FFmpeg
vf_edgedetect.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2012-2014 Clément Bœsch <u pkh me>
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20 
21 /**
22  * @file
23  * Edge detection filter
24  *
25  * @see https://en.wikipedia.org/wiki/Canny_edge_detector
26  */
27 
28 #include "libavutil/avassert.h"
29 #include "libavutil/imgutils.h"
30 #include "libavutil/opt.h"
31 #include "avfilter.h"
32 #include "formats.h"
33 #include "internal.h"
34 #include "video.h"
35 #include "edge_common.h"
36 
37 #define PLANE_R 0x4
38 #define PLANE_G 0x1
39 #define PLANE_B 0x2
40 #define PLANE_Y 0x1
41 #define PLANE_U 0x2
42 #define PLANE_V 0x4
43 #define PLANE_A 0x8
44 
45 enum FilterMode {
50 };
51 
52 struct plane_info {
53  uint8_t *tmpbuf;
54  uint16_t *gradients;
55  char *directions;
56  int width, height;
57 };
58 
59 typedef struct EdgeDetectContext {
60  const AVClass *class;
61  struct plane_info planes[3];
63  int nb_planes;
64  double low, high;
65  uint8_t low_u8, high_u8;
66  int mode;
68 
69 #define OFFSET(x) offsetof(EdgeDetectContext, x)
70 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
71 static const AVOption edgedetect_options[] = {
72  { "high", "set high threshold", OFFSET(high), AV_OPT_TYPE_DOUBLE, {.dbl=50/255.}, 0, 1, FLAGS },
73  { "low", "set low threshold", OFFSET(low), AV_OPT_TYPE_DOUBLE, {.dbl=20/255.}, 0, 1, FLAGS },
74  { "mode", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=MODE_WIRES}, 0, NB_MODE-1, FLAGS, "mode" },
75  { "wires", "white/gray wires on black", 0, AV_OPT_TYPE_CONST, {.i64=MODE_WIRES}, INT_MIN, INT_MAX, FLAGS, "mode" },
76  { "colormix", "mix colors", 0, AV_OPT_TYPE_CONST, {.i64=MODE_COLORMIX}, INT_MIN, INT_MAX, FLAGS, "mode" },
77  { "canny", "detect edges on planes", 0, AV_OPT_TYPE_CONST, {.i64=MODE_CANNY}, INT_MIN, INT_MAX, FLAGS, "mode" },
78  { "planes", "set planes to filter", OFFSET(filter_planes), AV_OPT_TYPE_FLAGS, {.i64=7}, 1, 0x7, FLAGS, "flags" },
79  { "y", "filter luma plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_Y}, 0, 0, FLAGS, "flags" },
80  { "u", "filter u plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_U}, 0, 0, FLAGS, "flags" },
81  { "v", "filter v plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_V}, 0, 0, FLAGS, "flags" },
82  { "r", "filter red plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_R}, 0, 0, FLAGS, "flags" },
83  { "g", "filter green plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_G}, 0, 0, FLAGS, "flags" },
84  { "b", "filter blue plane", 0, AV_OPT_TYPE_CONST, {.i64=PLANE_B}, 0, 0, FLAGS, "flags" },
85  { NULL }
86 };
87 
89 
91 {
93 
94  edgedetect->low_u8 = edgedetect->low * 255. + .5;
95  edgedetect->high_u8 = edgedetect->high * 255. + .5;
96  return 0;
97 }
98 
100 {
101  const EdgeDetectContext *edgedetect = ctx->priv;
102  static const enum AVPixelFormat wires_pix_fmts[] = {AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE};
104  static const enum AVPixelFormat colormix_pix_fmts[] = {AV_PIX_FMT_GBRP, AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE};
105  const enum AVPixelFormat *pix_fmts = NULL;
106 
107  if (edgedetect->mode == MODE_WIRES) {
108  pix_fmts = wires_pix_fmts;
109  } else if (edgedetect->mode == MODE_COLORMIX) {
110  pix_fmts = colormix_pix_fmts;
111  } else if (edgedetect->mode == MODE_CANNY) {
112  pix_fmts = canny_pix_fmts;
113  } else {
114  av_assert0(0);
115  }
117 }
118 
120 {
121  int p;
122  AVFilterContext *ctx = inlink->dst;
125 
126  edgedetect->nb_planes = inlink->format == AV_PIX_FMT_GRAY8 ? 1 : 3;
127  for (p = 0; p < edgedetect->nb_planes; p++) {
128  struct plane_info *plane = &edgedetect->planes[p];
129  int vsub = p ? desc->log2_chroma_h : 0;
