This CogEye toolkit offers an unique ready-to-use computer vision-based software solution. This has been tested in real-time, optimized, field-tested, and verified as a reliable software solution. It is based on a set of algorithms in computer vision and machine learning.


Features:

• Minimize the light variation changing the infrastructure
• Capture the Image in Real-time
• Arduino-based IoT device (Camera +IR Sensor+ Arduino) captures the image and controls the flow of the software
• Mark the area of interest
• Extract the area of interest from each capture
• Minimize the printing bed vibration effect
• Identify noise in the images (Implemented a deep-learning neural network to remove the noise factor from each pixel)
• Distinguish defects from noise
• Identify printing defects
• Visualize defects in a user-friendly manner

The time taken to inspect a garment is less than 4 seconds, much faster than conventional human inspection. It has a high efficiency and accuracy rate. Generally, for plain garments, Accuracy has been monitored at 99.99%, while for AOP (All-Over-Print) fabrics, Accuracy is determined around 85% – 90%.