PCB缺陷检测与分类数据集
收藏资源简介:
本数据集由北京大学深圳研究生院机器感知重点实验室创建,包含1386张印刷电路板(PCB)的图像,涵盖6种常见缺陷类型。数据集分为两部分,一半图像为正确放置的PCB,另一半为随机旋转的PCB,以模拟实际检测中的不同情况。每张图像均提供缺陷的边界框和坐标信息,便于研究者定位缺陷。该数据集主要用于PCB缺陷的检测、分类和注册任务,旨在解决传统手动检测效率低下和易出错的问题。
This dataset was created by the Key Laboratory of Machine Perception, Peking University Shenzhen Graduate School. It contains 1386 printed circuit board (PCB) images covering 6 common defect types. The dataset is split into two subsets: half of the images are correctly positioned PCBs, while the other half are randomly rotated PCBs, aiming to simulate various scenarios in actual PCB inspection. Bounding boxes and coordinate information of defects are provided for each image to help researchers accurately locate defects. This dataset is primarily designed for PCB defect detection, classification and registration tasks, aiming to address the issues of low efficiency and high error rate in traditional manual inspection.

- 1A PCB Dataset for Defects Detection and Classification北京大学深圳研究生院机器感知重点实验室 · 2019年



