Ask Your Question

Revision history [back]

Different ways to detect LED Screen from the image.

raw_image

I want to extract the LED screen from the image above. Some approaches that I have tried include:
1. I first converted the image to HSV and made a trackbar GUI through which I noted at what value of HSV our mask filters out our ROI.
2. using canny edge detection, contours extraction, and selecting the contour with 4 vertices and area greater than 100

Both of the solutions do work.

The problem with the first approach is that it only works on a pinkish screen. While the second approach is more generic, but both approaches need a lot of fine-tuning to get the required result.

My question is that if there's any other more generalised approach that I can try which require less tuning of parameters ?

Different ways to detect LED Screen from the image.

raw_image

I want to extract the LED screen from the image above. Some approaches that I have tried include:
1. I first converted the image to HSV and made a trackbar GUI through which I noted at what value of HSV our mask filters out our ROI.
2. using canny edge detection, contours extraction, and selecting the contour with 4 vertices and area greater than 100

Both of the solutions do work.

The problem with the first approach is that it only works on a pinkish screen. While the second approach is more generic, but both approaches need a lot of fine-tuning to get the required result.

My question is that if there's any other more generalised approach that I can try which require less tuning of parameters ?