Robust Local Optical Flow Libary Documentation
V 1.2
|
RLOF Interface. More...
#include "include\OpticalFlow\IFeatureTracker.h"
Public Member Functions | |
virtual void | run (const uchar *imgPtr1, unsigned int step1, const uchar *imgPtr2, unsigned int step2, unsigned int nRows, unsigned int nCols, std::vector< CRPoint > &prevPoints, std::vector< CRPoint > &currPoints, const FeatureTrackerContext *context) |
virtual void | run (const cv::Mat &prevImage, const cv::Mat &currImage, std::vector< cv::Point2f > &prevPoints, std::vector< cv::Point2f > &currPoints, const FeatureTrackerContext *context)=0 |
virtual FeatureTrackerContext * | createDefaultParameter ()=0 |
RLOF Interface.
This class provides a general interface for feature tracking purposes. The RLOF method is a inherent class that is instantiated by the RLOFFactory. To parametrize the RLOF tracking method the FeatureTrackerContext is used. Default parameters could be taken from the createDefaultParameter function.
Definition at line 150 of file IFeatureTracker.h.
|
pure virtual |
Create a default context containing the default parameter for the related tracking methods.
|
inlinevirtual |
Track a set of features points (prevPoints) that are specified by a 2D point list and store the tracked positions at the (currPoints) point list by using the image informations providen by the prevImage and currImage. If these images has to be gray scales images. Parameters are set by the context argument. This interface is used to ensure compatibility down to OpenCV 2.0 und upon new OpenCV versions. To ensure such compatibility OpenCV defined classes are replaced by wrapper classes, see CRPoint.(If use used OpenCV 2.4.6 the second interface function could be used)
imgPtr1 | Data pointer to an rgb or gray scaled image at time t. |
step1 | Specifies the number of elements used to allocate a row of image 1. |
imgPtr2 | Data pointer to an rgb or gray scaled image at time t+1. |
step2 | Specifies the number of elements used to allocate a row of image 2. |
nRows | Number of rows of the input images. |
nCols | Number of columns of the input images. |
prevPoints | Features as point list at time t. |
currPoints | Tracked features as point list to time t+1. |
context | Context including parameters, if NULL default parameters will be used. |
Definition at line 170 of file IFeatureTracker.h.
Referenced by demo(), and denseMotionField().
|
pure virtual |
ATTENTION - should be only just with OpenCV 2.4.6. Track a set of features points (prevPoints) that are specified by a 2D point list and store the tracked positions at the (currPoints) point list by using the image informations provided by the prevImage and currImage. If these images are RGB images, they will be transformed to grav scaled images internally. Parameters are set by the context argument.
prevImage | Image at time t (CV_8UC3 or CV_8UC1 depending from method). |
currImage | Image at time t+1 (CV_8UC3 or CV_8UC1 depending from method). |
prevPoints | Features as point list at time t. |
currPoints | Tracked features as point list to time t+1. |
context | Context including parameters, if NULL default parameters will be used. |