3 * Copyright (C) 2013 Miguel Casas-Sanchez <miguelecasassanchez@gmail.com>
4 * Except: Parts of code inside the preprocessor define CODE_FROM_OREILLY_BOOK,
5 * which are downloaded from O'Reilly website
6 * [http://examples.oreilly.com/9780596516130/]
7 * and adapted. Its license reads:
9 * Right to use this code in any way you want without warrenty, support or
10 * any guarentee of it working. "
13 * Permission is hereby granted, free of charge, to any person obtaining a
14 * copy of this software and associated documentation files (the "Software"),
15 * to deal in the Software without restriction, including without limitation
16 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
17 * and/or sell copies of the Software, and to permit persons to whom the
18 * Software is furnished to do so, subject to the following conditions:
20 * The above copyright notice and this permission notice shall be included in
21 * all copies or substantial portions of the Software.
23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
24 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
25 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
26 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
27 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
28 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
29 * DEALINGS IN THE SOFTWARE.
31 * Alternatively, the contents of this file may be used under the
32 * GNU Lesser General Public License Version 2.1 (the "LGPL"), in
33 * which case the following provisions apply instead of the ones
36 * This library is free software; you can redistribute it and/or
37 * modify it under the terms of the GNU Library General Public
38 * License as published by the Free Software Foundation; either
39 * version 2 of the License, or (at your option) any later version.
41 * This library is distributed in the hope that it will be useful,
42 * but WITHOUT ANY WARRANTY; without even the implied warranty of
43 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
44 * Library General Public License for more details.
46 * You should have received a copy of the GNU Library General Public
47 * License along with this library; if not, write to the
48 * Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
49 * Boston, MA 02110-1301, USA.
51 #define CODE_FROM_OREILLY_BOOK
54 * SECTION:element-segmentation
56 * This element creates and updates a fg/bg model using one of several approaches.
57 * The one called "codebook" refers to the codebook approach following the opencv
58 * O'Reilly book [1] implementation of the algorithm described in K. Kim,
59 * T. H. Chalidabhongse, D. Harwood and L. Davis [2]. BackgroundSubtractorMOG [3],
60 * or MOG for shorts, refers to a Gaussian Mixture-based Background/Foreground
61 * Segmentation Algorithm. OpenCV MOG implements the algorithm described in [4].
62 * BackgroundSubtractorMOG2 [5], refers to another Gaussian Mixture-based
63 * Background/Foreground segmentation algorithm. OpenCV MOG2 implements the
64 * algorithm described in [6] and [7].
66 * [1] Learning OpenCV: Computer Vision with the OpenCV Library by Gary Bradski
67 * and Adrian Kaehler, Published by O'Reilly Media, October 3, 2008
68 * [2] "Real-time Foreground-Background Segmentation using Codebook Model",
69 * Real-time Imaging, Volume 11, Issue 3, Pages 167-256, June 2005.
70 * [3] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog
71 * [4] P. KadewTraKuPong and R. Bowden, "An improved adaptive background
72 * mixture model for real-time tracking with shadow detection", Proc. 2nd
73 * European Workshop on Advanced Video-Based Surveillance Systems, 2001
74 * [5] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog2
75 * [6] Z.Zivkovic, "Improved adaptive Gausian mixture model for background
76 * subtraction", International Conference Pattern Recognition, UK, August, 2004.
77 * [7] Z.Zivkovic, F. van der Heijden, "Efficient Adaptive Density Estimation
78 * per Image Pixel for the Task of Background Subtraction", Pattern Recognition
79 * Letters, vol. 27, no. 7, pages 773-780, 2006.
