Automated Optical Inspection
Automated Optical Inspection (AOI) has become a powerful solution in the steel making industry. AOI techniques are used to inspect the surfaces of objects to reduce defective products with high precision and speed. In the inspection process, an AOI device acquires image of the object under inspection using an optical imaging system, and then identifies the type of defects by matching the input image with known template. Recently, deep learning has becoming a powerful solution for industrial inspection applications. In particular, deep learning-based object detection techniques are used to improve the speed and accuracy of an AOI device.
Steel Defect Detection
In any manufacturing process, it is not unusual to have defects on the products. But when it is too small to be noticed, checking the products precisely would take a lot of manpower and time. By training machines to distinguish the defects, it can quickly warn manufacturer the types of defect and the level of damage, which could help trace back the process to locate the source of problem efficiently. As shown on the left, these types of defects usually occurred in steel manufacturing, such as sponge, scratched, scaling, water droplets and etc.
Serial Number Identification
The identification of serial number on a product using video camera can be useful in an industrial setting. However, the area's lighting condition and the material could make identification difficult due to brightness, shadows or distortion from texture. By locating the serial number on the item, it is possible to train the machine to identify the numbers by undergoing training, which will help understand the pattern of numbers while getting minimal disruption from shadows or light reflection.