It comes down to how noisy your data is. In practice not all your lines will intersect at a single point, so you may get many intersections and may have to do some nonmaximal supression or if you like linear algebra can rely on some SVD solution to minimize error. It'll require some tuning depending on what images you're working with. Are you trying to implement the algorithm?
Yes you could try setting your own threshold or use maximal suppression. It’s a similar concept I talked about in my Good Corner Detection video. Or you can try to take the average as another approach
how do we decide what should be the ideal number of lines like when you changed some threshold values and we could see the number of lines changing?
It comes down to how noisy your data is. In practice not all your lines will intersect at a single point, so you may get many intersections and may have to do some nonmaximal supression or if you like linear algebra can rely on some SVD solution to minimize error. It'll require some tuning depending on what images you're working with. Are you trying to implement the algorithm?
Thank you for sharing this,
Is there a solution to consolidate multiple closely parallel lines detected by my code into a single line?
Yes you could try setting your own threshold or use maximal suppression. It’s a similar concept I talked about in my Good Corner Detection video. Or you can try to take the average as another approach