82 * <title>Example launch line</title>
84 * gst-launch-1.0 v4l2src device=/dev/video0 ! videoconvert ! video/x-raw,width=320,height=240 ! videoconvert ! segmentation test-mode=true method=2 ! videoconvert ! ximagesink
95 #include "gstsegmentation.h"
96 #include <opencv2/video/background_segm.hpp>
98 GST_DEBUG_CATEGORY_STATIC (gst_segmentation_debug);
99 #define GST_CAT_DEFAULT gst_segmentation_debug
101 /* Filter signals and args */
120 } GstSegmentationMethod;
122 #define DEFAULT_TEST_MODE FALSE
123 #define DEFAULT_METHOD METHOD_MOG2
124 #define DEFAULT_LEARNING_RATE 0.01
126 #define GST_TYPE_SEGMENTATION_METHOD (gst_segmentation_method_get_type ())
128 gst_segmentation_method_get_type (void)
130 static GType etype = 0;
132 static const GEnumValue values[] = {
133 {METHOD_BOOK, "Codebook-based segmentation (Bradski2008)", "codebook"},
134 {METHOD_MOG, "Mixture-of-Gaussians segmentation (Bowden2001)", "mog"},
135 {METHOD_MOG2, "Mixture-of-Gaussians segmentation (Zivkovic2004)", "mog2"},
138 etype = g_enum_register_static ("GstSegmentationMethod", values);
143 G_DEFINE_TYPE (GstSegmentation, gst_segmentation, GST_TYPE_VIDEO_FILTER);
144 static GstStaticPadTemplate sink_factory = GST_STATIC_PAD_TEMPLATE ("sink",
147 GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("RGBA")));
149 static GstStaticPadTemplate src_factory = GST_STATIC_PAD_TEMPLATE ("src",
152 GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("RGBA")));
155 static void gst_segmentation_set_property (GObject * object, guint prop_id,
156 const GValue * value, GParamSpec * pspec);
157 static void gst_segmentation_get_property (GObject * object, guint prop_id,
158 GValue * value, GParamSpec * pspec);
160 static GstFlowReturn gst_segmentation_transform_ip (GstVideoFilter * btrans,
161 GstVideoFrame * frame);
163 static gboolean gst_segmentation_stop (GstBaseTransform * basesrc);
164 static gboolean gst_segmentation_set_info (GstVideoFilter * filter,
165 GstCaps * incaps, GstVideoInfo * in_info,
166 GstCaps * outcaps, GstVideoInfo * out_info);
167 static void gst_segmentation_release_all_pointers (GstSegmentation * filter);
169 /* Codebook algorithm + connected components functions*/
170 static int update_codebook (unsigned char *p, codeBook * c,
171 unsigned *cbBounds, int numChannels);
172 static int clear_stale_entries (codeBook * c);
173 static unsigned char background_diff (unsigned char *p, codeBook * c,
174 int numChannels, int *minMod, int *maxMod);
175 static void find_connected_components (IplImage * mask, int poly1_hull0,
176 float perimScale, CvMemStorage * mem_storage, CvSeq * contours);
178 /* MOG (Mixture-of-Gaussians functions */
179 static int initialise_mog (GstSegmentation * filter);
180 static int run_mog_iteration (GstSegmentation * filter);
181 static int run_mog2_iteration (GstSegmentation * filter);
182 static int finalise_mog (GstSegmentation * filter);
184 /* initialize the segmentation's class */
186 gst_segmentation_class_init (GstSegmentationClass * klass)
188 GObjectClass *gobject_class;
189 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
190 GstBaseTransformClass *basesrc_class = GST_BASE_TRANSFORM_CLASS (klass);
191 GstVideoFilterClass *video_class = (GstVideoFilterClass *) klass;
193 gobject_class = (GObjectClass *) klass;
195 gobject_class->set_property = gst_segmentation_set_property;
196 gobject_class->get_property = gst_segmentation_get_property;
198 basesrc_class->stop = gst_segmentation_stop;
200 video_class->transform_frame_ip = gst_segmentation_transform_ip;
201 video_class->set_info = gst_segmentation_set_info;
203 g_object_class_install_property (gobject_class, PROP_METHOD,
204 g_param_spec_enum ("method",
205 "Segmentation method to use",
206 "Segmentation method to use",
207 GST_TYPE_SEGMENTATION_METHOD, DEFAULT_METHOD,
208 (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS)));
210 g_object_class_install_property (gobject_class, PROP_TEST_MODE,
211 g_param_spec_boolean ("test-mode", "test-mode",
212 "If true, the output RGB is overwritten with the calculated foreground (white color)",
213 DEFAULT_TEST_MODE, (GParamFlags)
214 (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS)));
216 g_object_class_install_property (gobject_class, PROP_LEARNING_RATE,
217 g_param_spec_float ("learning-rate", "learning-rate",
218 "Speed with which a motionless foreground pixel would become background (inverse of number of frames)",
219 0, 1, DEFAULT_LEARNING_RATE, (GParamFlags) (G_PARAM_READWRITE)));
221 gst_element_class_set_static_metadata (element_class,
222 "Foreground/background video sequence segmentation",
