Add FITS Decoder

Signed-off-by: Paras Chadha <paraschadha18@gmail.com>

1 files changed, 318 insertions, 0 deletions

diff --git a/libavcodec/fitsdec.c b/libavcodec/fitsdec.c
new file mode 100644
index 0000000000..8ede0bcbfa
--- /dev/null
+++ b/libavcodec/fitsdec.c
@@ -0,0 +1,318 @@
+/*
+ * FITS image decoder
+ * Copyright (c) 2017 Paras Chadha
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * FITS image decoder
+ *
+ * Specification: https://fits.gsfc.nasa.gov/fits_standard.html Version 3.0
+ *
+ * Support all 2d images alongwith, bzero, bscale and blank keywords.
+ * RGBA images are supported as NAXIS3 = 3 or 4 i.e. Planes in RGBA order. Also CTYPE = 'RGB ' should be present.
+ * Also to interpret data, values are linearly scaled using min-max scaling but not RGB images.
+ */
+
+#include "avcodec.h"
+#include "internal.h"
+#include <float.h>
+#include "libavutil/intreadwrite.h"
+#include "libavutil/intfloat.h"
+#include "libavutil/dict.h"
+#include "libavutil/opt.h"
+#include "fits.h"
+
+typedef struct FITSContext {
+    const AVClass *class;
+    int blank_val;
+} FITSContext;
+
+/**
+ * Calculate the data_min and data_max values from the data.
+ * This is called if the values are not present in the header.
+ * @param ptr8 pointer to the data
+ * @param header pointer to the header
+ * @param end pointer to end of packet
+ * @return 0 if calculated successfully otherwise AVERROR_INVALIDDATA
+ */
+static int fill_data_min_max(const uint8_t *ptr8, FITSHeader *header, const uint8_t *end)
+{
+    uint8_t t8;
+    int16_t t16;
+    int32_t t32;
+    int64_t t64;
+    float tflt;
+    double tdbl;
+    int i, j;
+
+    header->data_min = DBL_MAX;
+    header->data_max = DBL_MIN;
+    switch (header->bitpix) {
+#define CASE_N(a, t, rd) \
+    case a: \
+        for (i = 0; i < header->naxisn[1]; i++) { \
+            for (j = 0; j < header->naxisn[0]; j++) { \
+                t = rd; \
+                if (!header->blank_found || t != header->blank) { \
+                    if (t > header->data_max) \
+                        header->data_max = t; \
+                    if (t < header->data_min) \
+                        header->data_min = t; \
+                } \
+                ptr8 += abs(a) >> 3; \
+            } \
+        } \
+        break
+
+        CASE_N(-64, tdbl, av_int2double(AV_RB64(ptr8)));
+        CASE_N(-32, tflt, av_int2float(AV_RB32(ptr8)));
+        CASE_N(8, t8, ptr8[0]);
+        CASE_N(16, t16, AV_RB16(ptr8));
+        CASE_N(32, t32, AV_RB32(ptr8));
+        CASE_N(64, t64, AV_RB64(ptr8));
+        default:
+            return AVERROR_INVALIDDATA;
+    }
+    return 0;
+}
+
+/**
+ * Read the fits header and store the values in FITSHeader pointed by header
+ * @param avctx AVCodec context
+ * @param ptr pointer to pointer to the data
+ * @param header pointer to the FITSHeader
+ * @param end pointer to end of packet
+ * @param metadata pointer to pointer to AVDictionary to store metadata
+ * @return 0 if calculated successfully otherwise AVERROR_INVALIDDATA
+ */
+static int fits_read_header(AVCodecContext *avctx, const uint8_t **ptr, FITSHeader *header,
+                            const uint8_t *end, AVDictionary **metadata)
+{
+    const uint8_t *ptr8 = *ptr;
+    int lines_read, bytes_left, i, ret;
+    size_t size;
+
+    lines_read = 1; // to account for first header line, SIMPLE or XTENSION which is not included in packet...
+    avpriv_fits_header_init(header, STATE_BITPIX);
+    do {
+        if (end - ptr8 < 80)
+            return AVERROR_INVALIDDATA;
+        ret = avpriv_fits_header_parse_line(avctx, header, ptr8, &metadata);
+        ptr8 += 80;
+        lines_read++;
+    } while (!ret);
+    if (ret < 0)
+        return ret;
+
+    bytes_left = (((lines_read + 35) / 36) * 36 - lines_read) * 80;
+    if (end - ptr8 < bytes_left)
+        return AVERROR_INVALIDDATA;
+    ptr8 += bytes_left;
+
+    if (header->rgb && (header->naxis != 3 || (header->naxisn[2] != 3 && header->naxisn[2] != 4))) {
+        av_log(avctx, AV_LOG_ERROR, "File contains RGB image but NAXIS = %d and NAXIS3 = %d\n", header->naxis, header->naxisn[2]);
