localization.cpp

Go to the documentation of this file.

#include <agile_grasp/localization.h>

std::vector<GraspHypothesis> Localization::localizeHands(const PointCloud::Ptr& cloud_in, int size_left,
        const std::vector<int>& indices, bool calculates_antipodal, bool uses_clustering)
{
        double t0 = omp_get_wtime();
        std::vector<GraspHypothesis> hand_list;

        // set camera source for all points (0 = left, 1 = right)
        std::cout << "Generating camera sources for " << cloud_in->size() << " points ...\n";
        Eigen::VectorXi pts_cam_source(cloud_in->size());
        if (size_left == cloud_in->size())
                pts_cam_source << Eigen::VectorXi::Zero(size_left);
        else
                pts_cam_source << Eigen::VectorXi::Zero(size_left), Eigen::VectorXi::Ones(cloud_in->size() - size_left);
                
        // remove NAN points from the cloud
        std::vector<int> nan_indices;
        pcl::removeNaNFromPointCloud(*cloud_in, *cloud_in, nan_indices);

        // reduce point cloud to workspace
        std::cout << "Filtering workspace ...\n";
        PointCloud::Ptr cloud(new PointCloud);
        filterWorkspace(cloud_in, pts_cam_source, cloud, pts_cam_source);
        std::cout << " " << cloud->size() << " points left\n";

        // store complete cloud for later plotting
        PointCloud::Ptr cloud_plot(new PointCloud);
        *cloud_plot = *cloud;
        *cloud_ = *cloud;

        // voxelize point cloud
        std::cout << "Voxelizing point cloud\n";
        double t1_voxels = omp_get_wtime();
        voxelizeCloud(cloud, pts_cam_source, cloud, pts_cam_source, 0.003);
        double t2_voxels = omp_get_wtime() - t1_voxels;
        std::cout << " Created " << cloud->points.size() << " voxels in " << t2_voxels << " sec\n";

        // plot camera source for each point in the cloud
        if (plots_camera_sources_)
                plot_.plotCameraSource(pts_cam_source, cloud);

        if (uses_clustering)
        {
    std::cout << "Finding point cloud clusters ... \n";
        
                // Create the segmentation object for the planar model and set all the parameters
                pcl::SACSegmentation<pcl::PointXYZ> seg;
                pcl::PointIndices::Ptr inliers(new pcl::PointIndices);
                pcl::ModelCoefficients::Ptr coefficients(new pcl::ModelCoefficients);
                pcl::PointCloud<pcl::PointXYZ>::Ptr cloud_plane(new pcl::PointCloud<pcl::PointXYZ>());
                seg.setOptimizeCoefficients(true);
                seg.setModelType(pcl::SACMODEL_PLANE);
                seg.setMethodType(pcl::SAC_RANSAC);
                seg.setMaxIterations(100);
                seg.setDistanceThreshold(0.01);

                // Segment the largest planar component from the remaining cloud
                seg.setInputCloud(cloud);
                seg.segment(*inliers, *coefficients);
                if (inliers->indices.size() == 0)
                {
                        std::cout << " Could not estimate a planar model for the given dataset." << std::endl;
                        hand_list.resize(0);
                        return hand_list;
                }
    
    std::cout << " PointCloud representing the planar component: " << inliers->indices.size()
      << " data points." << std::endl;

                // Extract the nonplanar inliers
                pcl::ExtractIndices<pcl::PointXYZ> extract;
                extract.setInputCloud(cloud);
                extract.setIndices(inliers);
                extract.setNegative(true);
                std::vector<int> indices_cluster;
                extract.filter(indices_cluster);
                PointCloud::Ptr cloud_cluster(new PointCloud);
                cloud_cluster->points.resize(indices_cluster.size());
                Eigen::VectorXi cluster_cam_source(indices_cluster.size());
                for (int i = 0; i < indices_cluster.size(); i++)
                {
                        cloud_cluster->points[i] = cloud->points[indices_cluster[i]];
                        cluster_cam_source[i] = pts_cam_source[indices_cluster[i]];
                }
                cloud = cloud_cluster;
                *cloud_plot = *cloud;
                std::cout << " PointCloud representing the non-planar component: " << cloud->points.size()
      << " data points." << std::endl;
        }

