aditya_sanghi's profile - activity

I have tried the following methods in c++ opencv to accurately detect the lines before using probabilistic hough line transform in opencv:

Forming skeleton using erosion, dilation and subtraction

Morphological closing

Canny Edge detection

Gaussian Blur

Thresholding I have used a lot of combinations and permutations of these methods as well as played around with the parameters of each of these functions as well as parameters of probabilistic hough line transform!

The skeleton method would have been ideal but it results a straight thick line being converted to shorter crooked lines

I am yet to get the ideal results. What improvements can be made or are there some methods which I have missed out on?