The RANSAC algorithm on the basis of the search tree and space.
Algorithms ideas:
1, point cloud pumping Greece, the normal estimate
2, points out index storage statements
3, the detection plane
for (size_t i = 0; i < cloudTemp->points.size(); i++)
{
Out point is determined;
if (near retrieve a certain amount of points)
{
To determine whether the search point set to meet the requirements;
Points stored search;
}
RANSAC plane fitting (RANSAC plane calculated model parameters);
Determining the correctness of fitting plane;
/*
* Calculated using the fit plane to the surface from the point cloud, the tolerance setting is determined whether the point cloud in a plane
*/
for (size_t j = 0; j < tr_cloud->points.size(); j++)
{
Analyzing out point;
Analyzing the plane normal to the consistency point method (the angle between two spatial vectors Solution);
Point set within the memory plane;
Update point out;
}
// plane noise outliers removed
//...
Storing data plane;
}
Results are as follows:
