Struct PointCloud 

pub struct PointCloud {
    pub points: Vec<[f64; 3]>,
    pub normals: Vec<[f64; 3]>,
    pub colors: Vec<[f32; 3]>,
}

A collection of 3-D points with optional per-point normals and colors.

Fields

points: Vec<[f64; 3]>

XYZ positions.

normals: Vec<[f64; 3]>

Per-point surface normals (may be empty).

colors: Vec<[f32; 3]>

Per-point RGB colors in [0, 1] (may be empty).

Implementations

impl PointCloud

pub fn new() -> Self

Creates an empty point cloud.

pub fn add_point(&mut self, p: [f64; 3])

Appends a point.

pub fn len(&self) -> usize

Returns the number of points.

pub fn is_empty(&self) -> bool

Returns true if the cloud is empty.

pub fn bounding_box(&self) -> ([f64; 3], [f64; 3])

Returns the axis-aligned bounding box as (min, max).

pub fn centroid(&self) -> [f64; 3]

Returns the centroid (mean position) of all points.

pub fn translate(&mut self, offset: [f64; 3])

Translates all points by offset.

pub fn scale_uniform(&mut self, s: f64)

Uniformly scales all points around the origin.

pub fn from_grid( nx: usize, ny: usize, height_fn: impl Fn(f64, f64) -> f64, dx: f64, ) -> Self

Builds a height-field point cloud from a regular nx × ny grid.

Points are placed at (i * dx, j * dx, height_fn(i * dx, j * dx)).

impl PointCloud

pub fn from_points(points: Vec<[f64; 3]>) -> Self

Constructs a PointCloud from a vec of points (expansion API).

pub fn estimate_normals_pca(&mut self, k: usize)

Estimate per-point normals using PCA of the k nearest neighbors. Populates self.normals.

pub fn voxel_downsample(&self, voxel_size: f64) -> Self

Return a new downsampled cloud using a voxel grid of side voxel_size.

pub fn statistical_outlier_removal(&self, k: usize, std_ratio: f64) -> Self

Return a new cloud with statistical outliers removed.

impl PointCloud

pub fn farthest_point_sample(&self, k: usize) -> Self

Farthest Point Sampling: select k well-spread points from this cloud.

Returns a new PointCloud with the selected points (normals / colors are copied if present).

pub fn fit_plane_ransac( &self, n_iter: usize, threshold: f64, ) -> Option<RansacPlaneResult>

Fit a plane to the point cloud using RANSAC.

Returns None if fewer than 3 points are present.

pub fn pca_obb(&self) -> ([[f64; 3]; 3], [f64; 3], [f64; 3])

Compute the PCA-based OBB of this cloud.

Returns (principal_axes, half_extents, center).

pub fn compute_principal_curvatures(&self, k: usize) -> Vec<(f64, f64)>

Compute PCA-based principal curvatures for each point.

For each point a k-NN neighbourhood is gathered, the 3×3 covariance matrix is formed, and the two smallest eigenvalues (λ₁ ≤ λ₂) are returned as the approximate principal curvatures κ₁ and κ₂.

Returns a Vec<(f64, f64)> of (kappa1, kappa2) for every point.

pub fn icp_register(&self, target: &PointCloud) -> IcpResult

Align this cloud onto target using ICP (point-to-point) and return the registration result (rotation + translation + error).

Calls the existing IcpRegistration::align helper.

Trait Implementations

impl Default for PointCloud

fn default() -> Self

Returns the “default value” for a type.