LiangliangNan
commented
9 months ago The basic usage is
auto cloud = dynamic_cast<PointCloud *>(model);
auto normals = cloud->get_vertex_property<vec3>("v:normal");
if (!normals) {
std::cerr << "Plane extraction using RANSAC requires normal information but it is not available" << std::endl;
return false;
}
PrimitivesRansac algo;
algo.add_primitive_type(PrimitivesRansac::PLANE); // Tell the algorithm to extract planes only (you can also add other types)
// you can try different parameters of RANSAC (usually you don't need to tune them)
const int num = algo.detect(cloud, 200, 0.005f, 0.02f, 0.8f, 0.001f);
if (num > 0) {
std::cout << num << " primitives extracted" << std::endl;
... Then the results are stored as per-point properties:
- "v:primitive_type" (one of PLANE, SPHERE, CYLINDER, CONE, TORUS, and UNKNOWN)
- "v:primitive_index" (-1, 0, 1, 2...). -1 meaning a vertex does not belong to any primitive (thus its primitive_type must be UNKNOWN.
}Please check the example at tutorials/Tutorial_703_Cloud_PlaneExtraction and also point_cloud_ransac.h for more details.
zeyadmahmoudd
Hello there!
I hope this message finds you well. I'm currently working on a 3D reconstruction project, and I'm in need of RANSAC primitive extraction from Mapple. I'm running reconstructions on multiple files and aiming to automate this process.
I'd greatly appreciate your guidance on how to integrate RANSAC primitive extraction into my project. If you have any insights or recommendations on how I can achieve automation, I would be grateful for your assistance.