223 "Filter/Effect/Video",
224 "Create a Foregound/Background mask applying a particular algorithm",
225 "Miguel Casas-Sanchez <miguelecasassanchez@gmail.com>");
227 gst_element_class_add_pad_template (element_class,
228 gst_static_pad_template_get (&src_factory));
229 gst_element_class_add_pad_template (element_class,
230 gst_static_pad_template_get (&sink_factory));
234 /* initialize the new element
235 * instantiate pads and add them to element
236 * set pad calback functions
237 * initialize instance structure
240 gst_segmentation_init (GstSegmentation * filter)
242 filter->method = DEFAULT_METHOD;
243 filter->test_mode = DEFAULT_TEST_MODE;
244 filter->framecount = 0;
245 filter->learning_rate = DEFAULT_LEARNING_RATE;
246 gst_base_transform_set_in_place (GST_BASE_TRANSFORM (filter), TRUE);
251 gst_segmentation_set_property (GObject * object, guint prop_id,
252 const GValue * value, GParamSpec * pspec)
254 GstSegmentation *filter = GST_SEGMENTATION (object);
258 filter->method = g_value_get_enum (value);
261 filter->test_mode = g_value_get_boolean (value);
263 case PROP_LEARNING_RATE:
264 filter->learning_rate = g_value_get_float (value);
267 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
273 gst_segmentation_get_property (GObject * object, guint prop_id,
274 GValue * value, GParamSpec * pspec)
276 GstSegmentation *filter = GST_SEGMENTATION (object);
280 g_value_set_enum (value, filter->method);
283 g_value_set_boolean (value, filter->test_mode);
285 case PROP_LEARNING_RATE:
286 g_value_set_float (value, filter->learning_rate);
289 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
294 /* GstElement vmethod implementations */
295 /* this function handles the link with other elements */
297 gst_segmentation_set_info (GstVideoFilter * filter,
298 GstCaps * incaps, GstVideoInfo * in_info,
299 GstCaps * outcaps, GstVideoInfo * out_info)
301 GstSegmentation *segmentation = GST_SEGMENTATION (filter);
304 size = cvSize (in_info->width, in_info->height);
305 segmentation->width = in_info->width;
306 segmentation->height = in_info->height;
307 /* If cvRGB is already allocated, it means there's a cap modification, */
308 /* so release first all the images. */
309 if (NULL != segmentation->cvRGBA)
310 gst_segmentation_release_all_pointers (segmentation);
312 segmentation->cvRGBA = cvCreateImageHeader (size, IPL_DEPTH_8U, 4);
314 segmentation->cvRGB = cvCreateImage (size, IPL_DEPTH_8U, 3);
315 segmentation->cvYUV = cvCreateImage (size, IPL_DEPTH_8U, 3);
317 segmentation->cvFG = cvCreateImage (size, IPL_DEPTH_8U, 1);
318 cvZero (segmentation->cvFG);
320 segmentation->ch1 = cvCreateImage (size, IPL_DEPTH_8U, 1);
321 segmentation->ch2 = cvCreateImage (size, IPL_DEPTH_8U, 1);
322 segmentation->ch3 = cvCreateImage (size, IPL_DEPTH_8U, 1);
324 /* Codebook method */
325 segmentation->TcodeBook = (codeBook *)
326 g_malloc (sizeof (codeBook) *
327 (segmentation->width * segmentation->height + 1));
328 for (int j = 0; j < segmentation->width * segmentation->height; j++) {
329 segmentation->TcodeBook[j].numEntries = 0;
330 segmentation->TcodeBook[j].t = 0;
332 segmentation->learning_interval = (int) (1.0 / segmentation->learning_rate);
334 /* Mixture-of-Gaussians (mog) methods */
335 initialise_mog (segmentation);
342 gst_segmentation_stop (GstBaseTransform * basesrc)
344 GstSegmentation *filter = GST_SEGMENTATION (basesrc);
346 if (filter->cvRGBA != NULL)
347 gst_segmentation_release_all_pointers (filter);
353 gst_segmentation_release_all_pointers (GstSegmentation * filter)
355 cvReleaseImage (&filter->cvRGBA);
356 cvReleaseImage (&filter->cvRGB);
357 cvReleaseImage (&filter->cvYUV);
358 cvReleaseImage (&filter->cvFG);
359 cvReleaseImage (&filter->ch1);
360 cvReleaseImage (&filter->ch2);
361 cvReleaseImage (&filter->ch3);
363 g_free (filter->TcodeBook);
364 finalise_mog (filter);
368 gst_segmentation_transform_ip (GstVideoFilter * btrans, GstVideoFrame * frame)
370 GstSegmentation *filter = GST_SEGMENTATION (btrans);
373 /* get image data from the input, which is RGBA */
374 filter->cvRGBA->imageData = (char *) GST_VIDEO_FRAME_COMP_DATA (frame, 0);
375 filter->cvRGBA->widthStep = GST_VIDEO_FRAME_COMP_STRIDE (frame, 0);
376 filter->framecount++;
378 /* Image preprocessing: color space conversion etc */
379 cvCvtColor (filter->cvRGBA, filter->cvRGB, CV_RGBA2RGB);
380 cvCvtColor (filter->cvRGB, filter->cvYUV, CV_RGB2YCrCb);
382 /* Create and update a fg/bg model using a codebook approach following the
383 * opencv O'Reilly book [1] implementation of the algo described in [2].
