Thread Counting Using Image Processing

Aim

The main aim is to count the number of threads using image processing as the camera keeps moving. It is to be noted that the processing of image is done till approximately half of the image actual width since the camera is moving, the number of thread count keeps getting updated.

Details

The project was taken as an initiative for milling factory to detect wherever thread is cut loose due to high string tension.

Hardware

• Raspberry pi
• Webcam
• 4 lasers (low intensity) for pinpointing from where to where the image is to be processed mobile power bank with 1.5 A steady output
• 8GB SD card for storage memory.

Software

The software involved was OpenCV 2.4.11, visual studio (for simulation on computer) and Raspberry pi GUC viewer.

Advice from Makers

OpenCV is much better than almost all other visual processing softwares as it is not only available for simulation on computers but also for Android phones in the form of OpenCV SDK. The only problem we could probably face is the failure of initiation of functions as mentioned in OpenCV 2.4.11 documentation for which other source like stack overflow come into play. Having a microprocessor such as raspberry pi as an interface between software and hardware is really advised.