Durian Plantation Disease Leaf Rot Detection Image Dataset

Ensures high-quality data through multiple rounds of annotation and automated consistency checks, combined with reviews by agricultural pathology experts. The annotation team consists of 10 professionals in agriculture and computer vision. Pre-processing steps include noise reduction, size adjustments, and color normalization to enhance the model's recognition capabilities. Data is stored in JPG format, organized hierarchically in clear file directories. The core advantage of this dataset is an annotation accuracy of up to 95% and excellent consistency, with innovative application of data augmentation techniques such as random cropping and color jittering to enrich training samples and improve recognition robustness. Integrates the latest quality assessment methods, improving model accuracy by 10% over other public agricultural datasets based on industry standards, and significantly reduces false alarm rates. Notably, it includes representative disease images from major durian growing areas, providing the model with a comprehensive perspective. Supports transplantation in different scenarios, reflecting the dataset's scalability and versatility.