The constructed defect-detection dataset including neutron and x-ray images

The uploaded dataset is built for neutron radiography images defect detection and segmentation, covering four typical defect types: blow holes, cracks, inclusions, and residual cores.It contains neutron/X-ray imaging data with labels (exceeding 10,000 samples through synthetic data generation and augmentation techniques) and is organized into four specialized subsets:N/X-Det: For defect detection tasks in neutron radiography images.N/X-Seg: For pixel-level defect segmentation in neutron radiography images.N/X-Seg1: An optimized dataset with high defect-to-background ratios, created through cropping of raw data.N/X-Seg2: An enhanced dataset derived from 4× super-resolution reconstruction of N/X-Seg1 to refine texture details.