Pineapple Leaf Disease Dataset for Deep Learning-Based Classification and Precision Agriculture
Description
Pineapple leaf diseases are some of the most important factors limiting pineapple yield globally․ They cause reduced crop productivity‚ declining fruit quality and major economic loss to the agricultural industry․ It is vital to detect diseases and identify pathogens at an early stage to minimize damage and make agricultural practices more sustainable․ However‚ most methods used to identify a disease visually are time-consuming‚ labor-intensive‚ and prone to human error․ Thus‚ automatic detection systems for diseases based on deep learning and computer vision technologies have become prominent in precision agriculture․ To support this work on automated detection of pineapple diseases‚ we constructed a large Pineapple Leaf Disease Dataset that consists of healthy and diseased pineapple leaf images acquired from natural pineapple fields of Kerala‚ India․ This dataset contains images of both healthy and diseased pineapple leaves obtained from various fields under different environmental and lighting conditions and viewpoints․ The dataset contains 4 classes of pineapple leaf disease: Pineapple Leaf Blight‚ Pineapple Fusarium ‚ Pineapple Mealybug Wilt‚ and Healthy Pineapple Leaf․ The labels for the disease images were validated using experts‚ to avoid imprecise annotation and classification of the dataset images․ The images were taken with smartphone cameras using images of various sizes to diversify the image acquisition conditions․ The images were pre-processed to adjust the dataset to be more suitable for deep learning‚ by resizing to 224 x 224‚ removing low-quality and blurry images‚ and normalizing the pixel values․ The dataset was augmented by applying transformations to the images‚ such as rotation‚ horizontal flipping‚ zooming‚ width shifting‚ and height shifting‚ so that the model could generalize better to unseen data․ The Pineapple Leaf Disease Dataset may be used by researchers‚ agriculture researchers and developers to create an efficient automated detection system through the early detection of diseases and health monitoring of crops․ It may also be used in further research on plant disease classification‚ monitoring crops through computer vision techniques‚ precision agriculture and deep learning applications․ 1․ Original Dataset: Images: 2368 Data format: ․jpg 2․ Preprocessed Dataset: Dimensions: 224 × 224 pixels Data format: ․jpg 3․ Augmented Dataset: No․ of images: 7000 Data format: ․jpg
Files
Institutions
- The NorthCap UniversityHaryana, Gurugram