tracking.hpp

00001 /*M///////////////////////////////////////////////////////////////////////////////////////
00002 //
00003 //  IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
00004 //
00005 //  By downloading, copying, installing or using the software you agree to this license.
00006 //  If you do not agree to this license, do not download, install,
00007 //  copy or use the software.
00008 //
00009 //
00010 //                          License Agreement
00011 //                For Open Source Computer Vision Library
00012 //
00013 // Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
00014 // Copyright (C) 2009, Willow Garage Inc., all rights reserved.
00015 // Copyright (C) 2013, OpenCV Foundation, all rights reserved.
00016 // Third party copyrights are property of their respective owners.
00017 //
00018 // Redistribution and use in source and binary forms, with or without modification,
00019 // are permitted provided that the following conditions are met:
00020 //
00021 //   * Redistribution's of source code must retain the above copyright notice,
00022 //     this list of conditions and the following disclaimer.
00023 //
00024 //   * Redistribution's in binary form must reproduce the above copyright notice,
00025 //     this list of conditions and the following disclaimer in the documentation
00026 //     and/or other materials provided with the distribution.
00027 //
00028 //   * The name of the copyright holders may not be used to endorse or promote products
00029 //     derived from this software without specific prior written permission.
00030 //
00031 // This software is provided by the copyright holders and contributors "as is" and
00032 // any express or implied warranties, including, but not limited to, the implied
00033 // warranties of merchantability and fitness for a particular purpose are disclaimed.
00034 // In no event shall the Intel Corporation or contributors be liable for any direct,
00035 // indirect, incidental, special, exemplary, or consequential damages
00036 // (including, but not limited to, procurement of substitute goods or services;
00037 // loss of use, data, or profits; or business interruption) however caused
00038 // and on any theory of liability, whether in contract, strict liability,
00039 // or tort (including negligence or otherwise) arising in any way out of
00040 // the use of this software, even if advised of the possibility of such damage.
00041 //
00042 //M*/
00043 
00044 #ifndef __OPENCV_TRACKING_HPP__
00045 #define __OPENCV_TRACKING_HPP__
00046 
00047 #include "opencv2/core.hpp"
00048 #include "opencv2/imgproc.hpp"
00049 
00050 namespace cv
00051 {
00052 
00053 //! @addtogroup video_track
00054 //! @{
00055 
00056 enum { OPTFLOW_USE_INITIAL_FLOW     = 4,
00057        OPTFLOW_LK_GET_MIN_EIGENVALS = 8,
00058        OPTFLOW_FARNEBACK_GAUSSIAN   = 256
00059      };
00060 
00061 /** @brief Finds an object center, size, and orientation.
00062 
00063 @param probImage Back projection of the object histogram. See calcBackProject.
00064 @param window Initial search window.
00065 @param criteria Stop criteria for the underlying meanShift.
00066 returns
00067 (in old interfaces) Number of iterations CAMSHIFT took to converge
00068 The function implements the CAMSHIFT object tracking algorithm @cite Bradski98 . First, it finds an
00069 object center using meanShift and then adjusts the window size and finds the optimal rotation. The
00070 function returns the rotated rectangle structure that includes the object position, size, and
00071 orientation. The next position of the search window can be obtained with RotatedRect::boundingRect()
00072 
00073 See the OpenCV sample camshiftdemo.c that tracks colored objects.
00074 
00075 @note
00076 -   (Python) A sample explaining the camshift tracking algorithm can be found at
00077     opencv_source_code/samples/python/camshift.py
00078  */
00079 CV_EXPORTS_W RotatedRect CamShift( InputArray probImage, CV_IN_OUT Rect& window,
00080                                    TermCriteria criteria );
00081 
00082 /** @brief Finds an object on a back projection image.
00083 
00084 @param probImage Back projection of the object histogram. See calcBackProject for details.
00085 @param window Initial search window.
00086 @param criteria Stop criteria for the iterative search algorithm.
00087 returns
00088 :   Number of iterations CAMSHIFT took to converge.
