This is the code corresponding our paper "Fast and Regularized Reconstruction of Building Fa\c{c}ades from Street-View Images using Binary Integer Programming" submitted to ISPRS2020 Congress.
The only requirement is Mosek, which has a free one-year trial version for academic use.
IN_PATH represents the path of the original box file, OUT_PATH represents the output path of the box after regularization.The file suffix is .xywh, it means each line in the file represents the coordinates of the upper left corner point, the width of the box and the height of the box. delta_x, delta_y, delta_w, delta_h represent the step size of the pre-clustering, alpha_x, alpha_y, alpha_w, alpha_h represent each weight, you should change the values of these parameters to get the result you want.
std::vector<double> x, y, w, h, X, Y, W, H;
std::vector<int> r;
auto s1 = read_file(IN_PATH);
get_xywh(s1, x, y, w, h);
N = x.size();
m = X.size();
n = Y.size();
m_ = W.size();
n_ = H.size();
pre_cluster(x, X, delta_x);
pre_cluster(y, Y, delta_y);
pre_cluster(w, W, delta_w);
pre_cluster(h, H, delta_h);
regularize(x, y, w, h, X, Y, W, H, r);
write_file(X, Y, W, H, r, OUT_PATH);MIT License