The ContourDetector extracts all contours of a given image. More...

#include <ContourDetector.h>

Inheritance diagram for icl::cv::ContourDetector:

Public Types
enum	Algorithm { Fast, Accurate, AccurateWithHierarchy }
	contour tracing algorithm used More...
Public Member Functions
	ContourDetector (const icl8u thresh=128, Algorithm a=Fast)
virtual	~ContourDetector ()
void	drawAllContours (core::ImgBase *img, const icl64f &value)
const std::vector< Contour > &	detect (const core::ImgBase *image)
const std::vector< Contour > &	detect (core::Img8u &image)
void	setThreshold (const icl8u &threshold)
	sets new binarization threshold
void	setAlgorithm (Algorithm a)
	sets whether a contour hierarchy is created
Private Attributes
Data *	m_data
	internal data pointer

Detailed Description

The ContourDetector extracts all contours of a given image.

Internally, the implementation works on binary images only, but a compatiblity layer is provided that allows for working on arbitraryly-typed images by internally converting the input image before the compuation takes place. The algorithms alters the values of the input image for performance reasons. If a const image is passed to the ContourDetector::detect method, the image is copied/converted before

Contour Hierarchy

The ContourDetector can be set up to also extract a countour hierarchy.

Algorithms

Internally 2 different contour tracing algorithms are implemented. While the "Fast" method uses a 4-point neighbourhood, the Accurate method uses a 8-point neighborhood an can also optionally be used to obtain a region hierarchy. The fast method uses its own memory allocator to improve runtime performance.

Member Enumeration Documentation

enum icl::cv::ContourDetector::Algorithm

contour tracing algorithm used

Enumerator:

Fast	fast contour detection algorithm (using 4-point neighbourhood)
Accurate	accurate contour detection algorithm (using 8-point neighborhood)
AccurateWithHierarchy	same as Accurate, but with Hierarchy estimation

Constructor & Destructor Documentation

icl::cv::ContourDetector::ContourDetector	(	const icl8u	thresh = `128`,
		Algorithm	a = `Fast`
	)

Parameters:

thesh	threshold for creating the binary image
hierarchy	shows if the relationship of the contours should be calculated

virtual icl::cv::ContourDetector::~ContourDetector ( ) [virtual]

Member Function Documentation

const std::vector<Contour>& icl::cv::ContourDetector::detect ( const core::ImgBase * image )

const std::vector<Contour>& icl::cv::ContourDetector::detect ( core::Img8u & image )

void icl::cv::ContourDetector::drawAllContours	(	core::ImgBase *	img,
		const icl64f &	value
	)

void icl::cv::ContourDetector::setAlgorithm ( Algorithm a )

sets whether a contour hierarchy is created

void icl::cv::ContourDetector::setThreshold ( const icl8u & threshold )

sets new binarization threshold

Member Data Documentation

Data* icl::cv::ContourDetector::m_data [private]

internal data pointer

The documentation for this class was generated from the following file:

/home/jenkins/workspace/icl-manual-trunk/icl-manual/ICLCV/src/ICLCV/ContourDetector.h

Public Types

Public Member Functions