PlantDiseases Dataset for Edge - AI Based Real - Time Plant Disease Recognition

Description

The PlantDiseases dataset is constructed for edge-AI-based real-time plant disease recognition. It encompasses 28 distinct disease types across 16 plant species, totaling 46 diagnostic categories. After preprocessing and augmentation, the dataset contains 259,135 images, including 220,498 training images, 19,419 validation images, and 19,218 test images. The images were collected under varying environmental conditions using different imaging devices and from multiple perspectives. This dataset provides essential data support for algorithm research in edge-computing scenarios for plant disease recognition.

Steps to reproduce

To reproduce the results related to this dataset: 1. Download the dataset from the provided link. 2. Preprocess the images by removing invalid or low-quality samples, resizing images to 224 × 224 pixels, and applying data augmentation such as contrast enhancement, mirror flipping, and scale transformation. 3. Use a deep learning framework such as PyTorch or TensorFlow to build and train the model. 4. Train the model with the training subset of the dataset using a batch size of 32, a learning rate of 0.0001, and the Adam optimizer. 5. Evaluate the trained model on the validation and test subsets to obtain performance metrics such as accuracy.

Institutions

• Henan Normal University

  Xinxiang

Categories

Funders

• Key Scientific Research Projects of Higher Education Institutions in Henan Province

  Grant ID: 25B413001
• Henan Science and Technology Department

  Henan, China

  Grant ID: 242102320175

Licence