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
query_formats(AVFilterContext * ctx)49 static int query_formats(AVFilterContext *ctx)
50 {
51 static const enum AVPixelFormat pix_fmts[] = {
52 AV_PIX_FMT_GRAY8, AV_PIX_FMT_GRAY9,
53 AV_PIX_FMT_GRAY10, AV_PIX_FMT_GRAY12, AV_PIX_FMT_GRAY14,
54 AV_PIX_FMT_GRAY16,
55 AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV411P,
56 AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P,
57 AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV444P,
58 AV_PIX_FMT_YUVJ420P, AV_PIX_FMT_YUVJ422P,
59 AV_PIX_FMT_YUVJ440P, AV_PIX_FMT_YUVJ444P,
60 AV_PIX_FMT_YUVJ411P,
61 AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
62 AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
63 AV_PIX_FMT_YUV440P10,
64 AV_PIX_FMT_YUV444P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV420P12,
65 AV_PIX_FMT_YUV440P12,
66 AV_PIX_FMT_YUV444P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV420P14,
67 AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
68 AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
69 AV_PIX_FMT_YUVA444P9, AV_PIX_FMT_YUVA444P10, AV_PIX_FMT_YUVA444P12, AV_PIX_FMT_YUVA444P16,
70 AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA422P16,
71 AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA420P16,
72 AV_PIX_FMT_NONE,
73 };
74
75 AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts);
76 if (!fmts_list)
77 return AVERROR(ENOMEM);
78 return ff_set_common_formats(ctx, fmts_list);
79 }
80
81 #define SET_META(key, value) \
82 av_dict_set_int(metadata, key, value, 0);
83
filter_frame(AVFilterLink * inlink,AVFrame * frame)84 static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
85 {
86 AVFilterContext *ctx = inlink->dst;
87 BBoxContext *bbox = ctx->priv;
88 FFBoundingBox box;
89 int has_bbox, w, h;
90
91 has_bbox =
92 ff_calculate_bounding_box(&box,
93 frame->data[0], frame->linesize[0],
94 inlink->w, inlink->h, bbox->min_val, bbox->depth);
95 w = box.x2 - box.x1 + 1;
96 h = box.y2 - box.y1 + 1;
97
98 av_log(ctx, AV_LOG_INFO,
99 "n:%"PRId64" pts:%s pts_time:%s", inlink->frame_count_out,
100 av_ts2str(frame->pts), av_ts2timestr(frame->pts, &inlink->time_base));
101
102 if (has_bbox) {
103 AVDictionary **metadata = &frame->metadata;
104
105 SET_META("lavfi.bbox.x1", box.x1)
106 SET_META("lavfi.bbox.x2", box.x2)
107 SET_META("lavfi.bbox.y1", box.y1)
108 SET_META("lavfi.bbox.y2", box.y2)
109 SET_META("lavfi.bbox.w", w)
110 SET_META("lavfi.bbox.h", h)
111
112 av_log(ctx, AV_LOG_INFO,
113 " x1:%d x2:%d y1:%d y2:%d w:%d h:%d"
114 " crop=%d:%d:%d:%d drawbox=%d:%d:%d:%d",
115 box.x1, box.x2, box.y1, box.y2, w, h,
116 w, h, box.x1, box.y1, /* crop params */
117 box.x1, box.y1, w, h); /* drawbox params */
118 }
119 av_log(ctx, AV_LOG_INFO, "\n");
120
121 return ff_filter_frame(inlink->dst->outputs[0], frame);
122 }
123
config_output(AVFilterLink * outlink)124 static int config_output(AVFilterLink *outlink)
125 {
126 AVFilterContext *ctx = outlink->src;
127 BBoxContext *s = ctx->priv;
128 const AVPixFmtDescriptor *desc;
129
130 desc = av_pix_fmt_desc_get(outlink->format);
131 if (!desc)
132 return AVERROR_BUG;
133 s->depth = desc->comp[0].depth;
134
135 return 0;
136 }
137
138 static const AVFilterPad bbox_inputs[] = {
139 {
140 .name = "default",
141 .type = AVMEDIA_TYPE_VIDEO,
142 .filter_frame = filter_frame,
143 },
144 { NULL }
145 };
146
147 static const AVFilterPad bbox_outputs[] = {
148 {
149 .name = "default",
150 .type = AVMEDIA_TYPE_VIDEO,
151 .config_props = config_output,
152 },
153 { NULL }
154 };
155
156 AVFilter ff_vf_bbox = {
157 .name = "bbox",
158 .description = NULL_IF_CONFIG_SMALL("Compute bounding box for each frame."),
159 .priv_size = sizeof(BBoxContext),
160 .priv_class = &bbox_class,
161 .query_formats = query_formats,
162 .inputs = bbox_inputs,
163 .outputs = bbox_outputs,
164 .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,
165 .process_command = ff_filter_process_command,
166 };
167