Struct FloatImageView 

pub struct FloatImageView<'a> { /* private fields */ }

A borrowed, validated, row-major f32 image, e.g. a thermal frame in °C.

The f32 detector and tracker work in the image’s own unit: no range normalization and no 8-bit conversion takes place. Validation guarantees that every value is finite and at most MAX_ABS_VALUE in magnitude, so no intermediate can overflow and no NaN can reach the score comparisons. Replace invalid pixels (e.g. dead sensor pixels) before creating the view.

Implementations

impl<'a> FloatImageView<'a>

pub const MAX_ABS_VALUE: f32 = 16_777_216.0

Largest accepted pixel magnitude, 2^24. Within it, Scharr gradients, structure-tensor sums and their determinant stay finite.

pub fn new( pixels: &'a [f32], width: u32, height: u32, ) -> Result<Self, FloatImageError>

Validates pixels as a width x height row-major image.

Costs one pass over the pixels.

pub fn width(&self) -> u32

Width in pixels.

pub fn height(&self) -> u32

Height in pixels.

pub fn dimensions(&self) -> (u32, u32)

(width, height) in pixels.

pub fn pixels(&self) -> &'a [f32]

The validated pixels, row by row.

Trait Implementations

impl<'a> Clone for FloatImageView<'a>

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<'a> Copy for FloatImageView<'a>

impl<'a> Debug for FloatImageView<'a>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.