1 /*
2 * Copyright (c) 2012 Stefano Sabatini
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 * bounding box detection filter
24 */
25
26 #include "libavutil/opt.h"
27 #include "libavutil/pixdesc.h"
28 #include "libavutil/timestamp.h"
29 #include "avfilter.h"
30 #include "bbox.h"
31 #include "internal.h"
32
33 typedef struct BBoxContext {
34 const AVClass *class;
35 int min_val;
36 int depth;
37 } BBoxContext;
38
39 #define OFFSET(x) offsetof(BBoxContext, x)
40 #define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
41
42 static const AVOption bbox_options[] = {
43 { "min_val", "set minimum luminance value for bounding box", OFFSET(min_val), AV_OPT_TYPE_INT, { .i64 = 16 }, 0, UINT16_MAX, FLAGS },
44 { NULL }
45 };
46
47 AVFILTER_DEFINE_CLASS(bbox);
48
49 static const enum AVPixelFormat pix_fmts[] = {
50 AV_PIX_FMT_GRAY8, AV_PIX_FMT_GRAY9,
51 AV_PIX_FMT_GRAY10, AV_PIX_FMT_GRAY12, AV_PIX_FMT_GRAY14,
52 AV_PIX_FMT_GRAY16,
53 AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV411P,
54 AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P,
55 AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV444P,
56 AV_PIX_FMT_YUVJ420P, AV_PIX_FMT_YUVJ422P,
57 AV_PIX_FMT_YUVJ440P, AV_PIX_FMT_YUVJ444P,
58 AV_PIX_FMT_YUVJ411P,
59 AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
60 AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
61 AV_PIX_FMT_YUV440P10,
62 AV_PIX_FMT_YUV444P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV420P12,
63 AV_PIX_FMT_YUV440P12,
64 AV_PIX_FMT_YUV444P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV420P14,
65 AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
66 AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
67 AV_PIX_FMT_YUVA444P9, AV_PIX_FMT_YUVA444P10, AV_PIX_FMT_YUVA444P12, AV_PIX_FMT_YUVA444P16,
68 AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA422P16,
69 AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA420P16,
70 AV_PIX_FMT_NONE,
71 };
72
73 #define SET_META(key, value) \
74 av_dict_set_int(metadata, key, value, 0);
75
filter_frame(AVFilterLink * inlink,AVFrame * frame)76 static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
77 {
78 AVFilterContext *ctx = inlink->dst;
79 BBoxContext *bbox = ctx->priv;
80 FFBoundingBox box;
81 int has_bbox, w, h;
82
83 has_bbox =
84 ff_calculate_bounding_box(&box,
85 frame->data[0], frame->linesize[0],
86 inlink->w, inlink->h, bbox->min_val, bbox->depth);
87 w = box.x2 - box.x1 + 1;
88 h = box.y2 - box.y1 + 1;
89
90 av_log(ctx, AV_LOG_INFO,
91 "n:%"PRId64" pts:%s pts_time:%s", inlink->frame_count_out,
92 av_ts2str(frame->pts), av_ts2timestr(frame->pts, &inlink->time_base));
93
94 if (has_bbox) {
95 AVDictionary **metadata = &frame->metadata;
96
97 SET_META("lavfi.bbox.x1", box.x1)
98 SET_META("lavfi.bbox.x2", box.x2)
99 SET_META("lavfi.bbox.y1", box.y1)
100 SET_META("lavfi.bbox.y2", box.y2)
101 SET_META("lavfi.bbox.w", w)
102 SET_META("lavfi.bbox.h", h)
103
104 av_log(ctx, AV_LOG_INFO,
105 " x1:%d x2:%d y1:%d y2:%d w:%d h:%d"
106 " crop=%d:%d:%d:%d drawbox=%d:%d:%d:%d",
107 box.x1, box.x2, box.y1, box.y2, w, h,
108 w, h, box.x1, box.y1, /* crop params */
109 box.x1, box.y1, w, h); /* drawbox params */
110 }
111 av_log(ctx, AV_LOG_INFO, "\n");
112
113 return ff_filter_frame(inlink->dst->outputs[0], frame);
114 }
115
config_output(AVFilterLink * outlink)116 static int config_output(AVFilterLink *outlink)
117 {
118 AVFilterContext *ctx = outlink->src;
119 BBoxContext *s = ctx->priv;
120 const AVPixFmtDescriptor *desc;
121
122 desc = av_pix_fmt_desc_get(outlink->format);
123 if (!desc)
124 return AVERROR_BUG;
125 s->depth = desc->comp[0].depth;
126
127 return 0;
128 }
129
130 static const AVFilterPad bbox_inputs[] = {
131 {
132 .name = "default",
133 .type = AVMEDIA_TYPE_VIDEO,
134 .filter_frame = filter_frame,
135 },
136 };
137
138 static const AVFilterPad bbox_outputs[] = {
139 {
140 .name = "default",
141 .type = AVMEDIA_TYPE_VIDEO,
142 .config_props = config_output,
143 },
144 };
145
146 const AVFilter ff_vf_bbox = {
147 .name = "bbox",
148 .description = NULL_IF_CONFIG_SMALL("Compute bounding box for each frame."),
149 .priv_size = sizeof(BBoxContext),
150 .priv_class = &bbox_class,
151 FILTER_INPUTS(bbox_inputs),
152 FILTER_OUTPUTS(bbox_outputs),
153 FILTER_PIXFMTS_ARRAY(pix_fmts),
154 .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC | AVFILTER_FLAG_METADATA_ONLY,
155 .process_command = ff_filter_process_command,
156 };
157