/EDBV

Visually classifying ports on computers using image processing techniques.

Primary LanguageMATLAB

EDBV

Usage

Open the App.mlapp by double clicking on it.

Male Ports

For male ports, click 'Load Male Plug Image', and choose an image from the Images folder named port[1-4].jpg. After the image appears in the gui, click 'Calculate'. If the button's line turnes blue, it is calculating. This may take a while. After the detection is done, the ports in the image should be highlighted.

Female Ports

For female ports, click 'Load Female Plug Image', and choose an image from the Images folder named image[0-5].jpg to see good examples. The other images can be chosen as well (image[7-26].jpg), but they might not deliver good results, as they don't comply with our requirements. After the image appeared in the gui, click 'Calculate'. Now a new window will open, where you have to draw a rectangle on the standardized part where the ports are. After this step, the window closes and the results will be shown in the gui.

The meaning of the colors is explained in the gui, and there you can also see a status of the program.

Using own dataset

If you want to use your own dataset, you have to consider the following tips to get useable results from the program:

  • for male ports, the rotation of the port has to be exactly the same as in our given examples
  • for male ports, we only detect hdmi and usb ports
  • for male ports, the ports must each be held vertically and must be about 10-15cm away from the camera
  • for female ports, we only support standardized computer backs like in our examples
  • for female and male ports, the images must be taken in good lighting and without any rotation or distortions
  • for female ports, you must be able to draw the rectangle onto the standardized part where the ports are without cutting off anything, and without distortions