mp2p_icp

A repertory of multi primitive-to-primitive (MP2P) ICP algorithms in C++

• Links
  • Rosindex
  • Website
  • Repository
• C++ API
  • Page Hierarchy
  • Class Hierarchy
  • File Hierarchy
  • Full C++ API

• Standard Documents
  • PACKAGE
  • CHANGELOG
  • README
• ROS Package Dependencies
  • mola_common
  • mrpt_libbase
  • mrpt_libobs
  • mrpt_libposes
  • mrpt_libmaps
  • mrpt_libgui
  • mrpt_libtclap
  • mola_imu_preintegration

README

Distro	Build dev	Build releases	Stable version
ROS 2 Humble (u22.04)		amd64 arm64
ROS 2 Jazzy @ u24.04		amd64 arm64
ROS 2 Kilted @ u24.04		amd64 arm64
ROS 2 Rolling (u24.04)		amd64 arm64

EOL Distro	Last release
ROS1 Noetic (u20.04)
ROS 2 Iron (u22.04)

mp2p_icp

C++ libraries for multi primitive-to-primitive (MP2P) ICP algorithms and flexible point cloud processing pipelines. mp2p_icp is used in the mola_lidar_odometry framework:

• Main library documentation and C++ API

• License: New BSD 3-Clause (Note that each module of MOLA has its own license)

Introduction

The project provides these C++ libraries:

• mp2p_icp_map: Provides the mp2p_icp::metric_map_t generic metric map container and related utilities for working with point cloud data structures. Metric map files with extension *.mm are serializations of instances of this class.

• mp2p_icp: With ICP algorithms, matchers (point-to-point, point-to-plane, point-to-line, covariance-to-covariance, adaptive), solvers (Horn, OLAE, Gauss-Newton), and quality evaluators. It depends on mp2p_icp_map.

• mp2p_icp_filters: With point cloud filtering and manipulation algorithms including decimation, voxelization, statistical outlier removal, edge/plane extraction, and various geometric transformations. It depends on mp2p_icp_map.

Command-Line Applications

The full repository also includes these applications (see apps/ directory):

Data Conversion Tools

• kitti2mm: Converts KITTI dataset LIDAR binary files (.bin) with (X,Y,Z, Intensity) data into mp2p_icp metric map files (.mm). Supports custom layer names, numeric IDs, and label strings.

• txt2mm: Converts plain-text point cloud data (TXT/CSV) into metric map (.mm) files. Supports multiple formats: XYZ, XYZI, XYZIRT, XYZRGB, with configurable column mapping and layer names.

• sm2mm: Converts a simple map (.simplemap) from a SLAM mapping session into a metric map (.mm) using a configurable processing pipeline. Essential for post-processing SLAM outputs into structured metric representations.

Data Export Tools

• mm2ply: Exports metric map layers to PLY point cloud files. Supports both ASCII and binary formats, selective field export, and preserves all point attributes (coordinates, colors, intensities, etc.).

• mm2txt: Exports metric map layers as space-delimited CSV/TXT files with header rows. Ideal for data analysis in spreadsheet applications or custom processing pipelines. Supports selective layer and field export.

• mm2las: Exports metric maps to industry-standard LAS 1.4 format with Point Format 8 support. Includes automatic color mapping, extra dimensions for custom fields. Compatible with CloudCompare, QGIS, and ArcGIS.

Map Processing & Analysis Tools

• mm-filter: Applies mp2p_icp_filters pipelines to metric map files. Can operate in pipeline mode (applying complete YAML-defined filter chains) or rename mode (simply renaming layers). Supports custom plugin loading.

• mm-info: Displays a summary of metric map contents including layers, point counts, and metadata.

• mm-georef: Extracts or injects geo-referencing information between metric map files (.mm) and standalone georeferencing files (.georef).

Visualization Tools

• mm-viewer: GUI application to visualize metric map (.mm) files. Supports loading additional 3D scenes and trajectory files in TUM format.

• icp-log-viewer: Interactive GUI for debugging ICP pipelines. Visualizes ICP log files (.icplog) with autoplay mode and detailed inspection of registration results. Essential for understanding and optimizing ICP algorithm performance.

ICP Execution Tools

• icp-run: Standalone program to execute ICP pipelines from the command line.

SimpleMaps Manipulation

• sm-cli: Swiss-army knife for simple map (.simplemap) manipulation. Commands include: info (analyze contents), cut (extract by keyframe index), trim (extract by bounding box), join (merge maps), level (make horizontal), tf (apply SE(3) transform), export-keyframes (save trajectories as TUM), and export-rawlog (convert to RawLog format).

Key Features and Components

Key C++ classes provided by this project (see full documentation):

Core Data Structures

• mp2p_icp::metric_map_t: A generic data type to store raw or processed point clouds, including support for multiple layers, segmented data, and discrete extracted features.

ICP Algorithms and Matchers

• mp2p_icp::ICP: A uniform API for matching generic point clouds with support for:

  • Point-to-Point matching (various distance metrics)

  • Point-to-Plane matching (planes extracted from point clouds)

  • Point-to-Line matching (edges/lines extracted from point clouds)

  • Covariance-to-Covariance matching (Gaussian distributions)

  • Adaptive matching (automatically selects best matcher)

Solvers

• Horn’s method: Closed-form solution for point-to-point alignment

• OLAE (Optimal Linear Attitude Estimator): For attitude/rotation estimation

• Gauss-Newton: Iterative solver for complex matching scenarios

Filters and Generators

The mp2p_icp_filters library provides extensive filtering capabilities:

• Decimation: FilterDecimate, FilterDecimateVoxels, FilterDecimateAdaptive

• Outlier removal: FilterSOR (Statistical Outlier Removal), FilterVoxelSOR

• Geometric filtering: FilterByRange, FilterBoundingBox, FilterByRing

• Feature extraction: FilterEdgesPlanes, FilterCurvature

• Data manipulation: FilterDeskew, FilterMerge, FilterAdjustTimestamps

• Generators: GeneratorEdgesFromCurvature, GeneratorEdgesFromRangeImage

Quality Evaluators

• QualityEvaluator_PairedRatio: Evaluates alignment quality based on paired point ratios

• QualityEvaluator_Voxels: Voxel-based quality metrics

• QualityEvaluator_RangeImageSimilarity: Range image comparison metrics

Multi-Primitive Pairings

The library supports matching between different geometric primitives (points, planes, lines, covariances), allowing for more robust and accurate registration compared to traditional point-to-point ICP. See the documentation for more details on the pairing strategies and matching algorithms.