130  int hsub = p ? desc->log2_chroma_w : 0;
131 
132  plane->width = AV_CEIL_RSHIFT(inlink->w, hsub);
133  plane->height = AV_CEIL_RSHIFT(inlink->h, vsub);
134  plane->tmpbuf = av_malloc(plane->width * plane->height);
135  plane->gradients = av_calloc(plane->width * plane->height, sizeof(*plane->gradients));
136  plane->directions = av_malloc(plane->width * plane->height);
137  if (!plane->tmpbuf || !plane->gradients || !plane->directions)
138  return AVERROR(ENOMEM);
139  }
140  return 0;
141 }
142 
143 static void color_mix(int w, int h,
144  uint8_t *dst, int dst_linesize,
145  const uint8_t *src, int src_linesize)
146 {
147  int i, j;
148 
149  for (j = 0; j < h; j++) {
150  for (i = 0; i < w; i++)
151  dst[i] = (dst[i] + src[i]) >> 1;
152  dst += dst_linesize;
153  src += src_linesize;
154  }
155 }
156 
158 {
159  AVFilterContext *ctx = inlink->dst;
161  AVFilterLink *outlink = ctx->outputs[0];
162  int p, direct = 0;
163  AVFrame *out;
164 
165  if (edgedetect->mode != MODE_COLORMIX && av_frame_is_writable(in)) {
166  direct = 1;
167  out = in;
168  } else {
169  out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
170  if (!out) {
171  av_frame_free(&in);
172  return AVERROR(ENOMEM);
173  }
175  }
176 
177  for (p = 0; p < edgedetect->nb_planes; p++) {
178  struct plane_info *plane = &edgedetect->planes[p];
179  uint8_t *tmpbuf = plane->tmpbuf;
180  uint16_t *gradients = plane->gradients;
181  int8_t *directions = plane->directions;
182  const int width = plane->width;
183  const int height = plane->height;
184 
185  if (!((1 << p) & edgedetect->filter_planes)) {
186  if (!direct)
187  av_image_copy_plane(out->data[p], out->linesize[p],
188  in->data[p], in->linesize[p],
189  width, height);
190  continue;
191  }
192 
193  /* gaussian filter to reduce noise */
194  ff_gaussian_blur_8(width, height,
195  tmpbuf, width,
196  in->data[p], in->linesize[p], 1);
197 
198  /* compute the 16-bits gradients and directions for the next step */
199  ff_sobel_8(width, height,
200  gradients, width,
202  tmpbuf, width, 1);
203 
204  /* non_maximum_suppression() will actually keep & clip what's necessary and
205  * ignore the rest, so we need a clean output buffer */
206  memset(tmpbuf, 0, width * height);
208  tmpbuf, width,
210  gradients, width);
211 
212  /* keep high values, or low values surrounded by high values */
213  ff_double_threshold(edgedetect->low_u8, edgedetect->high_u8,
214  width, height,
215  out->data[p], out->linesize[p],
216  tmpbuf, width);
217 
218  if (edgedetect->mode == MODE_COLORMIX) {
220  out->data[p], out->linesize[p],
221  in->data[p], in->linesize[p]);
222  }
223  }
224 
225  if (!direct)
226  av_frame_free(&in);
227  return ff_filter_frame(outlink, out);
228 }
229 
231 {
232  int p;
234 
235  for (p = 0; p < edgedetect->nb_planes; p++) {
236  struct plane_info *plane = &edgedetect->planes[p];
237  av_freep(&plane->tmpbuf);
238  av_freep(&plane->gradients);
239  av_freep(&plane->directions);
240  }
241 }
242 
243 static const AVFilterPad edgedetect_inputs[] = {
244  {
245  .name = "default",
246  .type = AVMEDIA_TYPE_VIDEO,
247  .config_props = config_props,
248  .filter_frame = filter_frame,
249  },
250 };
251 
252 static const AVFilterPad edgedetect_outputs[] = {
253  {
254  .name = "default",
255  .type = AVMEDIA_TYPE_VIDEO,
256  },
257 };
258 
260  .name = "edgedetect",
261  .description = NULL_IF_CONFIG_SMALL("Detect and draw edge."),
262  .priv_size = sizeof(EdgeDetectContext),
263  .init = init,
264  .uninit = uninit,
268  .priv_class = &edgedetect_class,
270 };
ff_get_video_buffer
AVFrame * ff_get_video_buffer(AVFilterLink *link, int w, int h)
Request a picture buffer with a specific set of permissions.
