A Dual-Dataset Framework for Apple Leaf Analysis: Lifecycle-Based Pest Detection and Large-Scale Disease Classification

This dataset presents a comprehensive dual-dataset framework designed to support research in automated apple leaf analysis, with a focus on both pest lifecycle detection and large-scale disease classification. The dataset is particularly relevant for developing artificial intelligence (AI) and deep learning models for precision agriculture and plant health monitoring. The dataset consists of two complementary components: 1. Lifecycle-Based Apple Leaf Miner DatasetThis dataset captures the biological progression of apple blotch leaf miner infestation, collected from the Zainapora region of Shopian district in Jammu and Kashmir, India—one of the earliest reported outbreak locations. It contains 1,200 high-quality images categorized into four biologically meaningful classes: Healthy leaves Dormant pupal stage Overwintering pupal stage Larval infestation stage This dataset enables fine-grained analysis of pest lifecycle stages, supporting early detection and targeted pest management strategies. 2. Large-Scale Apple Leaf Disease DatasetThis dataset consists of 33,914 images representing five major classes relevant to apple leaf health: Healthy Apple scab Mites infestation Alternaria leaf spot Apple rust The dataset was constructed by combining publicly available datasets with field-collected images to enhance diversity and real-world applicability. Class imbalance was addressed using data augmentation techniques, ensuring robustness for machine learning applications. Key Features: Dual-dataset structure addressing both pest lifecycle and disease classification Real-field data collected from Jammu and Kashmir orchards Large-scale dataset suitable for training deep learning models Balanced class distribution through augmentation Supports tasks such as classification, detection, and model benchmarking Potential Applications: Deep learning-based plant disease detection Pest lifecycle monitoring and prediction Smart agriculture systems and decision support tools Mobile-based crop health diagnostic applications Data Format: Images in standard formats (JPG/PNG) Organized into class-wise directories for ease of use Compatible with popular machine learning frameworks (TensorFlow, PyTorch, etc.) Geographical Context:Data is collected from apple-growing regions of Jammu and Kashmir, India, making it particularly relevant for temperate horticulture ecosystems.