API설명 상세 : https://nlesc.github.io/python-pcl/
1. Modules (Library)
pcl_filters - 3D 점군 데이터에서 이상값과 노이즈 제거 등의 필터링 pcl_features - 점군 데이터로부터 3D 특징 추정 (feature estimation) 을 위한 수많은 자료 구조와 방법들 (surface normals, curvatures, boundary point estimation, moment invariants, principal curvatures, PFH & FPFH descriptors, spin images, integral images, NARF descriptors, RIFT, RSD, VFH, SHOT 등) pcl_keypoints - Keypoint (or interest point) 을 검출하는 알고리즘 구현 (BRISK, Harris Corner, NARF, SIFT, SUSAN 등) pcl_registration - 여러 데이터셋을 합쳐 큰 모델로 만드는 registration 작업 (ICP 등) pcl_kdtree - 빠른 최근거리 이웃을 탐색하는 FLANN 라이브러리를 사용한 kdtree 자료 구조 pcl_octree - 점군 데이터로부터 계층 트리 구조를 구성하는 방법 pcl_segmentation - 점군으로부터 클러스터들로 구분하는 알고리즘들 pcl_sample_consensus - 선, 평면, 실린더 등의 모델 계수 추정을 위한 RANSAC 등의 알고리즘들 pcl_surface - 3D 표면 복원 기법들 (meshing, convex hulls, Moving Least Squares 등) pcl_range_image - range image (or depth map) 을 나타내고 처리하는 방법 pcl_io - OpenNI 호환 depth camera 로부터 점군 데이터를 읽고 쓰는 방법 pcl_visualization - 3D 점군 데이터를 처리하는 알고리즘의 결과를 시각화
Module common : common data structures, computing distances/norms, means and covariances, angular conversions, geometric transformations
Module filters: outlier and noise removal mechanisms (eg. PassThrough, voxel grid)
- How 3D Features work in PCL
- Filtering a PointCloud using a PassThrough filter: simple filtering along a specified dimension – that is, cut off values that are either inside or outside a given user range.
Downsampling a PointCloud using a VoxelGrid filter: reduce the number of points – a point cloud dataset, using a voxelized grid approach
Projecting points using a parametric model: project points onto a parametric model (e.g., plane, sphere, etc)
- Extracting indices from a PointCloud: extract a subset of points by a segmentation algorithm
Removing outliers using a StatisticalOutlierRemoval filter: remove noisy measurements, e.g. outliers
Removing outliers using a Conditional or RadiusOutlier removal: remove outliers by ConditionalRemoval filter & RadiusOutlierRemoval filter
Module features: data structures and mechanisms for 3D feature estimation from point cloud data
Module keypoints: keypoint detection algorithms
Module registration: Combining several datasets into a global consistent model is usually performed using a technique called registration
Module kdtree: kd-tree data-structure, Allows for fast nearest neighbor searches.
Module octree: efficient methods for creating a hierarchical tree data structure from point cloud data
- Point Cloud Compression:
- The
pcl_octreeimplementation provides efficient nearest neighbor search routines, such as- "Neighbors within Voxel Search”,
- “K Nearest Neighbor Search”
- “Neighbors within Radius Search”.
Module segmentation: algorithms for segmenting a point cloud into distinct clusters
Module sample_consensus: SAmple Consensus (SAC) methods like RANSAC / models like planes and cylinders
Module surface: deals with reconstructing the original surfaces from 3D scans(eg. hull, a mesh representation or a smoothed/resampled)
Module recognition: algorithms used for Object Recognition
- 3D Object Recognition based on Correspondence Grouping:use Correspondence Grouping algorithms in order to cluster the set of point-to-point correspondences
Module io: reading and writing point cloud data (PCD) files
- The PCD (Point Cloud Data) file format
- Reading PointCloud data from PCD files
- Writing PointCloud data to PCD files
- The OpenNI Grabber Framework in PCL
- Grabbing point clouds from Ensenso cameras
Module visualization: visualize the results
Module Range Image : depth map
- How to create a range image from a point cloud
- How to extract borders from range images: ?NOT Lidar data?
Module common : common data structures, computing distances/norms, means and covariances, angular conversions, geometric transformations
Module search : methods for searching for nearest neighbors using different data structures