Definition: video.c:101
AVPixelFormat
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
AVERROR
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
opt.h
out
FILE * out
Definition: movenc.c:54
ff_filter_frame
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
Definition: avfilter.c:969
av_pix_fmt_desc_get
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
Definition: pixdesc.c:2888
PLANE_B
#define PLANE_B
Definition: vf_edgedetect.c:39
edgedetect
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied your filters are likely to have a very short lifetime due to more or less regular internal API and a limited and testing changes the pixels in whatever fashion you and outputs the modified frame The most simple way of doing this is to take a similar filter We ll pick edgedetect
Definition: writing_filters.txt:16
inlink
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is inlink
Definition: filter_design.txt:212
av_frame_free
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
Definition: frame.c:99
AVFrame
This structure describes decoded (raw) audio or video data.
Definition: frame.h:330
w
uint8_t w
Definition: llviddspenc.c:38
edge_common.h
AVOption
AVOption.
Definition: opt.h:251
FILTER_QUERY_FUNC
#define FILTER_QUERY_FUNC(func)
Definition: internal.h:171
AVFilter::name
const char * name
Filter name.
Definition: avfilter.h:165
video.h
AVFrame::data
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:351
av_malloc
#define av_malloc(s)
Definition: tableprint_vlc.h:30
av_image_copy_plane
void av_image_copy_plane(uint8_t *dst, int dst_linesize, const uint8_t *src, int src_linesize, int bytewidth, int height)
Copy image plane from src to dst.
Definition: imgutils.c:374
hsub
static void hsub(htype *dst, const htype *src, int bins)
Definition: vf_median.c:74
formats.h
plane_info::height
int height
Definition: vf_edgedetect.c:56
FilterMode
FilterMode
Definition: vp9.h:64
MODE_COLORMIX
@ MODE_COLORMIX
Definition: vf_edgedetect.c:47
AVFilterPad
A filter pad used for either input or output.
Definition: internal.h:49
PLANE_V
#define PLANE_V
Definition: vf_edgedetect.c:42
avassert.h
av_cold
#define av_cold
Definition: attributes.h:90
width
#define width
AV_CEIL_RSHIFT
#define AV_CEIL_RSHIFT(a, b)
Definition: common.h:50
AV_OPT_TYPE_DOUBLE
@ AV_OPT_TYPE_DOUBLE
Definition: opt.h:227
ff_set_common_formats_from_list
int ff_set_common_formats_from_list(AVFilterContext *ctx, const int *fmts)
Equivalent to ff_set_common_formats(ctx, ff_make_format_list(fmts))
Definition: formats.c:755
MODE_CANNY
@ MODE_CANNY
Definition: vf_edgedetect.c:48
av_assert0
#define av_assert0(cond)
assert() equivalent, that is always enabled.
Definition: avassert.h:37
pix_fmts
static enum AVPixelFormat pix_fmts[]
Definition: libkvazaar.c:296
MODE_WIRES
@ MODE_WIRES
Definition: vf_edgedetect.c:46
ctx
AVFormatContext * ctx
Definition: movenc.c:48
AV_PIX_FMT_YUV420P
@ AV_PIX_FMT_YUV420P
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
Definition: pixfmt.h:66
FILTER_INPUTS
#define FILTER_INPUTS(array)
Definition: internal.h:194
edgedetect_inputs
static const AVFilterPad edgedetect_inputs[]
Definition: vf_edgedetect.c:243
init
static av_cold int init(AVFilterContext *ctx)
Definition: vf_edgedetect.c:90
AVClass
Describe the class of an AVClass context structure.
Definition: log.h:66
NULL
#define NULL
Definition: coverity.c:32
av_frame_copy_props
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
Definition: frame.c:594
EdgeDetectContext::high
double high
Definition: vf_edgedetect.c:64
PLANE_G
#define PLANE_G
Definition: vf_edgedetect.c:38
ff_vf_edgedetect
const AVFilter ff_vf_edgedetect
Definition: vf_edgedetect.c:259
ff_non_maximum_suppression
void ff_non_maximum_suppression(int w, int h, uint8_t *dst, int dst_linesize, const int8_t *dir, int dir_linesize, const uint16_t *src, int src_linesize)
Filters rounded gradients to drop all non-maxima pixels in the magnitude image Expects gradients gene...
Definition: edge_common.c:60
plane_info::directions
char * directions
Definition: vf_edgedetect.c:55
AV_PIX_FMT_GRAY8
@ AV_PIX_FMT_GRAY8
Y , 8bpp.
Definition: pixfmt.h:74
filter_frame
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
Definition: vf_edgedetect.c:157
uninit
static av_cold void uninit(AVFilterContext *ctx)
Definition: vf_edgedetect.c:230
NULL_IF_CONFIG_SMALL
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification.
Definition: internal.h:115
ff_double_threshold
void ff_double_threshold(int low, int high, int w, int h, uint8_t *dst, int dst_linesize, const uint8_t *src, int src_linesize)
Filters all pixels in src to keep all pixels > high, and keep all pixels > low where all surrounding ...
