Apple Defect

This is a self-built apple surface defect detection dataset containing 2792 real-scene apple images with six categories: normal apple, disease spot, bruise, blemish, crack, and russet. It covers common apple surface defects including pathological lesions, mechanical damages, and physiological rind imperfections. All images are manually collected, filtered, and accurately annotated with reliable quality, fitting actual fruit sorting and quality inspection scenarios. The dataset is suitable for visual algorithm research and model training on apple defect classification, object detection, and quality grading in smart agriculture. It is open for non-commercial academic and educational use only. Unauthorized resale and commercial usage are prohibited. Please cite this dataset in your research.