00089 The function implements the iterative object search algorithm. It takes the input back projection of
00090 an object and the initial position. The mass center in window of the back projection image is
00091 computed and the search window center shifts to the mass center. The procedure is repeated until the
00092 specified number of iterations criteria.maxCount is done or until the window center shifts by less
00093 than criteria.epsilon. The algorithm is used inside CamShift and, unlike CamShift , the search
00094 window size or orientation do not change during the search. You can simply pass the output of
00095 calcBackProject to this function. But better results can be obtained if you pre-filter the back
00096 projection and remove the noise. For example, you can do this by retrieving connected components
00097 with findContours , throwing away contours with small area ( contourArea ), and rendering the
00098 remaining contours with drawContours.
00099 
00100 @note
00101 -   A mean-shift tracking sample can be found at opencv_source_code/samples/cpp/camshiftdemo.cpp
00102  */
00103 CV_EXPORTS_W int meanShift( InputArray probImage, CV_IN_OUT Rect& window, TermCriteria criteria );
00104 
00105 /** @brief Constructs the image pyramid which can be passed to calcOpticalFlowPyrLK.
00106 
00107 @param img 8-bit input image.
00108 @param pyramid output pyramid.
00109 @param winSize window size of optical flow algorithm. Must be not less than winSize argument of
00110 calcOpticalFlowPyrLK. It is needed to calculate required padding for pyramid levels.
00111 @param maxLevel 0-based maximal pyramid level number.
00112 @param withDerivatives set to precompute gradients for the every pyramid level. If pyramid is
00113 constructed without the gradients then calcOpticalFlowPyrLK will calculate them internally.
00114 @param pyrBorder the border mode for pyramid layers.
00115 @param derivBorder the border mode for gradients.
00116 @param tryReuseInputImage put ROI of input image into the pyramid if possible. You can pass false
00117 to force data copying.
00118 @return number of levels in constructed pyramid. Can be less than maxLevel.
00119  */
00120 CV_EXPORTS_W int buildOpticalFlowPyramid( InputArray img, OutputArrayOfArrays pyramid,
00121                                           Size winSize, int maxLevel, bool withDerivatives = true,
00122                                           int pyrBorder = BORDER_REFLECT_101,
00123                                           int derivBorder = BORDER_CONSTANT,
00124                                           bool tryReuseInputImage = true );
00125 
00126 /** @brief Calculates an optical flow for a sparse feature set using the iterative Lucas-Kanade method with
00127 pyramids.
00128 
00129 @param prevImg first 8-bit input image or pyramid constructed by buildOpticalFlowPyramid.
00130 @param nextImg second input image or pyramid of the same size and the same type as prevImg.
00131 @param prevPts vector of 2D points for which the flow needs to be found; point coordinates must be
00132 single-precision floating-point numbers.
00133 @param nextPts output vector of 2D points (with single-precision floating-point coordinates)
00134 containing the calculated new positions of input features in the second image; when
00135 OPTFLOW_USE_INITIAL_FLOW flag is passed, the vector must have the same size as in the input.
00136 @param status output status vector (of unsigned chars); each element of the vector is set to 1 if
00137 the flow for the corresponding features has been found, otherwise, it is set to 0.
00138 @param err output vector of errors; each element of the vector is set to an error for the
00139 corresponding feature, type of the error measure can be set in flags parameter; if the flow wasn't
00140 found then the error is not defined (use the status parameter to find such cases).
00141 @param winSize size of the search window at each pyramid level.
00142 @param maxLevel 0-based maximal pyramid level number; if set to 0, pyramids are not used (single
00143 level), if set to 1, two levels are used, and so on; if pyramids are passed to input then
00144 algorithm will use as many levels as pyramids have but no more than maxLevel.
00145 @param criteria parameter, specifying the termination criteria of the iterative search algorithm
00146 (after the specified maximum number of iterations criteria.maxCount or when the search window
00147 moves by less than criteria.epsilon.
00148 @param flags operation flags:
00149  -   **OPTFLOW_USE_INITIAL_FLOW** uses initial estimations, stored in nextPts; if the flag is
00150      not set, then prevPts is copied to nextPts and is considered the initial estimate.