Definition: edge_common.c:89
EdgeDetectContext::mode
int mode
Definition: vf_edgedetect.c:66
av_frame_is_writable
int av_frame_is_writable(AVFrame *frame)
Check if the frame data is writable.
Definition: frame.c:524
height
#define height
plane_info::gradients
uint16_t * gradients
Definition: vf_edgedetect.c:54
plane_info::width
int width
Definition: vf_edgedetect.c:56
internal.h
AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC
#define AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC
Some filters support a generic "enable" expression option that can be used to enable or disable a fil...
Definition: avfilter.h:142
EdgeDetectContext::low
double low
Definition: vf_edgedetect.c:64
i
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:269
PLANE_R
#define PLANE_R
Definition: vf_edgedetect.c:37
EdgeDetectContext::planes
struct plane_info planes[3]
Definition: vf_edgedetect.c:61
EdgeDetectContext::filter_planes
int filter_planes
Definition: vf_edgedetect.c:62
plane_info
Definition: vf_edgedetect.c:52
EdgeDetectContext::low_u8
uint8_t low_u8
Definition: vf_edgedetect.c:65
plane_info::tmpbuf
uint8_t * tmpbuf
Definition: vf_edgedetect.c:53
AVFilterPad::name
const char * name
Pad name.
Definition: internal.h:55
config_props
static int config_props(AVFilterLink *inlink)
Definition: vf_edgedetect.c:119
av_calloc
void * av_calloc(size_t nmemb, size_t size)
Definition: mem.c:262
AVFilter
Filter definition.
Definition: avfilter.h:161
EdgeDetectContext::high_u8
uint8_t high_u8
Definition: vf_edgedetect.c:65
edgedetect_options
static const AVOption edgedetect_options[]
Definition: vf_edgedetect.c:71
query_formats
static int query_formats(AVFilterContext *ctx)
Definition: vf_edgedetect.c:99
PLANE_Y
#define PLANE_Y
Definition: vf_edgedetect.c:40
mode
mode
Definition: ebur128.h:83
AV_PIX_FMT_NONE
@ AV_PIX_FMT_NONE
Definition: pixfmt.h:65
AV_OPT_TYPE_INT
@ AV_OPT_TYPE_INT
Definition: opt.h:225
avfilter.h
EdgeDetectContext::nb_planes
int nb_planes
Definition: vf_edgedetect.c:63
AVFILTER_DEFINE_CLASS
AVFILTER_DEFINE_CLASS(edgedetect)
AV_PIX_FMT_YUV444P
@ AV_PIX_FMT_YUV444P
planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples)
Definition: pixfmt.h:71
OFFSET
#define OFFSET(x)
Definition: vf_edgedetect.c:69
AVFilterContext
An instance of a filter.
Definition: avfilter.h:392
AV_PIX_FMT_GBRP
@ AV_PIX_FMT_GBRP
planar GBR 4:4:4 24bpp
Definition: pixfmt.h:158
desc
const char * desc
Definition: libsvtav1.c:83
AVMEDIA_TYPE_VIDEO
@ AVMEDIA_TYPE_VIDEO
Definition: avutil.h:201
FLAGS
#define FLAGS
Definition: vf_edgedetect.c:70
AV_PIX_FMT_YUV422P
@ AV_PIX_FMT_YUV422P
planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples)
Definition: pixfmt.h:70
AVPixFmtDescriptor
Descriptor that unambiguously describes how the bits of a pixel are stored in the up to 4 data planes...
Definition: pixdesc.h:69
FILTER_OUTPUTS
#define FILTER_OUTPUTS(array)
Definition: internal.h:195
av_freep
#define av_freep(p)
Definition: tableprint_vlc.h:34
src
INIT_CLIP pixel * src
Definition: h264pred_template.c:418
PLANE_U
#define PLANE_U
Definition: vf_edgedetect.c:41
AV_OPT_TYPE_FLAGS
@ AV_OPT_TYPE_FLAGS
Definition: opt.h:224
imgutils.h
AVFrame::linesize
int linesize[AV_NUM_DATA_POINTERS]
For video, a positive or negative value, which is typically indicating the size in bytes of each pict...
Definition: frame.h:375
NB_MODE
@ NB_MODE
Definition: vf_edgedetect.c:49
h
h
Definition: vp9dsp_template.c:2038
edgedetect_outputs
static const AVFilterPad edgedetect_outputs[]
Definition: vf_edgedetect.c:252
color_mix
static void color_mix(int w, int h, uint8_t *dst, int dst_linesize, const uint8_t *src, int src_linesize)
Definition: vf_edgedetect.c:143
AV_OPT_TYPE_CONST
@ AV_OPT_TYPE_CONST
Definition: opt.h:234
EdgeDetectContext
Definition: vf_edgedetect.c:59