how to find coordinates of intersection of lines after using "houghlines"

edit

BTW, to check that line1 and line2 actually intersect, verify that the angle between them (theta1-theta2) is high enough (around 90° or 270°)

thank you for fast answering,but there is a critical problem and as you can see from the picture there are 7 lines not four lines.Hence, i hanged on it. i just need to know the exact coordinates of largest polygon with each method that you think its fast and easier to do. Indeed, i need the coordinates for going on this code. Thank you.

 Point2f   inputQuad[4];  
 Point2f   outputQaud[4];


outputQaud[0] = Point2f( 0,0 );
    outputQaud[1] = Point2f( cropped.cols,-10);
    outputQaud[2] = Point2f( cropped.cols,cropped.rows);
    outputQaud[3] = Point2f( 0,cropped.rows  );



lambda = getPerspectiveTransform(inputQuad, outputQaud );

     warpPerspective(cropped,output,lambda,output.size() );

To answer directly your question, I use the following procedure to get the 4 points around an object:

1. get the center of the object (using image moments: xc=mom.m01/mom.m00; yc=mom.m10/mom.m00;
2. recalculate rho/theta for every line according to this point
3. create a vector of lines (rho/theta pairs). Add the first line.
4. continue adding lines to the vector. Add a line only if the theta of the current line is different of all the thetas in the vector (abs(theta(current)-theta(j))>45° for every j). Stop once you have 4 lines in your vector.
5. Sort the vector according to theta.
6. Get the intersection of these lines using the procedure from my previous post. The intersection of line(4) and line(1) will give the top-left corner, line(1) and line(2) the top right corner and so on.

However there is a simpler solution: use contour detection and the enclosing convex hull to get the polygon that approximates the best your object. Tutorial here

Please tell me which solution helped you so I can update my answer.

thank you for your answers, to be quiet honest , i am new one in opencv and image processing. i would like to "wrapaffine" the following picture which i have attached as same as the "card image"

i think the first solution is better for me is it possible for you describe it more ? Thank you.

I propose to try the second solution, it's much easier to implement. If it doesn't work, you can still try the first one.

actually my main question is obtaining the coordinates from the contours,as you told me i used the second one,but unfortunately i do not know how can i solve my problem,i really appreciate. is it possible bring me one example according solution one?