00151  -   **OPTFLOW_LK_GET_MIN_EIGENVALS** use minimum eigen values as an error measure (see
00152      minEigThreshold description); if the flag is not set, then L1 distance between patches
00153      around the original and a moved point, divided by number of pixels in a window, is used as a
00154      error measure.
00155 @param minEigThreshold the algorithm calculates the minimum eigen value of a 2x2 normal matrix of
00156 optical flow equations (this matrix is called a spatial gradient matrix in @cite Bouguet00), divided
00157 by number of pixels in a window; if this value is less than minEigThreshold, then a corresponding
00158 feature is filtered out and its flow is not processed, so it allows to remove bad points and get a
00159 performance boost.
00160 
00161 The function implements a sparse iterative version of the Lucas-Kanade optical flow in pyramids. See
00162 @cite Bouguet00 . The function is parallelized with the TBB library.
00163 
00164 @note
00165 
00166 -   An example using the Lucas-Kanade optical flow algorithm can be found at
00167     opencv_source_code/samples/cpp/lkdemo.cpp
00168 -   (Python) An example using the Lucas-Kanade optical flow algorithm can be found at
00169     opencv_source_code/samples/python/lk_track.py
00170 -   (Python) An example using the Lucas-Kanade tracker for homography matching can be found at
00171     opencv_source_code/samples/python/lk_homography.py
00172  */
00173 CV_EXPORTS_W void calcOpticalFlowPyrLK( InputArray prevImg, InputArray nextImg,
00174                                         InputArray prevPts, InputOutputArray nextPts,
00175                                         OutputArray status, OutputArray err,
00176                                         Size winSize = Size(21,21), int maxLevel = 3,
00177                                         TermCriteria criteria = TermCriteria(TermCriteria::COUNT+TermCriteria::EPS, 30, 0.01),
00178                                         int flags = 0, double minEigThreshold = 1e-4 );
00179 
00180 /** @brief Computes a dense optical flow using the Gunnar Farneback's algorithm.
00181 
00182 @param prev first 8-bit single-channel input image.
00183 @param next second input image of the same size and the same type as prev.
00184 @param flow computed flow image that has the same size as prev and type CV_32FC2.
00185 @param pyr_scale parameter, specifying the image scale (<1) to build pyramids for each image;
00186 pyr_scale=0.5 means a classical pyramid, where each next layer is twice smaller than the previous
00187 one.
00188 @param levels number of pyramid layers including the initial image; levels=1 means that no extra
00189 layers are created and only the original images are used.
00190 @param winsize averaging window size; larger values increase the algorithm robustness to image
00191 noise and give more chances for fast motion detection, but yield more blurred motion field.
00192 @param iterations number of iterations the algorithm does at each pyramid level.
00193 @param poly_n size of the pixel neighborhood used to find polynomial expansion in each pixel;
00194 larger values mean that the image will be approximated with smoother surfaces, yielding more
00195 robust algorithm and more blurred motion field, typically poly_n =5 or 7.
00196 @param poly_sigma standard deviation of the Gaussian that is used to smooth derivatives used as a
00197 basis for the polynomial expansion; for poly_n=5, you can set poly_sigma=1.1, for poly_n=7, a
00198 good value would be poly_sigma=1.5.
00199 @param flags operation flags that can be a combination of the following:
00200  -   **OPTFLOW_USE_INITIAL_FLOW** uses the input flow as an initial flow approximation.
00201  -   **OPTFLOW_FARNEBACK_GAUSSIAN** uses the Gaussian \f$\texttt{winsize}\times\texttt{winsize}\f$
00202      filter instead of a box filter of the same size for optical flow estimation; usually, this
00203      option gives z more accurate flow than with a box filter, at the cost of lower speed;
00204      normally, winsize for a Gaussian window should be set to a larger value to achieve the same
00205      level of robustness.
