自动化
计算机科学
一致性(知识库)
质量(理念)
人气
选择(遗传算法)
人工智能
人工神经网络
目标检测
机器视觉
图像处理
对象(语法)
计算机视觉
工程类
模式识别(心理学)
图像(数学)
认识论
机械工程
心理学
社会心理学
哲学
作者
Yunjie Tang,Kai Sun,Danhuai Zhao,Yan Lü,Jiaju Jiang,Hong Chen
标识
DOI:10.1109/dsc55868.2022.00091
摘要
To ensure the quality of products, it is crucial to inspect and assess their condition in quality control. Among all of the methods, surface inspection is a critical step to identify defective products. With the recent advancement in artificial intelligence and computer vision, a plethora of industries are expecting next level automation. The investments in automated defect detection systems are gaining popularity today as they not only reduce labor costs but also improve the consistency of the production line. This review paper presents some examples of defects in the first part. Then some basic but extensive introduction about industrial camera selection, lens selection, optical illumination are included. Since the images collected from factories are not always satisfying, common methods for image data processing are systematically discussed. Then, this survey comprehensively investigates two neural network algorithms vastly used in industrial object detection systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI