Struct KltTracker 

pub struct KltTracker {
    pub config: KltConfig,
}

Pyramidal KLT tracker.

Fields

config: KltConfig

Tracker configuration.

Implementations

impl KltTracker

pub fn new(config: KltConfig) -> Self

Create a new KLT tracker with the given configuration.

pub fn track_features( &self, prev_frame: &[u8], curr_frame: &[u8], width: u32, height: u32, points: &[(f32, f32)], ) -> AlignResult<Vec<Option<(f32, f32)>>>

Track sparse feature points from prev_frame to curr_frame using pyramidal Lucas-Kanade optical flow.

This is the primary high-level API for KLT tracking. It accepts f32 pixel coordinates (common in feature-detection pipelines) and returns Some((x, y)) for successfully tracked points or None when tracking failed (low texture, out-of-bounds, etc.).

Both frames must be single-channel (grayscale), row-major byte images of size width * height.

Errors

Returns an error if image dimensions are inconsistent or the frame is smaller than 8×8 pixels.

pub fn track( &self, prev_image: &[u8], curr_image: &[u8], width: usize, height: usize, points: &[Point2D], ) -> AlignResult<Vec<TrackResult>>

Track a set of points from prev_image to curr_image.

Both images must be single-channel (grayscale), row-major, with the same width and height.

Errors

Returns an error if the image dimensions are inconsistent.

Trait Implementations

impl Default for KltTracker

fn default() -> Self

Returns the “default value” for a type.