Fatal error with PCL ICP

Question

I'm new to PCL (Point Cloud Library) and just want to apply to ICP to two sets of points. The online code sample for ICP, however, throws a fatal error when I try to run it with Visual Studio 2010 64 bit. I've tried different ways of creating a point cloud but no luck. Fatal error occurs inside icp.setInputTarget, at the line target_ = target.makeShared ();

This is how I create both of my cloud points

pcl::PointCloud<pcl::PointXYZ>::Ptr cloud_in (new pcl::PointCloud<pcl::PointXYZ>);
cloud_in->width    = _head_width+1;
cloud_in->height   = _head_height+1;
cloud_in->is_dense = true;
for(int x=0; x<=width; x++) {
        for(int y=0; y<=height; y++)    {
            float z = depths[x][y];
            pcl::PointXYZ curr_point;
            curr_point.x = x;
            curr_point.y = y;
            curr_point.z = z;
            cloud_in->points.push_back(curr_point);
        }
    }

And this is where error occurs

pcl::IterativeClosestPoint<pcl::PointXYZ, pcl::PointXYZ> icp;
icp_dummy.setInputCloud(cloud_in);
icp_dummy.setInputTarget(cloud_out); /* This throws a fatal error */

Any help will be appreciated

How do you declare/create cloud_out? If you run the example at pointclouds.org/documentation/tutorials/… does it work? — Sassa
Not really, that's the code I tried to use :/ I suspect it might be a 64bit problem — Ege Akpinar

Martijn van Wezel Martijn van Wezel · Accepted Answer · 2015-09-28T08:40:15

I have a few questions what is not right in my eyes: -The depht map is not correct, the values of x,y are not real world coordinates

pcl::PointCloud<pcl::PointXYZ>::Ptr cloud (new pcl::PointCloud<pcl::PointXYZ>);
cloud->points.reserve(depthmap.rows*depthmap.cols);//for speeding up code, whitout it cost more time

cloud->is_dense = true;
//Don't know how it'd defined but try something like this by the depthmap
for(int x=0; x<depthmap.rows; x++) {
        for(int y=0; y<depthmap.cols; y++)    {
            float z = depths[x][y];
            pcl::PointXYZ curr_point;
            curr_point.x = (x - cx) * z / fx; //for speedup calculate inverse of fx and multiply this 
            curr_point.y = (y - cy) * z / fy;//for speedup calculate inverse of fy and multiply this 
            curr_point.z = z;
            cloud->points.push_back(curr_point);
        }
    }

- Also for speeding up thing use PTR (smart pointers)

pcl::PointCloud<pcl::PointXYZ>::Ptr cloud_out;

pcl::IterativeClosestPoint<pcl::PointXYZ, pcl::PointXYZ> icp; //hereyou have to define icp

icp.setInputCloud(cloud_in);//so here icp_dummy needs to be icp

icp.setInputTarget(cloud_out); //so here icp_dummy needs to be icp

// The fatal error must be gone, otherwise change cloud_in to same type
// as cloud_out

Fatal error with PCL ICP

1 Answers