385 * [1] Learning OpenCV: Computer Vision with the OpenCV Library by Gary
386 * Bradski and Adrian Kaehler, Published by O'Reilly Media, October 3, 2008
387 * [2] "Real-time Foreground-Background Segmentation using Codebook Model",
388 * Real-time Imaging, Volume 11, Issue 3, Pages 167-256, June 2005. */
389 if (METHOD_BOOK == filter->method) {
390 unsigned cbBounds[3] = { 10, 5, 5 };
391 int minMod[3] = { 20, 20, 20 }, maxMod[3] = {
394 if (filter->framecount < 30) {
395 /* Learning background phase: update_codebook on every frame */
396 for (j = 0; j < filter->width * filter->height; j++) {
397 update_codebook ((unsigned char *) filter->cvYUV->imageData + j * 3,
398 (codeBook *) & (filter->TcodeBook[j]), cbBounds, 3);
401 /* this updating is responsible for FG becoming BG again */
402 if (filter->framecount % filter->learning_interval == 0) {
403 for (j = 0; j < filter->width * filter->height; j++) {
404 update_codebook ((uchar *) filter->cvYUV->imageData + j * 3,
405 (codeBook *) & (filter->TcodeBook[j]), cbBounds, 3);
408 if (filter->framecount % 60 == 0) {
409 for (j = 0; j < filter->width * filter->height; j++)
410 clear_stale_entries ((codeBook *) & (filter->TcodeBook[j]));
413 for (j = 0; j < filter->width * filter->height; j++) {
415 ((uchar *) filter->cvYUV->imageData + j * 3,
416 (codeBook *) & (filter->TcodeBook[j]), 3, minMod, maxMod)) {
417 filter->cvFG->imageData[j] = 255;
419 filter->cvFG->imageData[j] = 0;
424 /* 3rd param is the smallest area to show: (w+h)/param , in pixels */
425 find_connected_components (filter->cvFG, 1, 10000,
426 filter->mem_storage, filter->contours);
429 /* Create the foreground and background masks using BackgroundSubtractorMOG [1],
430 * Gaussian Mixture-based Background/Foreground segmentation algorithm. OpenCV
431 * MOG implements the algorithm described in [2].
433 * [1] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog
434 * [2] P. KadewTraKuPong and R. Bowden, "An improved adaptive background
435 * mixture model for real-time tracking with shadow detection", Proc. 2nd
436 * European Workshop on Advanced Video-Based Surveillance Systems, 2001
438 else if (METHOD_MOG == filter->method) {
439 run_mog_iteration (filter);
441 /* Create the foreground and background masks using BackgroundSubtractorMOG2
442 * [1], Gaussian Mixture-based Background/Foreground segmentation algorithm.
443 * OpenCV MOG2 implements the algorithm described in [2] and [3].
445 * [1] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog2
446 * [2] Z.Zivkovic, "Improved adaptive Gausian mixture model for background
447 * subtraction", International Conference Pattern Recognition, UK, Aug 2004.
