy Mildew).
• Resized all images to 224 × 224 pixels for compatibility with standard deep learning models.

Main Results:

• A unique dataset of 2500 images representing healthy and diseased okra leaves under real-world conditions.
• The dataset incorporates natural variations in lighting, leaf positioning, and environmental factors.
• This dataset is one of the first publicly available Indian okra leaf disease datasets captured in natural settings.

Conclusions:

• The dataset serves as a valuable resource for researchers, particularly young scientists, in smart agriculture.
• It enables advancements in machine learning-based disease diagnosis and precision agriculture applications.
• Future work will expand the dataset to enhance model generalization and real-world applicability.