00206 
00207 The function finds an optical flow for each prev pixel using the @cite Farneback2003 algorithm so that
00208 
00209 \f[\texttt{prev} (y,x)  \sim \texttt{next} ( y + \texttt{flow} (y,x)[1],  x + \texttt{flow} (y,x)[0])\f]
00210 
00211 @note
00212 
00213 -   An example using the optical flow algorithm described by Gunnar Farneback can be found at
00214     opencv_source_code/samples/cpp/fback.cpp
00215 -   (Python) An example using the optical flow algorithm described by Gunnar Farneback can be
00216     found at opencv_source_code/samples/python/opt_flow.py
00217  */
00218 CV_EXPORTS_W void calcOpticalFlowFarneback( InputArray prev, InputArray next, InputOutputArray flow,
00219                                             double pyr_scale, int levels, int winsize,
00220                                             int iterations, int poly_n, double poly_sigma,
00221                                             int flags );
00222 
00223 /** @brief Computes an optimal affine transformation between two 2D point sets.
00224 
00225 @param src First input 2D point set stored in std::vector or Mat, or an image stored in Mat.
00226 @param dst Second input 2D point set of the same size and the same type as A, or another image.
00227 @param fullAffine If true, the function finds an optimal affine transformation with no additional
00228 restrictions (6 degrees of freedom). Otherwise, the class of transformations to choose from is
00229 limited to combinations of translation, rotation, and uniform scaling (5 degrees of freedom).
00230 
00231 The function finds an optimal affine transform *[A|b]* (a 2 x 3 floating-point matrix) that
00232 approximates best the affine transformation between:
00233 
00234 *   Two point sets
00235 *   Two raster images. In this case, the function first finds some features in the src image and
00236     finds the corresponding features in dst image. After that, the problem is reduced to the first
00237     case.
00238 In case of point sets, the problem is formulated as follows: you need to find a 2x2 matrix *A* and
00239 2x1 vector *b* so that:
00240 
00241 \f[[A^*|b^*] = arg  \min _{[A|b]}  \sum _i  \| \texttt{dst}[i] - A { \texttt{src}[i]}^T - b  \| ^2\f]
00242 where src[i] and dst[i] are the i-th points in src and dst, respectively
00243 \f$[A|b]\f$ can be either arbitrary (when fullAffine=true ) or have a form of
00244 \f[\begin{bmatrix} a_{11} & a_{12} & b_1  \\ -a_{12} & a_{11} & b_2  \end{bmatrix}\f]
00245 when fullAffine=false.
00246 
00247 @sa
00248 getAffineTransform, getPerspectiveTransform, findHomography
00249  */
00250 CV_EXPORTS_W Mat estimateRigidTransform( InputArray src, InputArray dst, bool fullAffine );
00251 
00252 
00253 enum
00254 {
00255     MOTION_TRANSLATION = 0,
00256     MOTION_EUCLIDEAN   = 1,
00257     MOTION_AFFINE      = 2,
00258     MOTION_HOMOGRAPHY  = 3
00259 };
00260 
00261 /** @brief Finds the geometric transform (warp) between two images in terms of the ECC criterion @cite EP08 .
00262 
00263 @param templateImage single-channel template image; CV_8U or CV_32F array.
00264 @param inputImage single-channel input image which should be warped with the final warpMatrix in
00265 order to provide an image similar to templateImage, same type as temlateImage.
00266 @param warpMatrix floating-point \f$2\times 3\f$ or \f$3\times 3\f$ mapping matrix (warp).
00267 @param motionType parameter, specifying the type of motion:
00268  -   **MOTION_TRANSLATION** sets a translational motion model; warpMatrix is \f$2\times 3\f$ with
00269      the first \f$2\times 2\f$ part being the unity matrix and the rest two parameters being
00270      estimated.
00271  -   **MOTION_EUCLIDEAN** sets a Euclidean (rigid) transformation as motion model; three
00272      parameters are estimated; warpMatrix is \f$2\times 3\f$.
00273  -   **MOTION_AFFINE** sets an affine motion model (DEFAULT); six parameters are estimated;
00274      warpMatrix is \f$2\times 3\f$.
00275  -   **MOTION_HOMOGRAPHY** sets a homography as a motion model; eight parameters are
00276      estimated;\`warpMatrix\` is \f$3\times 3\f$.
00277 @param criteria parameter, specifying the termination criteria of the ECC algorithm;
00278 criteria.epsilon defines the threshold of the increment in the correlation coefficient between two
00279 iterations (a negative criteria.epsilon makes criteria.maxcount the only termination criterion).
