The following case is a record of Intsoft Technology’s detection of surface defects on coils. It records various defects on the coil surface and analyzes which can be detected by conventional detection and which require AI intelligent detection.

visual detection items

coiled material defect detection

Width: 1.5M,

Accuracy: 0.05mm,

Speed: 50-70m/min.

Workflow:

  1. Enter basic product information, take OK products for modeling, and set inspection data.
  2. Install the camera at a fixed position. When the product reaches the photoelectric sensor position, a trigger signal will be generated immediately, and the camera will collect images and upload them to the database.
  3. After receiving the image information, Intsoft Technology visual recognition system will perform a series of processing, analysis, and good/bad product judgments, and the interface will output the inspection information in real time. If the product is detected to be inconsistent with the preset data, an NG signal will be output; if the inspection is OK, the product will enter the next process according to customer requirements.
  4. According to the preset parameters, product inspection and judgment are carried out, and inspection parameters can be set in different areas to flexibly respond to different inspection requirements of different areas of the same product;

Image acquisition effect:

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Process:

Bulge:

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Defects such as wrinkles and roller marks are not clear, and ordinary algorithms cannot extract and identify them, so AI processing is required:

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Impurity:

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Scratch

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Impurity

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Scratch and Impurity

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

Bulge:

Coiled Material Defect Detection-Using Deep Learning-industrial vision integrator | defect detection | machine vision inspection

System functions:

  1. When the system detects a defect, it outputs an alarm signal.
  2. Real-time information display and record detection information.
  3. Provide special tools such as system parameter adjustment and image saving. The system interface is friendly, operable and intuitive.
  4. In order to ensure the security of data, the system is equipped with permission management. Only administrators with permission can modify the corresponding system parameters.
  5. According to the selected product inspection items, matching the inspection program, the inspection area can be adjusted according to the actual situation.
  6. The inspection history records can be automatically counted, saved, queried, called and other functions.