448 * [3] Z.Zivkovic, F. van der Heijden, "Efficient Adaptive Density Estimation
449 * per Image Pixel for the Task of Background Subtraction", Pattern
450 * Recognition Letters, vol. 27, no. 7, pages 773-780, 2006. */
451 else if (METHOD_MOG2 == filter->method) {
452 run_mog2_iteration (filter);
455 /* if we want to test_mode, just overwrite the output */
456 if (filter->test_mode) {
457 cvCvtColor (filter->cvFG, filter->cvRGB, CV_GRAY2RGB);
459 cvSplit (filter->cvRGB, filter->ch1, filter->ch2, filter->ch3, NULL);
461 cvSplit (filter->cvRGBA, filter->ch1, filter->ch2, filter->ch3, NULL);
463 /* copy anyhow the fg/bg to the alpha channel in the output image */
464 cvMerge (filter->ch1, filter->ch2, filter->ch3, filter->cvFG, filter->cvRGBA);
470 /* entry point to initialize the plug-in
471 * initialize the plug-in itself
472 * register the element factories and other features
475 gst_segmentation_plugin_init (GstPlugin * plugin)
477 GST_DEBUG_CATEGORY_INIT (gst_segmentation_debug, "segmentation",
478 0, "Performs Foreground/Background segmentation in video sequences");
480 return gst_element_register (plugin, "segmentation", GST_RANK_NONE,
481 GST_TYPE_SEGMENTATION);
486 #ifdef CODE_FROM_OREILLY_BOOK /* See license at the beginning of the page */
488 int update_codebook(uchar *p, codeBook &c, unsigned cbBounds)
489 Updates the codebook entry with a new data point
491 p Pointer to a YUV or HSI pixel
492 c Codebook for this pixel
493 cbBounds Learning bounds for codebook (Rule of thumb: 10)
494 numChannels Number of color channels we¡¯re learning
497 cvBounds must be of length equal to numChannels
503 update_codebook (unsigned char *p, codeBook * c, unsigned *cbBounds,
507 unsigned int high[3], low[3];
511 for (n = 0; n < numChannels; n++) {
512 high[n] = *(p + n) + *(cbBounds + n);
515 low[n] = *(p + n) - *(cbBounds + n);
520 /* SEE IF THIS FITS AN EXISTING CODEWORD */
521 for (i = 0; i < c->numEntries; i++) {
523 for (n = 0; n < numChannels; n++) {
524 if ((c->cb[i]->learnLow[n] <= *(p + n)) &&
525 /* Found an entry for this channel */
526 (*(p + n) <= c->cb[i]->learnHigh[n])) {
530 if (matchChannel == numChannels) { /* If an entry was found */
531 c->cb[i]->t_last_update = c->t;
532 /* adjust this codeword for the first channel */
533 for (n = 0; n < numChannels; n++) {
534 if (c->cb[i]->max[n] < *(p + n)) {
535 c->cb[i]->max[n] = *(p + n);
536 } else if (c->cb[i]->min[n] > *(p + n)) {
537 c->cb[i]->min[n] = *(p + n);
543 /* OVERHEAD TO TRACK POTENTIAL STALE ENTRIES */
544 for (int s = 0; s < c->numEntries; s++) {
545 /* Track which codebook entries are going stale: */
546 int negRun = c->t - c->cb[s]->t_last_update;
547 if (c->cb[s]->stale < negRun)
548 c->cb[s]->stale = negRun;
550 /* ENTER A NEW CODEWORD IF NEEDED */
551 if (i == c->numEntries) { /* if no existing codeword found, make one */
553 (code_element **) g_malloc (sizeof (code_element *) *
554 (c->numEntries + 1));
555 for (int ii = 0; ii < c->numEntries; ii++) {
556 foo[ii] = c->cb[ii]; /* copy all pointers */
558 foo[c->numEntries] = (code_element *) g_malloc (sizeof (code_element));
562 for (n = 0; n < numChannels; n++) {
563 c->cb[c->numEntries]->learnHigh[n] = high[n];
564 c->cb[c->numEntries]->learnLow[n] = low[n];
565 c->cb[c->numEntries]->max[n] = *(p + n);
566 c->cb[c->numEntries]->min[n] = *(p + n);
568 c->cb[c->numEntries]->t_last_update = c->t;
569 c->cb[c->numEntries]->stale = 0;
572 /* SLOWLY ADJUST LEARNING BOUNDS */
573 for (n = 0; n < numChannels; n++) {
574 if (c->cb[i]->learnHigh[n] < high[n])
575 c->cb[i]->learnHigh[n] += 1;
576 if (c->cb[i]->learnLow[n] > low[n])
577 c->cb[i]->learnLow[n] -= 1;