00280 Default values are shown in the declaration above.
00281 @param inputMask An optional mask to indicate valid values of inputImage.
00282 
00283 The function estimates the optimum transformation (warpMatrix) with respect to ECC criterion
00284 (@cite EP08), that is
00285 
00286 \f[\texttt{warpMatrix} = \texttt{warpMatrix} = \arg\max_{W} \texttt{ECC}(\texttt{templateImage}(x,y),\texttt{inputImage}(x',y'))\f]
00287 
00288 where
00289 
00290 \f[\begin{bmatrix} x' \\ y' \end{bmatrix} = W \cdot \begin{bmatrix} x \\ y \\ 1 \end{bmatrix}\f]
00291 
00292 (the equation holds with homogeneous coordinates for homography). It returns the final enhanced
00293 correlation coefficient, that is the correlation coefficient between the template image and the
00294 final warped input image. When a \f$3\times 3\f$ matrix is given with motionType =0, 1 or 2, the third
00295 row is ignored.
00296 
00297 Unlike findHomography and estimateRigidTransform, the function findTransformECC implements an
00298 area-based alignment that builds on intensity similarities. In essence, the function updates the
00299 initial transformation that roughly aligns the images. If this information is missing, the identity
00300 warp (unity matrix) should be given as input. Note that if images undergo strong
00301 displacements/rotations, an initial transformation that roughly aligns the images is necessary
00302 (e.g., a simple euclidean/similarity transform that allows for the images showing the same image
00303 content approximately). Use inverse warping in the second image to take an image close to the first
00304 one, i.e. use the flag WARP_INVERSE_MAP with warpAffine or warpPerspective. See also the OpenCV
00305 sample image_alignment.cpp that demonstrates the use of the function. Note that the function throws
00306 an exception if algorithm does not converges.
00307 
00308 @sa
00309 estimateRigidTransform, findHomography
00310  */
00311 CV_EXPORTS_W double findTransformECC( InputArray templateImage, InputArray inputImage,
00312                                       InputOutputArray warpMatrix, int motionType = MOTION_AFFINE,
00313                                       TermCriteria criteria = TermCriteria(TermCriteria::COUNT+TermCriteria::EPS, 50, 0.001),
00314                                       InputArray inputMask = noArray());
00315 
00316 /** @brief Kalman filter class.
00317 
00318 The class implements a standard Kalman filter <http://en.wikipedia.org/wiki/Kalman_filter>,
00319 @cite Welch95 . However, you can modify transitionMatrix, controlMatrix, and measurementMatrix to get
00320 an extended Kalman filter functionality. See the OpenCV sample kalman.cpp.
00321 
00322 @note
00323 
00324 -   An example using the standard Kalman filter can be found at
00325     opencv_source_code/samples/cpp/kalman.cpp
00326  */
00327 class CV_EXPORTS_W KalmanFilter
00328 {
00329 public:
00330     /** @brief The constructors.
00331 
00332     @note In C API when CvKalman\* kalmanFilter structure is not needed anymore, it should be released
00333     with cvReleaseKalman(&kalmanFilter)
00334      */
00335     CV_WRAP KalmanFilter();
00336     /** @overload
00337     @param dynamParams Dimensionality of the state.
00338     @param measureParams Dimensionality of the measurement.
00339     @param controlParams Dimensionality of the control vector.
00340     @param type Type of the created matrices that should be CV_32F or CV_64F.
00341     */
00342     CV_WRAP KalmanFilter( int dynamParams, int measureParams, int controlParams = 0, int type = CV_32F );
00343 
00344     /** @brief Re-initializes Kalman filter. The previous content is destroyed.
00345 
00346     @param dynamParams Dimensionality of the state.
00347     @param measureParams Dimensionality of the measurement.
00348     @param controlParams Dimensionality of the control vector.
00349     @param type Type of the created matrices that should be CV_32F or CV_64F.
00350      */
00351     void init( int dynamParams, int measureParams, int controlParams = 0, int type = CV_32F );
00352 
00353     /** @brief Computes a predicted state.
