使用 AI 加速偵測瑕疵
車身焊接會產生熔融金屬悍渣,可能傷害作業人員或損壞線路。傳統上需由人工尋找並清除悍渣。然而在每 60 秒一個週期、每輛車身 2,000 個焊點、每日生產 1,000 輛車的情況下,此方式效率不足。基於視覺的 AI 解決方案更可靠且效率更高。Audi 開發自有 AI 模型,利用八台攝影機的影像在雲端進行訓練,並直接部署於廠房。
車身焊接會產生熔融金屬悍渣,可能傷害作業人員或損壞線路。傳統上需由人工尋找並清除悍渣。然而在每 60 秒一個週期、每輛車身 2,000 個焊點、每日生產 1,000 輛車的情況下,此方式效率不足。基於視覺的 AI 解決方案更可靠且效率更高。Audi 開發自有 AI 模型,利用八台攝影機的影像在雲端進行訓練,並直接部署於廠房。

在本地端執行 AI 模型讓 Audi 能將週期時間降至最低,因為 AI 幾乎可即時分析車身影像。

AI 能可靠偵測焊渣並協助作業人員將其清除,避免線路受損導致車輛不良品,同時可避免人員受傷。

由於 Audi 使用 Siemens 的標準化解決方案部署自有 AI 模型,因此能夠快速輕鬆將此技術推廣至全球其他工廠。
This e-book explains how AI can make your production process smarter and more adaptive. There’s no reason to wait any longer to pioneer new, smart manufacturing methods and make a real impact on the shop floor.

Explore how Audi’s AI model analyzes images of car bodies in near real time, spotting defects and guiding operators to correct them, improving quality and efficiency on the shop floor.
In the body shop, robots weld the car bodies. When they finish, eight cameras capture the underside. The images go straight to Audi’s AI model, running on the standardized Siemens Industrial Edge – bringing powerful intelligence directly to the shop floor


The AI analyzes the images and detects weld spatter in near real-time. The data is sent to the next cell, where the car body is rotated 90 degrees for inspection and processing.
The AI directs blue lights to mark weld spatter on the underbody, making it easy for operators to remove. Soon, robots will take over this task – freeing skilled workers to focus on more valuable jobs.

In the future, we will have many AI solutions on our shop floor, for example for predictive maintenance. The Siemens Industrial AI Suite and Industrial Edge will help us deploy these solutions in our factories.