Page 74 - 国外钢铁技术信息内参（ 2022年5月）

ᇏݓࣁඋ࿐߶

（七）如何利用人工智能简化金属工业自动光学表

面检测操作 ᇏݓࣁඋ࿐߶

ᇏݓࣁඋ࿐߶
Chris Miller , Joe hen Koenig, Mark Cornell

（美国申克博士有限责任公司）

摘要：虽然钢、铝、黄铜、铜、锡和涂层/涂漆产品等金属表面的自动光学检测

（AOI）功能越来越强大和多样化，但有时很难应用到日常操作中，尤其是在操 ᇏݓࣁඋ࿐߶

作和常驻工程人员缩减的情况下。因此，重要的是 AOI 制造商将开发重点放在
ᇏݓࣁඋ࿐߶
工具和方法上，目标定位为 AOI 系统基本上无需干预日常操作并尽可能简化其

初始应用。这是通过使用各种形式的人工智能（AI）来实现的，AI 可以自动执

行诸如调整亮度等级、检测水平度和缺陷分类等任务。本文将介绍这些功能通常

如何工作，并展示对其用法的研究。

How Artificial Intelligence Simplifies Automated
ᇏݓࣁඋ࿐߶
ᇏݓࣁඋ࿐߶
Optical Surface Inspection Deployment and

Operation in the Metals Industry

ABSTRACT：While automated optical inspection (AOI) of metal surfaces such as

steel, aluminum, brass, copper, tin and coated/painted products has grown

increasingly powerful and versatile, the deployment into day-to-day operations can at
ᇏݓࣁඋ࿐߶
times be difficult, particularly as operational and resident engineering staffing is

reduced: It has therefore been important to focus the development at AOI

manufacturers on tools and methods toward the goal of making AOI systems

essentially hands-off for day-to-day operation and simplify their initial deployment as
much as possible. This is done utilizing various forms of artificial intelligence (AI),

which automate tasks such as adjusting light levels, detection · levels and defect
ᇏݓࣁඋ࿐߶

71