Surface defect detection of steel, aluminum and metal strip


The purpose of this project is to use the VisionBank artificial intelligence vision system to detect the upper and lower surface defects of the steel belt.

System Solution

1. Product width: about 12mm, running speed 10m/min. The inspection is random inspection, and the preliminary design is to capture and analyze image once every second.


2. The detection range includes the top and bottom of the product.


3. The types of defects include: pits and abnormal variegated colors. There are also black textures on the surface of qualified products, which need to be filtered out to prevent false alarms.

Product Inspected

Acceptable Product

Unaccepted Product

Project Details

Project Name: Wuxi Sopa Steel Strip Surface Defect Detection


Item category: Defect detection – Dented product, variation in color


Field of view: 20mm*15mm


Detection accuracy: 0.01mm (visible to the human eye)


Movement speed: 100 m/min

Detection Results

All defect areas analyzed by the software (marked in red).

Display data by perimeter dimension and ignore areas with perimeter less than 20.

Low-precision detection, even if there is a small scratch, it is still qualified.

Thanks for reading!

Xem thêm nhiều ứng dụng thành công khác

Xem thêm những khó khăn của các ứng dụng Machine Vision và cách các sản phẩm của chúng tôi khắc phục điều đó.

Về Trang Chủ

Tìm hiểu thêm về Techvico và các sản phẩm của chúng tôi.