Defect Detection Dataset: Porosities in Machined Aluminum Holes
收藏资源简介：
This dataset comprises 302 JPEG images captured with an endoscopic camera, focusing on detecting porosities in the machined holes inner walls of cast aluminum parts. Each image has a resolution of 400x400 pixels in RGB color space, providing detailed views of potential defects. The dataset is intended for developing and evaluating algorithms for automated defect detection in industrial manufacturing, specifically targeting porosity defects in aluminum casting processes. It does not include annotations or labels. Researchers can use these images to: Train and test machine learning models for defect detection. Explore characteristics and distributions of porosity defects in machined holes. Develop algorithms for automated quality control in manufacturing settings. Preprocessing such as normalization and resizing may be necessary before applying the images to machine learning tasks. Data was collected using SF-CQ6USB-D2.0 Endoscopic camera.
本数据集包含302张由内窥镜拍摄的JPEG（Joint Photographic Experts Group）图像，核心任务为检测铸铝零件加工孔内壁的孔隙缺陷。每张图像均为400×400像素的RGB色彩空间格式，可清晰呈现潜在缺陷的细节特征。本数据集旨在开发与评估工业制造场景下的自动化缺陷检测算法，尤其针对铝铸造工序中的孔隙缺陷检测任务。本数据集未附带标注或标签信息。研究人员可利用该批图像开展以下工作： 1. 训练并测试用于缺陷检测的机器学习模型； 2. 探究加工孔内孔隙缺陷的特征与分布规律； 3. 开发面向制造场景的自动化质量控制算法。 在将图像应用于机器学习任务前，可能需要进行归一化、尺寸调整等预处理操作。本数据集采用SF-CQ6USB-D2.0型内窥镜采集所得。
