VideoDynamicAdaptedFeatureDetector Class Reference

#include <feature_adjuster.h>

Inheritance diagram for VideoDynamicAdaptedFeatureDetector:

List of all members.

Public Member Functions
virtual cv::Ptr < StatefulFeatureDetector >	clone () const
virtual bool	empty () const
	VideoDynamicAdaptedFeatureDetector (cv::Ptr< cv::AdjusterAdapter > adjuster, int min_features=400, int max_features=500, int max_iters=5)
Protected Member Functions
virtual void	detectImpl (const cv::Mat &image, std::vector< cv::KeyPoint > &keypoints, const cv::Mat &mask=cv::Mat()) const
Private Member Functions
VideoDynamicAdaptedFeatureDetector &	operator= (const VideoDynamicAdaptedFeatureDetector &)
	VideoDynamicAdaptedFeatureDetector (const VideoDynamicAdaptedFeatureDetector &)
Private Attributes
cv::Ptr< cv::AdjusterAdapter >	adjuster_
int	escape_iters_
int	max_features_
int	min_features_

---

Detailed Description

In contrast to the original DynamicAdaptedFeatureDetector, this variant is enhanced for processing of video sequences. It is meant to work with the DetectorAdjuster. It keeps the DetectorAdjuster alive, so that the final threshold will be retained throughout detection calls. For video sequences the "good" threshold will in general be similar for successive frames, therefore the last "good" threshold is a good starting point for the next frame.

In case of too many features, this variant will just decrease the threshold of the DetectorAdjuster without triggering redetection, so the user needs to get rid of the superfluous keypoints. Mostly the keypoints are scored anyway, so this avoids costly and unnecessary redetections.

Definition at line 53 of file feature_adjuster.h.

---

Constructor & Destructor Documentation

VideoDynamicAdaptedFeatureDetector::VideoDynamicAdaptedFeatureDetector	(	cv::Ptr< cv::AdjusterAdapter >	adjuster,
		int	min_features = 400,
		int	max_features = 500,
		int	max_iters = 5
	)

Parameters:

adjuster	an AdjusterAdapter that will do the detection and parameter adjustment
max_features	the maximum desired number of features
max_iters	the maximum number of times to try to adjust the feature detector params for the FastAdjuster this can be high, but with Star or Surf this can get time consuming
min_features	the minimum desired features

VideoDynamicAdaptedFeatureDetector::VideoDynamicAdaptedFeatureDetector

(

const VideoDynamicAdaptedFeatureDetector &

)

[private]

---

Member Function Documentation

cv::Ptr< StatefulFeatureDetector > VideoDynamicAdaptedFeatureDetector::clone

(

)

const [virtual]

Implements StatefulFeatureDetector.

Definition at line 166 of file feature_adjuster.cpp.

void VideoDynamicAdaptedFeatureDetector::detectImpl	(	const cv::Mat &	image,
		std::vector< cv::KeyPoint > &	keypoints,
		const cv::Mat &	mask = cv::Mat()
	)		const [protected, virtual]

Definition at line 189 of file feature_adjuster.cpp.

bool VideoDynamicAdaptedFeatureDetector::empty

(

)

const [virtual]

Definition at line 176 of file feature_adjuster.cpp.

VideoDynamicAdaptedFeatureDetector& VideoDynamicAdaptedFeatureDetector::operator=

(

const VideoDynamicAdaptedFeatureDetector &

)

[private]

---

Member Data Documentation

cv::Ptr<cv::AdjusterAdapter> VideoDynamicAdaptedFeatureDetector::adjuster_ [mutable, private]

Definition at line 77 of file feature_adjuster.h.

int VideoDynamicAdaptedFeatureDetector::escape_iters_ [private]

Definition at line 75 of file feature_adjuster.h.

int VideoDynamicAdaptedFeatureDetector::max_features_ [private]

Definition at line 76 of file feature_adjuster.h.

int VideoDynamicAdaptedFeatureDetector::min_features_ [private]

Definition at line 76 of file feature_adjuster.h.

---

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

• feature_adjuster.h
• feature_adjuster.cpp