        // draw down-sampled and workspace reduced cloud
        cloud_plot = cloud;
  
  // find hand configurations
  HandSearch hand_search(finger_width_, hand_outer_diameter_, hand_depth_, 
    hand_height_, init_bite_, num_threads_, num_samples_, plots_hands_);
        hand_list = hand_search.findHands(cloud, pts_cam_source, indices, cloud_plot, 
    calculates_antipodal, uses_clustering);

        // remove hands at boundaries of workspace
        if (filters_boundaries_)
  {
    std::cout << "Filtering out hands close to workspace boundaries ...\n";
    hand_list = filterHands(hand_list);
    std::cout << " # hands left: " << hand_list.size() << "\n";
  }

        double t2 = omp_get_wtime();
        std::cout << "Hand localization done in " << t2 - t0 << " sec\n";

        if (plots_hands_)
                plot_.plotHands(hand_list, cloud_plot, "");

        return hand_list;
}

std::vector<GraspHypothesis> Localization::predictAntipodalHands(
        const std::vector<GraspHypothesis>& hand_list, const std::string& svm_filename)
{
        double t0 = omp_get_wtime();
        std::vector<GraspHypothesis> antipodal_hands;
        Learning learn(num_threads_);
        Eigen::Matrix<double, 3, 2> cams_mat;
        cams_mat.col(0) = cam_tf_left_.block<3, 1>(0, 3);
        cams_mat.col(1) = cam_tf_right_.block<3, 1>(0, 3);
        antipodal_hands = learn.classify(hand_list, svm_filename, cams_mat);
        std::cout << " runtime: " << omp_get_wtime() - t0 << " sec\n";
        std::cout << antipodal_hands.size() << " antipodal hand configurations found\n"; 
  if (plots_hands_)
                plot_.plotHands(hand_list, antipodal_hands, cloud_, "Antipodal Hands");
        return antipodal_hands;
}

std::vector<GraspHypothesis> Localization::localizeHands(const std::string& pcd_filename_left,
        const std::string& pcd_filename_right, bool calculates_antipodal, bool uses_clustering)
{
        std::vector<int> indices(0);
        return localizeHands(pcd_filename_left, pcd_filename_right, indices, calculates_antipodal, uses_clustering);
}

std::vector<GraspHypothesis> Localization::localizeHands(const std::string& pcd_filename_left,
        const std::string& pcd_filename_right, const std::vector<int>& indices, bool calculates_antipodal,
        bool uses_clustering)
{
        double t0 = omp_get_wtime();

        // load point clouds
        PointCloud::Ptr cloud_left(new PointCloud);
        if (pcl::io::loadPCDFile<pcl::PointXYZ>(pcd_filename_left, *cloud_left) == -1) //* load the file
        {
                PCL_ERROR("Couldn't read pcd_filename_left file \n");
                std::vector<GraspHypothesis> hand_list(0);
                return hand_list;
        }
        if (pcd_filename_right.length() > 0)
                std::cout << "Loaded left point cloud with " << cloud_left->width * cloud_left->height << " data points.\n";
        else
                std::cout << "Loaded point cloud with " << cloud_left->width * cloud_left->height << " data points.\n";

        PointCloud::Ptr cloud_right(new PointCloud);
        if (pcd_filename_right.length() > 0)
        {
                if (pcl::io::loadPCDFile<pcl::PointXYZ>(pcd_filename_right, *cloud_right) == -1) //* load the file
                {
                        PCL_ERROR("Couldn't read pcd_filename_right file \n");
                        std::vector<GraspHypothesis> hand_list(0);
                        return hand_list;
                }
                std::cout << "Loaded right point cloud with " << cloud_right->width * cloud_right->height << " data points.\n";
                std::cout << "Loaded both clouds in " << omp_get_wtime() - t0 << " sec\n";
        }
        