00354 
00355     @param control The optional input control
00356      */
00357     CV_WRAP const Mat& predict( const Mat& control = Mat() );
00358 
00359     /** @brief Updates the predicted state from the measurement.
00360 
00361     @param measurement The measured system parameters
00362      */
00363     CV_WRAP const Mat& correct( const Mat& measurement );
00364 
00365     CV_PROP_RW Mat statePre;           //!< predicted state (x'(k)): x(k)=A*x(k-1)+B*u(k)
00366     CV_PROP_RW Mat statePost;          //!< corrected state (x(k)): x(k)=x'(k)+K(k)*(z(k)-H*x'(k))
00367     CV_PROP_RW Mat transitionMatrix;   //!< state transition matrix (A)
00368     CV_PROP_RW Mat controlMatrix;      //!< control matrix (B) (not used if there is no control)
00369     CV_PROP_RW Mat measurementMatrix;  //!< measurement matrix (H)
00370     CV_PROP_RW Mat processNoiseCov;    //!< process noise covariance matrix (Q)
00371     CV_PROP_RW Mat measurementNoiseCov;//!< measurement noise covariance matrix (R)
00372     CV_PROP_RW Mat errorCovPre;        //!< priori error estimate covariance matrix (P'(k)): P'(k)=A*P(k-1)*At + Q)*/
00373     CV_PROP_RW Mat gain;               //!< Kalman gain matrix (K(k)): K(k)=P'(k)*Ht*inv(H*P'(k)*Ht+R)
00374     CV_PROP_RW Mat errorCovPost;       //!< posteriori error estimate covariance matrix (P(k)): P(k)=(I-K(k)*H)*P'(k)
00375 
00376     // temporary matrices
00377     Mat temp1;
00378     Mat temp2;
00379     Mat temp3;
00380     Mat temp4;
00381     Mat temp5;
00382 };
00383 
00384 
00385 class CV_EXPORTS_W DenseOpticalFlow : public Algorithm
00386 {
00387 public:
00388     /** @brief Calculates an optical flow.
00389 
00390     @param I0 first 8-bit single-channel input image.
00391     @param I1 second input image of the same size and the same type as prev.
00392     @param flow computed flow image that has the same size as prev and type CV_32FC2.
00393      */
00394     CV_WRAP virtual void calc( InputArray I0, InputArray I1, InputOutputArray flow ) = 0;
00395     /** @brief Releases all inner buffers.
00396     */
00397     CV_WRAP virtual void collectGarbage() = 0;
00398 };
00399 
00400 /** @brief "Dual TV L1" Optical Flow Algorithm.
00401 
00402 The class implements the "Dual TV L1" optical flow algorithm described in @cite Zach2007 and
00403 @cite Javier2012 .
00404 Here are important members of the class that control the algorithm, which you can set after
00405 constructing the class instance:
00406 
00407 -   member double tau
00408     Time step of the numerical scheme.
00409 
00410 -   member double lambda
00411     Weight parameter for the data term, attachment parameter. This is the most relevant
00412     parameter, which determines the smoothness of the output. The smaller this parameter is,
00413     the smoother the solutions we obtain. It depends on the range of motions of the images, so
00414     its value should be adapted to each image sequence.
00415 
00416 -   member double theta
00417     Weight parameter for (u - v)\^2, tightness parameter. It serves as a link between the
00418     attachment and the regularization terms. In theory, it should have a small value in order
00419     to maintain both parts in correspondence. The method is stable for a large range of values
00420     of this parameter.
00421 
00422 -   member int nscales
00423     Number of scales used to create the pyramid of images.
00424 
00425 -   member int warps
00426     Number of warpings per scale. Represents the number of times that I1(x+u0) and grad(
00427     I1(x+u0) ) are computed per scale. This is a parameter that assures the stability of the
00428     method. It also affects the running time, so it is a compromise between speed and
00429     accuracy.
00430 
00431 -   member double epsilon
00432     Stopping criterion threshold used in the numerical scheme, which is a trade-off between
00433     precision and running time. A small value will yield more accurate solutions at the
00434     expense of a slower convergence.
