ThreeCrate Algorithms

High-performance algorithms for 3D point cloud and mesh processing.

Features

• Point Cloud Processing: Filtering, downsampling, and outlier removal
• Registration: ICP (Iterative Closest Point) algorithm for point cloud alignment
• Segmentation: RANSAC plane segmentation and clustering algorithms
• Spatial Queries: K-nearest neighbor search and spatial indexing
• Normal Estimation: Surface normal computation for point clouds
• Parallel Processing: Multi-threaded algorithms using Rayon

Usage

Add this to your Cargo.toml:

[dependencies]
threecrate-algorithms = "0.1.0"
threecrate-core = "0.1.0"

Example

use threecrate_core::{PointCloud, Point3f};
use threecrate_algorithms::{icp_registration, estimate_normals, plane_segmentation_ransac};

// Load or create point clouds
let source = PointCloud::from_points(vec![/* points */]);
let target = PointCloud::from_points(vec![/* points */]);

// ICP registration
let result = icp_registration(&source, &target, 50, 0.001, 1.0)?;
println!("Registration converged: {}", result.converged);

// Estimate normals
let normals = estimate_normals(&source, 10)?;

// RANSAC plane segmentation
let plane_result = plane_segmentation_ransac(&source, 1000, 0.01)?;
println!("Found {} inliers", plane_result.inliers.len());

Algorithms

• ICP Registration: Point cloud alignment using iterative closest point
• RANSAC Segmentation: Robust plane fitting and outlier detection
• Normal Estimation: Surface normal computation using local neighborhoods
• Filtering: Statistical outlier removal and radius filtering
• Spatial Indexing: KD-tree and R-tree based spatial queries

License

This project is licensed under either of

• Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.