        // concatenate point clouds
        std::cout << "Concatenating point clouds ...\n";
        PointCloud::Ptr cloud(new PointCloud);
        *cloud = *cloud_left + *cloud_right;

        return localizeHands(cloud, cloud_left->size(), indices, calculates_antipodal, uses_clustering);
}

void Localization::filterWorkspace(const PointCloud::Ptr& cloud_in, const Eigen::VectorXi& pts_cam_source_in,
        PointCloud::Ptr& cloud_out, Eigen::VectorXi& pts_cam_source_out)
{
        std::vector<int> indices(cloud_in->points.size());
        int c = 0;
        PointCloud::Ptr cloud(new PointCloud);
//  std::cout << "workspace_: " << workspace_.transpose() << "\n";

        for (int i = 0; i < cloud_in->points.size(); i++)
        {
//    std::cout << "i: " << i << "\n";
                const pcl::PointXYZ& p = cloud_in->points[i];
                if (p.x >= workspace_(0) && p.x <= workspace_(1) && p.y >= workspace_(2) && p.y <= workspace_(3)
                                && p.z >= workspace_(4) && p.z <= workspace_(5))
                {
                        cloud->points.push_back(p);
                        indices[c] = i;
                        c++;
                }
        }

        Eigen::VectorXi pts_cam_source(c);
        for (int i = 0; i < pts_cam_source.size(); i++)
        {
                pts_cam_source(i) = pts_cam_source_in(indices[i]);
        }

        cloud_out = cloud;
        pts_cam_source_out = pts_cam_source;
}

void Localization::voxelizeCloud(const PointCloud::Ptr& cloud_in, const Eigen::VectorXi& pts_cam_source_in,
        PointCloud::Ptr& cloud_out, Eigen::VectorXi& pts_cam_source_out, double cell_size)
{
        Eigen::Vector3d min_left, min_right;
        min_left << 10000, 10000, 10000;
        min_right << 10000, 10000, 10000;
        Eigen::Matrix3Xd pts(3, cloud_in->points.size());
        int num_left = 0;
        int num_right = 0;
        for (int i = 0; i < cloud_in->points.size(); i++)
        {
                if (pts_cam_source_in(i) == 0)
                {
                        if (cloud_in->points[i].x < min_left(0))
                                min_left(0) = cloud_in->points[i].x;
                        if (cloud_in->points[i].y < min_left(1))
                                min_left(1) = cloud_in->points[i].y;
                        if (cloud_in->points[i].z < min_left(2))
                                min_left(2) = cloud_in->points[i].z;
                        num_left++;
                }
                else if (pts_cam_source_in(i) == 1)
                {
                        if (cloud_in->points[i].x < min_right(0))
                                min_right(0) = cloud_in->points[i].x;
                        if (cloud_in->points[i].y < min_right(1))
                                min_right(1) = cloud_in->points[i].y;
                        if (cloud_in->points[i].z < min_right(2))
                                min_right(2) = cloud_in->points[i].z;
                        num_right++;
                }
                pts.col(i) = cloud_in->points[i].getVector3fMap().cast<double>();
        }

        // find the cell that each point falls into
        std::set<Eigen::Vector3i, Localization::UniqueVectorComparator> bins_left;
        std::set<Eigen::Vector3i, Localization::UniqueVectorComparator> bins_right;
        int prev;
        for (int i = 0; i < pts.cols(); i++)
        {
                if (pts_cam_source_in(i) == 0)
                {
                        Eigen::Vector3i v = floorVector((pts.col(i) - min_left) / cell_size);
                        bins_left.insert(v);
                        prev = bins_left.size();
                }
                else if (pts_cam_source_in(i) == 1)
                {
                        Eigen::Vector3i v = floorVector((pts.col(i) - min_right) / cell_size);
                        bins_right.insert(v);
                }
        }