00435 
00436 -   member int iterations
00437     Stopping criterion iterations number used in the numerical scheme.
00438 
00439 C. Zach, T. Pock and H. Bischof, "A Duality Based Approach for Realtime TV-L1 Optical Flow".
00440 Javier Sanchez, Enric Meinhardt-Llopis and Gabriele Facciolo. "TV-L1 Optical Flow Estimation".
00441 */
00442 class CV_EXPORTS_W DualTVL1OpticalFlow : public DenseOpticalFlow
00443 {
00444 public:
00445     //! @brief Time step of the numerical scheme
00446     /** @see setTau */
00447     virtual double getTau() const = 0;
00448     /** @copybrief getTau @see getTau */
00449     virtual void setTau(double val) = 0;
00450     //! @brief Weight parameter for the data term, attachment parameter
00451     /** @see setLambda */
00452     virtual double getLambda() const = 0;
00453     /** @copybrief getLambda @see getLambda */
00454     virtual void setLambda(double val) = 0;
00455     //! @brief Weight parameter for (u - v)^2, tightness parameter
00456     /** @see setTheta */
00457     virtual double getTheta() const = 0;
00458     /** @copybrief getTheta @see getTheta */
00459     virtual void setTheta(double val) = 0;
00460     //! @brief coefficient for additional illumination variation term
00461     /** @see setGamma */
00462     virtual double getGamma() const = 0;
00463     /** @copybrief getGamma @see getGamma */
00464     virtual void setGamma(double val) = 0;
00465     //! @brief Number of scales used to create the pyramid of images
00466     /** @see setScalesNumber */
00467     virtual int getScalesNumber() const = 0;
00468     /** @copybrief getScalesNumber @see getScalesNumber */
00469     virtual void setScalesNumber(int val) = 0;
00470     //! @brief Number of warpings per scale
00471     /** @see setWarpingsNumber */
00472     virtual int getWarpingsNumber() const = 0;
00473     /** @copybrief getWarpingsNumber @see getWarpingsNumber */
00474     virtual void setWarpingsNumber(int val) = 0;
00475     //! @brief Stopping criterion threshold used in the numerical scheme, which is a trade-off between precision and running time
00476     /** @see setEpsilon */
00477     virtual double getEpsilon() const = 0;
00478     /** @copybrief getEpsilon @see getEpsilon */
00479     virtual void setEpsilon(double val) = 0;
00480     //! @brief Inner iterations (between outlier filtering) used in the numerical scheme
00481     /** @see setInnerIterations */
00482     virtual int getInnerIterations() const = 0;
00483     /** @copybrief getInnerIterations @see getInnerIterations */
00484     virtual void setInnerIterations(int val) = 0;
00485     //! @brief Outer iterations (number of inner loops) used in the numerical scheme
00486     /** @see setOuterIterations */
00487     virtual int getOuterIterations() const = 0;
00488     /** @copybrief getOuterIterations @see getOuterIterations */
00489     virtual void setOuterIterations(int val) = 0;
00490     //! @brief Use initial flow
00491     /** @see setUseInitialFlow */
00492     virtual bool getUseInitialFlow() const = 0;
00493     /** @copybrief getUseInitialFlow @see getUseInitialFlow */
00494     virtual void setUseInitialFlow(bool val) = 0;
00495     //! @brief Step between scales (<1)
00496     /** @see setScaleStep */
00497     virtual double getScaleStep() const = 0;
00498     /** @copybrief getScaleStep @see getScaleStep */
00499     virtual void setScaleStep(double val) = 0;
00500     //! @brief Median filter kernel size (1 = no filter) (3 or 5)
00501     /** @see setMedianFiltering */
00502     virtual int getMedianFiltering() const = 0;
00503     /** @copybrief getMedianFiltering @see getMedianFiltering */
00504     virtual void setMedianFiltering(int val) = 0;
00505 };
00506 
00507 /** @brief Creates instance of cv::DenseOpticalFlow
00508 */
00509 CV_EXPORTS_W Ptr<DualTVL1OpticalFlow> createOptFlow_DualTVL1();
00510 
00511 //! @} video_track
00512 
00513 } // cv
00514 
00515 #endif
00516