Multi-Environment Comprehensive Image Dataset for Automated Brick Quality Grading and Class Classification in Structural Engineering Frameworks

This dataset contains exactly 10,000 high-resolution real-world images of bricks collected by the authors to enhance machine learning research on automated brick quality grading and class classification in structural engineering frameworks. The images represent diverse field environments and various levels of material quality. The dataset is organized hierarchically into folders corresponding to the collection locations (such as manufacturing brick fields and active construction sites) and is further granularized into six distinct categorical classes. All images were manually verified, filtered, and preprocessed to a uniform 1:1 aspect ratio to ensure accurate, artifact-free, and consistent annotations. This dataset is intended for deep learning automation, automated quality assessment platforms, and visual inspection models to train, validate, and evaluate computer vision classifiers. While the images reflect authentic real-world industrial and construction scenarios, extreme variations in environmental lighting, background conditions, and multi-angle perspectives are intentionally present to ensure field generalizability. Categories for this data: 1st Class: Premium units with uniform deep red/copper color and sharp 90-degree corners. 2nd Class: Standard units with minor shape irregularities, slight abrasions, or small chips. 2nd Class (Algae): Second-grade units with visible organic moss, algae, or biological films. 3rd Class: Under-burnt, pale pink/yellowish units with soft, crumbling boundaries. Pricate: Over-burnt, structurally warped units with bloated or fused shapes. Jhama: Highly over-burnt, metallic purplish-black units with porous, melted masses.