cvw::steps

Namespace containing step factory functions for pipeline composition. More...

Functions

	Name
auto	resize(int width, int height, int interpolation =cv::INTER_LINEAR) Creates a resize step with the given dimensions.
auto	flip(int axis) Creates a flip step along the given axis.
auto	rotate(double angle, bool expand =false) Creates a rotation step.
auto	to_gray() Creates a BGR-to-Gray conversion step.
auto	to_gray_rgb() Creates an RGB-to-Gray conversion step.
auto	to_bgr() Creates a Gray-to-BGR conversion step.
auto	yuyv_to_gray() Creates a YUYV-to-Gray conversion step.
auto	gaussian_blur(int kernel_size, double sigma =0.0) Creates a Gaussian blur step.
template <is_pixel_format F&gt; auto	median_blur(int kernel_size) Creates a median blur step for single-channel images.
auto	canny(double lo, double hi, int aperture =3) Creates a Canny edge detection step.
auto	sobel(int dx, int dy, int kernel_size =3) Creates a Sobel derivative step.
auto	threshold(double thresh, double max_val, int type =cv::THRESH_BINARY) Creates a global threshold step.
auto	adaptive_threshold(int block_size, double C, int method =cv::ADAPTIVE_THRESH_GAUSSIAN_C) Creates an adaptive threshold step.
auto	normalize(float min_val =0.0f, float max_val =1.0f) Creates a min-max normalization step.
auto	save(std::string path) Creates a save-to-file sink step.

Detailed Description

Namespace containing step factory functions for pipeline composition.

Each factory returns a callable that accepts an Image or Expected and returns an Expected with automatic error propagation.

Functions Documentation

function resize

inline auto resize(
    int width,
    int height,
    int interpolation =cv::INTER_LINEAR
)

Creates a resize step with the given dimensions.

Parameters:

• width Target width in pixels. Must be positive.
• height Target height in pixels. Must be positive.
• interpolation OpenCV interpolation flag. Default: LINEAR.

Return: A callable step that resizes an image.

Since: 1.0.0

function flip

inline auto flip(
    int axis
)

Creates a flip step along the given axis.

Parameters:

• axis Flip axis (0 = vertical, 1 = horizontal, -1 = both).

Return: A callable step that flips an image.

Since: 1.0.0

function rotate

inline auto rotate(
    double angle,
    bool expand =false
)

Creates a rotation step.

Parameters:

• angle Rotation angle in degrees. Positive = counter-clockwise.
• expand Whether to expand the output to hold the full rotated image. Default: false.

Return: A callable step that rotates an image.

Since: 1.0.0

function to_gray

inline auto to_gray()

Creates a BGR-to-Gray conversion step.

Return: A callable step converting Image<BGR> to Image<Gray>.

Since: 1.0.0

function to_gray_rgb

inline auto to_gray_rgb()

Creates an RGB-to-Gray conversion step.

Return: A callable step converting Image<RGB> to Image<Gray>.

Since: 1.0.0

function to_bgr

inline auto to_bgr()

Creates a Gray-to-BGR conversion step.

Return: A callable step converting Image<Gray> to Image<BGR>.

Since: 1.0.0

function yuyv_to_gray

inline auto yuyv_to_gray()

Creates a YUYV-to-Gray conversion step.

Return: A callable step converting ImageView<YUYV> to Image<Gray>.

Since: 1.0.0

function gaussian_blur

inline auto gaussian_blur(
    int kernel_size,
    double sigma =0.0
)

Creates a Gaussian blur step.

Parameters:

• kernel_size Side length of the square kernel. Must be odd and positive. Unit: pixels.
• sigma Standard deviation. 0.0 means computed from kernel size. Default: 0.0.

Return: A callable step that applies Gaussian blur.

Since: 1.0.0

function median_blur

template <is_pixel_format F>
inline auto median_blur(
    int kernel_size
)

Creates a median blur step for single-channel images.

Parameters:

• kernel_size Side length of the square kernel. Must be odd. Unit: pixels.

Template Parameters:

• F Pixel format with exactly 1 channel.

Return: A callable step that applies median blur.

Since: 1.0.0

function canny

inline auto canny(
    double lo,
    double hi,
    int aperture =3
)

Creates a Canny edge detection step.

Parameters:

• lo Lower hysteresis threshold.
• hi Upper hysteresis threshold.
• aperture Aperture size for the Sobel operator. Default: 3.

Return: A callable step that applies Canny edge detection.

Since: 1.0.0

function sobel

inline auto sobel(
    int dx,
    int dy,
    int kernel_size =3
)

Creates a Sobel derivative step.

Parameters:

• dx Order of the x-derivative.
• dy Order of the y-derivative.
• kernel_size Aperture size. Default: 3.

Return: A callable step that computes Sobel derivatives.

Since: 1.0.0

function threshold

inline auto threshold(
    double thresh,
    double max_val,
    int type =cv::THRESH_BINARY
)

Creates a global threshold step.

Parameters:

• thresh Threshold value.
• max_val Maximum value for THRESH_BINARY. Default: 0.
• type OpenCV threshold type. Default: THRESH_BINARY.

Return: A callable step that applies thresholding.

Since: 1.0.0

function adaptive_threshold

inline auto adaptive_threshold(
    int block_size,
    double C,
    int method =cv::ADAPTIVE_THRESH_GAUSSIAN_C
)

Creates an adaptive threshold step.

Parameters:

• block_size Size of the local neighborhood. Must be odd >= 3. Unit: pixels.
• C Constant subtracted from the mean. May be negative.
• method Adaptive method. Default: ADAPTIVE_THRESH_GAUSSIAN_C.

Return: A callable step that applies adaptive thresholding.

Since: 1.0.0

function normalize

inline auto normalize(
    float min_val =0.0f,
    float max_val =1.0f
)

Creates a min-max normalization step.

Parameters:

• min_val Lower bound of the output range. Default: 0.0f.
• max_val Upper bound of the output range. Default: 1.0f.

Return: A callable step normalizing Image<Gray> to Image<Float1>.

Since: 1.0.0

function save

inline auto save(
    std::string path
)

Creates a save-to-file sink step.

Parameters:

• path Output file path.

Return: A callable step that saves an image and returns expected<void, AlgorithmError>.

Since: 1.0.0

---

Updated on 2026-05-17 at 13:22:38 +0000