587 int clear_stale_entries(codeBook &c)
588 During learning, after you've learned for some period of time,
589 periodically call this to clear out stale codebook entries
591 c Codebook to clean up
594 number of entries cleared
597 clear_stale_entries (codeBook * c)
599 int staleThresh = c->t >> 1;
600 int *keep = (int *) g_malloc (sizeof (int) * (c->numEntries));
605 /* SEE WHICH CODEBOOK ENTRIES ARE TOO STALE */
606 for (int i = 0; i < c->numEntries; i++) {
607 if (c->cb[i]->stale > staleThresh)
608 keep[i] = 0; /* Mark for destruction */
610 keep[i] = 1; /* Mark to keep */
614 /* KEEP ONLY THE GOOD */
615 c->t = 0; /* Full reset on stale tracking */
616 foo = (code_element **) g_malloc (sizeof (code_element *) * keepCnt);
618 for (int ii = 0; ii < c->numEntries; ii++) {
621 /* We have to refresh these entries for next clearStale */
622 foo[k]->t_last_update = 0;
630 numCleared = c->numEntries - keepCnt;
631 c->numEntries = keepCnt;
638 uchar background_diff( uchar *p, codeBook &c,
639 int minMod, int maxMod)
640 Given a pixel and a codebook, determine if the pixel is
641 covered by the codebook
643 p Pixel pointer (YUV interleaved)
645 numChannels Number of channels we are testing
646 maxMod Add this (possibly negative) number onto
648 max level when determining if new pixel is foreground
649 minMod Subract this (possibly negative) number from
650 min level when determining if new pixel is foreground
653 minMod and maxMod must have length numChannels,
654 e.g. 3 channels => minMod[3], maxMod[3]. There is one min and
655 one max threshold per channel.
658 0 => background, 255 => foreground
661 background_diff (unsigned char *p, codeBook * c, int numChannels,
662 int *minMod, int *maxMod)
665 /* SEE IF THIS FITS AN EXISTING CODEWORD */
667 for (i = 0; i < c->numEntries; i++) {
669 for (int n = 0; n < numChannels; n++) {
670 if ((c->cb[i]->min[n] - minMod[n] <= *(p + n)) &&
671 (*(p + n) <= c->cb[i]->max[n] + maxMod[n])) {
672 matchChannel++; /* Found an entry for this channel */
677 if (matchChannel == numChannels) {
678 break; /* Found an entry that matched all channels */
681 if (i >= c->numEntries)
690 void find_connected_components(IplImage *mask, int poly1_hull0,
691 float perimScale, int *num,
692 CvRect *bbs, CvPoint *centers)
693 This cleans up the foreground segmentation mask derived from calls
696 mask Is a grayscale (8-bit depth) “rawâ€
\9d mask image that
700 poly1_hull0 If set, approximate connected component by
701 (DEFAULT) polygon, or else convex hull (0)
702 perimScale Len = image (width+height)/perimScale. If contour
703 len < this, delete that contour (DEFAULT: 4)
704 num Maximum number of rectangles and/or centers to
705 return; on return, will contain number filled
707 bbs Pointer to bounding box rectangle vector of
708 length num. (DEFAULT SETTING: NULL)
709 centers Pointer to contour centers vector of length
713 /* Approx.threshold - the bigger it is, the simpler is the boundary */
714 #define CVCONTOUR_APPROX_LEVEL 1
715 /* How many iterations of erosion and/or dilation there should be */
716 #define CVCLOSE_ITR 1
718 find_connected_components (IplImage * mask, int poly1_hull0, float perimScale,
719 CvMemStorage * mem_storage, CvSeq * contours)
721 CvContourScanner scanner;
724 /* Just some convenience variables */
725 const CvScalar CVX_WHITE = CV_RGB (0xff, 0xff, 0xff);
726 const CvScalar CVX_BLACK = CV_RGB (0x00, 0x00, 0x00);
728 /* CLEAN UP RAW MASK */
729 cvMorphologyEx (mask, mask, 0, 0, CV_MOP_OPEN, CVCLOSE_ITR);
730 cvMorphologyEx (mask, mask, 0, 0, CV_MOP_CLOSE, CVCLOSE_ITR);
731 /* FIND CONTOURS AROUND ONLY BIGGER REGIONS */
732 if (mem_storage == NULL) {