        // calculate the cell values
        Eigen::Matrix3Xd voxels_left(3, bins_left.size());
        Eigen::Matrix3Xd voxels_right(3, bins_right.size());
        int i = 0;
        for (std::set<Eigen::Vector3i, Localization::UniqueVectorComparator>::iterator it = bins_left.begin();
                        it != bins_left.end(); it++)
        {
                voxels_left.col(i) = (*it).cast<double>();
                i++;
        }
        i = 0;
        for (std::set<Eigen::Vector3i, Localization::UniqueVectorComparator>::iterator it = bins_right.begin();
                        it != bins_right.end(); it++)
        {
                voxels_right.col(i) = (*it).cast<double>();
                i++;
        }

        voxels_left.row(0) = voxels_left.row(0) * cell_size
                        + Eigen::VectorXd::Ones(voxels_left.cols()) * min_left(0);
        voxels_left.row(1) = voxels_left.row(1) * cell_size
                        + Eigen::VectorXd::Ones(voxels_left.cols()) * min_left(1);
        voxels_left.row(2) = voxels_left.row(2) * cell_size
                        + Eigen::VectorXd::Ones(voxels_left.cols()) * min_left(2);
        voxels_right.row(0) = voxels_right.row(0) * cell_size
                        + Eigen::VectorXd::Ones(voxels_right.cols()) * min_right(0);
        voxels_right.row(1) = voxels_right.row(1) * cell_size
                        + Eigen::VectorXd::Ones(voxels_right.cols()) * min_right(1);
        voxels_right.row(2) = voxels_right.row(2) * cell_size
                        + Eigen::VectorXd::Ones(voxels_right.cols()) * min_right(2);

        PointCloud::Ptr cloud(new PointCloud);
        cloud->resize(bins_left.size() + bins_right.size());
        Eigen::VectorXi pts_cam_source(bins_left.size() + bins_right.size());
        for (int i = 0; i < bins_left.size(); i++)
        {
                pcl::PointXYZ p;
                p.x = voxels_left(0, i);
                p.y = voxels_left(1, i);
                p.z = voxels_left(2, i);
                cloud->points[i] = p;
                pts_cam_source(i) = 0;
        }
        for (int i = 0; i < bins_right.size(); i++)
        {
                pcl::PointXYZ p;
                p.x = voxels_right(0, i);
                p.y = voxels_right(1, i);
                p.z = voxels_right(2, i);
                cloud->points[bins_left.size() + i] = p;
                pts_cam_source(bins_left.size() + i) = 1;
        }

        cloud_out = cloud;
        pts_cam_source_out = pts_cam_source;
}

Eigen::Vector3i Localization::floorVector(const Eigen::Vector3d& a)
{
        Eigen::Vector3i b;
        b << floor(a(0)), floor(a(1)), floor(a(2));
        return b;
}

std::vector<GraspHypothesis> Localization::filterHands(const std::vector<GraspHypothesis>& hand_list)
{
        const double MIN_DIST = 0.02;

        std::vector<GraspHypothesis> filtered_hand_list;

        for (int i = 0; i < hand_list.size(); i++)
        {
                const Eigen::Vector3d& center = hand_list[i].getGraspSurface();
                int k;
                for (k = 0; k < workspace_.size(); k++)
                {
                        if (fabs((center(floor(k / 2.0)) - workspace_(k))) < MIN_DIST)
                        {
                                break;
                        }
                }
                if (k == workspace_.size())
                {
                        filtered_hand_list.push_back(hand_list[i]);
                }
        }

        return filtered_hand_list;
}

std::vector<Handle> Localization::findHandles(const std::vector<GraspHypothesis>& hand_list, int min_inliers,
        double min_length, bool is_plotting)
{
        HandleSearch handle_search;
        std::vector<Handle> handles = handle_search.findHandles(hand_list, min_inliers, min_length);
        if (is_plotting)
                handle_search.plotHandles(handles, cloud_, "");
        return handles;
}