ContourDetector structure extracts polyline contours from 8-bit gray images. More...

#include <SimdContour.hpp>

Public Types
typedef A< uint8_t >	Allocator

typedef Simd::View< A >	View

typedef Simd::Point< ptrdiff_t >	Size

typedef Simd::Point< ptrdiff_t >	Point

typedef Rectangle< ptrdiff_t >	Rect

typedef std::vector< Point >	Contour

typedef std::vector< Contour >	Contours

Public Member Functions
void	Init (Size size)

bool	Detect (const View &src, Contours &contours, const View &mask=View(), uint8_t indexMin=3, const Rect &roi=Rect(), int gradientThreshold=40, int anchorThreshold=0, int anchorScanInterval=2, int minSegmentLength=2)

Detailed Description

template<template< class > class A>
struct Simd::ContourDetector< A >

ContourDetector structure extracts polyline contours from 8-bit gray images.

The structure is a C++ helper for contour extraction. It calculates packed Sobel metrics, optionally filters them by a mask and a rectangular region of interest, selects contour anchors, and routes connected edge points into contours. Call Init() once for the required image size, then call Detect() for one or more images of the same size.

Each contour is an ordered vector of image points. Adjacent points can be connected with line segments, as shown in the example below.

Using example:

#include "Simd/SimdContour.hpp"
#include "Simd/SimdDrawing.hpp"

int main()
{
    typedef Simd::ContourDetector<Simd::Allocator> ContourDetector;

    ContourDetector::View image;
    image.Load("../../data/image/face/lena.pgm");

    ContourDetector contourDetector;

    contourDetector.Init(image.Size());

    ContourDetector::Contours contours;
    contourDetector.Detect(image, contours);

    for (size_t i = 0; i < contours.size(); ++i)
    {
        for (size_t j = 1; j < contours[i].size(); ++j)
            Simd::DrawLine(image, contours[i][j - 1], contours[i][j], uint8_t(255));
    }
    image.Save("result.pgm");

    return 0;
}

Member Typedef Documentation

◆ Allocator

typedef A<uint8_t> Allocator

Allocator type used by internal and external images.

◆ View

typedef Simd::View<A> View

Image type used by the detector. Input and mask images for Detect() must be View::Gray8.

◆ Size

typedef Simd::Point<ptrdiff_t> Size

Image size type.

◆ Point

typedef Simd::Point<ptrdiff_t> Point

Image point type.

◆ Rect

typedef Rectangle<ptrdiff_t> Rect

Rectangle type used to specify a region of interest.

◆ Contour

typedef std::vector<Point> Contour

Ordered chain of contour points in source image coordinates.

◆ Contours

typedef std::vector<Contour> Contours

Vector of detected contours.

Member Function Documentation

◆ Init()

void Init ( Size size )

Allocates internal buffers and prepares ContourDetector to process Gray8 images of the given size.

The same initialized detector can be reused for several images with the same size.

Parameters

[in] size - a size of input images.

◆ Detect()

bool Detect	(	const View &	src,
		Contours &	contours,
		const View &	mask = `View()`,
		uint8_t	indexMin = `3`,
		const Rect &	roi = `Rect()`,
		int	gradientThreshold = `40`,
		int	anchorThreshold = `0`,
		int	anchorScanInterval = `2`,
		int	minSegmentLength = `2`
	)

Detects contours in the given Gray8 image.

The source image must have the size passed to Init(). If a mask is specified, it must also be a Gray8 image of this size. Detected contours are appended to contours, so clear this vector before the call if previous results must be discarded.