當你走進一間先進的工廠，看到產品一件件快速產出時，有沒有想過——這些產品到底是怎麼確保品質沒問題？答案就在你這學期要學的這門課：《AI 品質檢測運用技術與實務分享－影像辨識與自動缺陷檢測》！
別被這長長的課名嚇到，其實這是一門活潑又超級實用的課程。我們會帶你從生活中出發，一步步進入產業現場，看 AI 是如何變身成為「品管達人」，幫忙挑出每一個小瑕疵，讓產品出廠前都能保持完美！
為什麼產業需要 AI 來幫忙檢查品質？
在過去，檢查產品有沒有問題，全靠人眼掃、人手摸。但人眼會累、會漏看、會誤判，尤其在大規模生產時，人力難以負荷，錯誤也難以避免。
這時候，AI 就成了產業界的超級助手！透過影像辨識技術，AI 可以幫忙看產品的外觀、分辨有沒有刮傷、錯字、掉漆、異物等等。舉個例子：
* 在食品業，AI 可以偵測包裝有沒有破損或標示錯誤；
* 在面板或晶圓產業，AI 可以找到肉眼都難以察覺的微裂痕；
* 在車用零件產線上，AI 可以即時檢查焊接點是否正確。
這些應用，現在在很多台灣重要的科技、製造企業都已經上線了，你學會這套技術，**未來進入職場馬上能派上用場**！
我們會學哪些實用技能？
這門課不是那種只會放一堆 PPT 給你抄筆記的課，我們要帶你「做中學」，讓你真正理解並體驗 AI 怎麼在產業中發揮功能。
你會學到的技能包括：
* 用 Python 操作影像資料（不用擔心，我們會從最簡單的教起）
* 建立 AI 模型，例如 CNN、U-Net、YOLO 等，專門處理影像分類、偵測、分割
* 實作 AOI（自動光學檢查）模擬，讓 AI 幫忙比對正常圖與瑕疵圖
* 學習如何在 X-ray 影像中找出隱藏的異常點
* 更進一步了解如何把這些 AI 技術**實際導入產線流程**、與工程師、品保部門合作
產業界怎麼看待這樣的技能？
現在，越來越多企業都在推動「智慧製造」，也就是所謂的工業 4.0，而 AI 品質檢測正是其中非常重要的一環。能夠懂 AI、懂影像、又懂品質控管的人才，對企業來說非常搶手！
舉幾個真實例子：
* 有半導體公司導入 AI AOI 檢測後，把誤判率從 15% 降到 3%，節省大量人力與時間；
* 傳統機械加工廠在導入 AI 缺陷偵測後，報廢率大幅下降，品質穩定性上升；
* 食品包裝廠透過 AI 影像比對，能即時發現印刷錯字，避免整批商品下架。
這些例子都說明了：**AI 品質檢測不只是新科技，它是產業升級的關鍵武器。**
這門課的風格是？
我們會用案例講故事、用圖片講理論、用簡單的程式讓你動手玩，讓「品質檢測」這件事變得不再抽象。你不需要有強大的數學或程式基礎，只要願意嘗試，我們會一步步帶你進入 AI 與產業結合的世界。
上完課後，你會收穫什麼？
* 理解 AI 在實際工廠環境中的應用
* 具備實作簡單影像辨識缺陷檢測系統的能力
* 熟悉 Python 與電腦視覺的基本操作
* 能與工程師/品保部門討論導入 AI 檢測的可行性
* 對未來就業方向（品質、製造、AI 技術整合）有更明確的想像與準備
這不是一門冷冰冰的理論課，而是一門結合了科技、創意與產業需求的實戰課程。
**讓我們一起打開 AI 品質檢測的世界，用技術改變製造未來！**
When you enter an advanced factory and see products being produced one by one, have you ever thought about how these products are sure to ensure quality? The answer is in the course you will learn in this period: "Sharing of AI Quality Testing Application Technology and Practical Affairs - Image Identification and Automatic Defect Testing"!
Don't be scared by this long class title, it is actually a lively and super practical course. We will take you out of life and step by step into the industry scene to see how AI becomes a "quality control master", helping you pick out every small flaw so that the product can be perfect before it is shipped!
Why does industry need AI to help check quality?
In the past, checking whether there were any problems with the product was entirely up to human eyes and touching it with human hands. However, human eyes will be exhausted, missed, and misjudged. Especially when large-scale production, manpower will be difficult to bear the burden and errors will be avoided.
At this time, AI became a super assistant in the industry! Through image recognition technology, AI can help you see the appearance of the product, identify whether there are scratches, errors, paint drops, strange objects, etc. An example:
* In the food industry, AI can detect whether the packaging is broken or marked error;
* In panel or crystal industry, AI can find micro cracks that are difficult to detect with the naked eye;
* On the vehicle parts line, AI can check whether the welding points are correct immediately.
These applications are now available in many important technology and manufacturing companies in Taiwan. If you learn this set of technology, **If you enter the job market in the future, you will be sent to the market immediately**!
What practical skills will we learn?
This class is not the kind that only puts a bunch of PPTs to copy you. We want to take you to "learn by doing" so that you can truly understand and experience how AI can develop functions in your industry.
The skills you will learn include:
* Use Python to operate image data (no need to worry, we will start with the simplest teaching)
* Establish AI models, such as CNN, U-Net, YOLO, etc., and specialize in image classification, detection, and segmentation.
* Implement AOI (automatic optical inspection) simulation to allow AI to compare normal and defective pictures
* Learn how to find hidden anomalies in X-ray images
* Learn more about how to actually introduce these AI technologies into line processes**, and cooperate with engineers and quality assurance departments
How does the industry view such skills?
Nowadays, more and more companies are promoting "smart manufacturing", which is the so-called Industry 4.0, and AI quality testing is a very important part of it. Talents who understand AI, imaging, and quality control are very skilled in enterprises!
Here are some real examples:
* After semiconductor companies introduced AI AOI testing, the error judgment rate was reduced from 15% to 3%, saving a lot of manpower and time;
* After traditional mechanical processing factories introduced AI defect detection, the reporting rate dropped significantly and the quality increased steadily;
* Food packaging factories can instantly detect printing errors through AI image comparison to avoid the entire batch of products being removed from the shelves.
These examples illustrate: **AI quality detection is not just a new technology, it is a key weapon for industry upgrades. **
What is the style of this class?
We will use cases to tell stories, use pictures to explain, and use simple programs to make you play with it, so that "quality testing" is no longer abstract. You don't need to have a strong mathematical or programming foundation. As long as you are willing to try, we will take you step by step into the world of AI and industry.
What will you receive after the class?
* Understand the application of AI in an international factory environment
* Ability to implement a simple image identification defect detection system
* Be familiar with the basic operations of Python and computer vision
* Can discuss and introduce the feasibility of AI testing with engineers/quality insurance departments
* Have clearer imagination and preparation for future career directions (quality, manufacturing, AI technology integration)
This is not a cold theoretical course, but a practical course that combines technology, creativity and industry needs.
**Let us open the world of AI quality testing together and use technological changes to create the future! **
講義
Talk
| 評分項目 Grading Method | 配分比例 Grading percentage | 說明 Description |
|---|---|---|
| 課程參與小組討論與互動課程參與小組討論與互動 Course participation and group discussion and interaction |
10 | |
| 期末報告期末報告 Final report |
30 | |
| 作業與討論點名作業與討論點名 Business and discussion point names |
30 | |
| 期中報告期中報告 Midterm Report |
30 |