733 mem_storage = cvCreateMemStorage (0);
735 cvClearMemStorage (mem_storage);
738 scanner = cvStartFindContours (mask, mem_storage, sizeof (CvContour),
739 CV_RETR_TREE, CV_CHAIN_APPROX_SIMPLE, cvPoint (0, 0));
741 while ((c = cvFindNextContour (scanner)) != NULL) {
742 double len = cvContourArea (c, CV_WHOLE_SEQ, 0);
743 /* calculate perimeter len threshold: */
744 double q = (mask->height + mask->width) / perimScale;
745 /* Get rid of blob if its perimeter is too small: */
747 cvSubstituteContour (scanner, NULL);
749 /* Smooth its edges if its large enough */
752 /* Polygonal approximation */
754 cvApproxPoly (c, sizeof (CvContour), mem_storage, CV_POLY_APPROX_DP,
755 CVCONTOUR_APPROX_LEVEL, 0);
757 /* Convex Hull of the segmentation */
758 c_new = cvConvexHull2 (c, mem_storage, CV_CLOCKWISE, 1);
760 cvSubstituteContour (scanner, c_new);
764 contours = cvEndFindContours (&scanner);
766 /* PAINT THE FOUND REGIONS BACK INTO THE IMAGE */
768 /* DRAW PROCESSED CONTOURS INTO THE MASK */
769 for (c = contours; c != NULL; c = c->h_next)
770 cvDrawContours (mask, c, CVX_WHITE, CVX_BLACK, -1, CV_FILLED, 8, cvPoint (0,
773 #endif /*ifdef CODE_FROM_OREILLY_BOOK */
777 initialise_mog (GstSegmentation * filter)
779 filter->img_input_as_cvMat = (void *) new cv::Mat (filter->cvYUV, false);
780 filter->img_fg_as_cvMat = (void *) new cv::Mat (filter->cvFG, false);
782 filter->mog = (void *) new cv::BackgroundSubtractorMOG ();
783 filter->mog2 = (void *) new cv::BackgroundSubtractorMOG2 ();
789 run_mog_iteration (GstSegmentation * filter)
791 ((cv::Mat *) filter->img_input_as_cvMat)->data =
792 (uchar *) filter->cvYUV->imageData;
793 ((cv::Mat *) filter->img_fg_as_cvMat)->data =
794 (uchar *) filter->cvFG->imageData;
797 BackgroundSubtractorMOG [1], Gaussian Mixture-based Background/Foreground
798 Segmentation Algorithm. OpenCV MOG implements the algorithm described in [2].
800 [1] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog
801 [2] P. KadewTraKuPong and R. Bowden, "An improved adaptive background
802 mixture model for real-time tracking with shadow detection", Proc. 2nd
803 European Workshop on Advanced Video-Based Surveillance Systems, 2001
806 (*((cv::BackgroundSubtractorMOG *) filter->mog)) (*((cv::Mat *) filter->
807 img_input_as_cvMat), *((cv::Mat *) filter->img_fg_as_cvMat),
808 filter->learning_rate);
814 run_mog2_iteration (GstSegmentation * filter)
816 ((cv::Mat *) filter->img_input_as_cvMat)->data =
817 (uchar *) filter->cvYUV->imageData;
818 ((cv::Mat *) filter->img_fg_as_cvMat)->data =
819 (uchar *) filter->cvFG->imageData;
822 BackgroundSubtractorMOG2 [1], Gaussian Mixture-based Background/Foreground
823 segmentation algorithm. OpenCV MOG2 implements the algorithm described in
826 [1] http://opencv.itseez.com/modules/video/doc/motion_analysis_and_object_tracking.html#backgroundsubtractormog2
827 [2] Z.Zivkovic, "Improved adaptive Gausian mixture model for background
828 subtraction", International Conference Pattern Recognition, UK, August, 2004.
829 [3] Z.Zivkovic, F. van der Heijden, "Efficient Adaptive Density Estimation per
830 Image Pixel for the Task of Background Subtraction", Pattern Recognition
831 Letters, vol. 27, no. 7, pages 773-780, 2006.
834 (*((cv::BackgroundSubtractorMOG *) filter->mog2)) (*((cv::Mat *) filter->
835 img_input_as_cvMat), *((cv::Mat *) filter->img_fg_as_cvMat),
836 filter->learning_rate);
842 finalise_mog (GstSegmentation * filter)
844 delete (cv::Mat *) filter->img_input_as_cvMat;
845 delete (cv::Mat *) filter->img_fg_as_cvMat;
846 delete (cv::BackgroundSubtractorMOG *) filter->mog;
847 delete (cv::BackgroundSubtractorMOG2 *) filter->mog2;