vf_cropdetect.c 6.39 KB
Newer Older
1
/*
2
 * Copyright (c) 2002 A'rpi
3
 * This file is part of Libav.
4
 *
5
 * Libav is free software; you can redistribute it and/or modify
6 7 8 9
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
10
 * Libav is distributed in the hope that it will be useful,
11 12 13 14 15
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License along
16
 * with Libav; if not, write to the Free Software Foundation, Inc.,
17 18 19 20 21 22 23 24 25
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

/**
 * @file
 * border detection filter
 * Ported from MPlayer libmpcodecs/vf_cropdetect.c.
 */

26 27
#include <stdio.h>

28
#include "libavutil/imgutils.h"
29
#include "libavutil/internal.h"
30
#include "avfilter.h"
31
#include "formats.h"
32
#include "internal.h"
33
#include "video.h"
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

typedef struct {
    int x1, y1, x2, y2;
    int limit;
    int round;
    int reset_count;
    int frame_nb;
    int max_pixsteps[4];
} CropDetectContext;

static int query_formats(AVFilterContext *ctx)
{
    static const enum PixelFormat pix_fmts[] = {
        PIX_FMT_YUV420P, PIX_FMT_YUVJ420P,
        PIX_FMT_YUV422P, PIX_FMT_YUVJ422P,
        PIX_FMT_YUV444P, PIX_FMT_YUVJ444P,
        PIX_FMT_YUV411P, PIX_FMT_GRAY8,
        PIX_FMT_NV12,    PIX_FMT_NV21,
        PIX_FMT_NONE
    };

55
    ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
    return 0;
}

static int checkline(void *ctx, const unsigned char *src, int stride, int len, int bpp)
{
    int total = 0;
    int div = len;

    switch (bpp) {
    case 1:
        while (--len >= 0) {
            total += src[0];
            src += stride;
        }
        break;
    case 3:
    case 4:
        while (--len >= 0) {
            total += src[0] + src[1] + src[2];
            src += stride;
        }
        div *= 3;
        break;
    }
    total /= div;

    av_log(ctx, AV_LOG_DEBUG, "total:%d\n", total);
    return total;
}

86
static av_cold int init(AVFilterContext *ctx, const char *args)
87 88 89 90 91 92 93 94 95 96 97
{
    CropDetectContext *cd = ctx->priv;

    cd->limit = 24;
    cd->round = 0;
    cd->reset_count = 0;
    cd->frame_nb = -2;

    if (args)
        sscanf(args, "%d:%d:%d", &cd->limit, &cd->round, &cd->reset_count);

98
    av_log(ctx, AV_LOG_VERBOSE, "limit:%d round:%d reset_count:%d\n",
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
           cd->limit, cd->round, cd->reset_count);

    return 0;
}

static int config_input(AVFilterLink *inlink)
{
    AVFilterContext *ctx = inlink->dst;
    CropDetectContext *cd = ctx->priv;

    av_image_fill_max_pixsteps(cd->max_pixsteps, NULL,
                               &av_pix_fmt_descriptors[inlink->format]);

    cd->x1 = inlink->w - 1;
    cd->y1 = inlink->h - 1;
    cd->x2 = 0;
    cd->y2 = 0;

    return 0;
}

120
static int end_frame(AVFilterLink *inlink)
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
{
    AVFilterContext *ctx = inlink->dst;
    CropDetectContext *cd = ctx->priv;
    AVFilterBufferRef *picref = inlink->cur_buf;
    int bpp = cd->max_pixsteps[0];
    int w, h, x, y, shrink_by;

    // ignore first 2 frames - they may be empty
    if (++cd->frame_nb > 0) {
        // Reset the crop area every reset_count frames, if reset_count is > 0
        if (cd->reset_count > 0 && cd->frame_nb > cd->reset_count) {
            cd->x1 = picref->video->w-1;
            cd->y1 = picref->video->h-1;
            cd->x2 = 0;
            cd->y2 = 0;
            cd->frame_nb = 1;
        }

        for (y = 0; y < cd->y1; y++) {
            if (checkline(ctx, picref->data[0] + picref->linesize[0] * y, bpp, picref->video->w, bpp) > cd->limit) {
                cd->y1 = y;
                break;
            }
        }

        for (y = picref->video->h-1; y > cd->y2; y--) {
            if (checkline(ctx, picref->data[0] + picref->linesize[0] * y, bpp, picref->video->w, bpp) > cd->limit) {
                cd->y2 = y;
                break;
            }
        }

        for (y = 0; y < cd->x1; y++) {
            if (checkline(ctx, picref->data[0] + bpp*y, picref->linesize[0], picref->video->h, bpp) > cd->limit) {
                cd->x1 = y;
                break;
            }
        }

        for (y = picref->video->w-1; y > cd->x2; y--) {
            if (checkline(ctx, picref->data[0] + bpp*y, picref->linesize[0], picref->video->h, bpp) > cd->limit) {
                cd->x2 = y;
                break;
            }
        }

        // round x and y (up), important for yuv colorspaces
        // make sure they stay rounded!
        x = (cd->x1+1) & ~1;
        y = (cd->y1+1) & ~1;

        w = cd->x2 - x + 1;
        h = cd->y2 - y + 1;

        // w and h must be divisible by 2 as well because of yuv
        // colorspace problems.
        if (cd->round <= 1)
            cd->round = 16;
        if (cd->round % 2)
            cd->round *= 2;

        shrink_by = w % cd->round;
        w -= shrink_by;
        x += (shrink_by/2 + 1) & ~1;

        shrink_by = h % cd->round;
        h -= shrink_by;
        y += (shrink_by/2 + 1) & ~1;

        av_log(ctx, AV_LOG_INFO,
191 192
               "x1:%d x2:%d y1:%d y2:%d w:%d h:%d x:%d y:%d pos:%"PRId64" pts:%"PRId64" t:%f crop=%d:%d:%d:%d\n",
               cd->x1, cd->x2, cd->y1, cd->y2, w, h, x, y, picref->pos, picref->pts,
193
               picref->pts == AV_NOPTS_VALUE ? -1 : picref->pts * av_q2d(inlink->time_base),
194 195 196
               w, h, x, y);
    }

197
    return ff_end_frame(inlink->dst->outputs[0]);
198 199 200 201 202 203 204 205 206 207 208
}

AVFilter avfilter_vf_cropdetect = {
    .name        = "cropdetect",
    .description = NULL_IF_CONFIG_SMALL("Auto-detect crop size."),

    .priv_size = sizeof(CropDetectContext),
    .init      = init,

    .query_formats = query_formats,

209 210 211 212 213 214 215 216 217 218 219
    .inputs    = (const AVFilterPad[]) {{ .name = "default",
                                          .type             = AVMEDIA_TYPE_VIDEO,
                                          .config_props     = config_input,
                                          .get_video_buffer = ff_null_get_video_buffer,
                                          .start_frame      = ff_null_start_frame,
                                          .end_frame        = end_frame, },
                                        { .name = NULL}},

    .outputs   = (const AVFilterPad[]) {{ .name             = "default",
                                          .type             = AVMEDIA_TYPE_VIDEO },
                                        { .name = NULL}},
220
};