+        return AVERROR_INVALIDDATA;
+    }
+
+    if (!header->rgb && header->naxis != 2) {
+        av_log(avctx, AV_LOG_ERROR, "unsupported number of dimensions, NAXIS = %d\n", header->naxis);
+        return AVERROR_INVALIDDATA;
+    }
+
+    if (header->blank_found && (header->bitpix == -32 || header->bitpix == -64)) {
+        av_log(avctx, AV_LOG_WARNING, "BLANK keyword found but BITPIX = %d\n. Ignoring BLANK", header->bitpix);
+        header->blank_found = 0;
+    }
+
+    size = abs(header->bitpix) >> 3;
+    for (i = 0; i < header->naxis; i++) {
+        if (header->naxisn[i] > SIZE_MAX / size) {
+            av_log(avctx, AV_LOG_ERROR, "unsupported size of FITS image");
+            return AVERROR_INVALIDDATA;
+        }
+        size *= header->naxisn[i];
+    }
+
+    if (end - ptr8 < size)
+        return AVERROR_INVALIDDATA;
+    *ptr = ptr8;
+
+    if (!header->rgb && (!header->data_min_found || !header->data_max_found)) {
+        ret = fill_data_min_max(ptr8, header, end);
+        if (ret < 0) {
+            av_log(avctx, AV_LOG_ERROR, "invalid BITPIX, %d\n", header->bitpix);
+            return ret;
+        }
+    } else {
+        /*
+         * instead of applying bscale and bzero to every element,
+         * we can do inverse transformation on data_min and data_max
+         */
+        header->data_min = (header->data_min - header->bzero) / header->bscale;
+        header->data_max = (header->data_max - header->bzero) / header->bscale;
+    }
+
+    return 0;
+}
+
+static int fits_decode_frame(AVCodecContext *avctx, void *data, int *got_frame, AVPacket *avpkt)
+{
+    AVFrame *p=data;
+    const uint8_t *ptr8 = avpkt->data, *end;
+    uint8_t t8;
+    int16_t t16;
+    int32_t t32;
+    int64_t t64;
+    float   tflt;
+    double  tdbl;
+    int ret, i, j, k;
+    const int map[] = {2, 0, 1, 3}; // mapping from GBRA -> RGBA as RGBA is to be stored in FITS file..
+    uint8_t *dst8;
+    uint16_t *dst16;
+    uint64_t t;
+    FITSHeader header;
+    FITSContext * fitsctx = avctx->priv_data;
+
+    end = ptr8 + avpkt->size;
+    p->metadata = NULL;
+    ret = fits_read_header(avctx, &ptr8, &header, end, &p->metadata);
+    if (ret < 0)
+        return ret;
+
+    if (header.rgb) {
+        if (header.bitpix == 8) {
+            if (header.naxisn[2] == 3) {
+                avctx->pix_fmt = AV_PIX_FMT_GBRP;
+            } else {
+                avctx->pix_fmt = AV_PIX_FMT_GBRAP;
+            }
+        } else if (header.bitpix == 16) {
+            if (header.naxisn[2] == 3) {
+                avctx->pix_fmt = AV_PIX_FMT_GBRP16;
+            } else {
+                avctx->pix_fmt = AV_PIX_FMT_GBRAP16;
+            }
+        } else {
+            av_log(avctx, AV_LOG_ERROR, "unsupported BITPIX = %d\n", header.bitpix);
+            return AVERROR_INVALIDDATA;
+        }
+    } else {
+        if (header.bitpix == 8) {
+            avctx->pix_fmt = AV_PIX_FMT_GRAY8;
+        } else {
+            avctx->pix_fmt = AV_PIX_FMT_GRAY16;
+        }
+    }
+
+    if ((ret = ff_set_dimensions(avctx, header.naxisn[0], header.naxisn[1])) < 0)
+        return ret;
+
+    if ((ret = ff_get_buffer(avctx, p, 0)) < 0)
+        return ret;
+
+    /*
+     * FITS stores images with bottom row first. Therefore we have
+     * to fill the image from bottom to top.
+     */
+    if (header.rgb) {
+        switch(header.bitpix) {
+#define CASE_RGB(cas, dst, type, dref) \
+    case cas: \
+        for (k = 0; k < header.naxisn[2]; k++) { \
+            for (i = 0; i < avctx->height; i++) { \
+                dst = (type *) (p->data[map[k]] + (avctx->height - i - 1) * p->linesize[map[k]]); \
+                for (j = 0; j < avctx->width; j++) { \
+                    t32 = dref(ptr8); \
+                    if (!header.blank_found || t32 != header.blank) { \
+                        t = t32 * header.bscale + header.bzero; \
+                    } else { \
+                        t = fitsctx->blank_val; \
+                    } \
+                    *dst++ = (type) t; \
+                    ptr8 += cas >> 3; \
+                } \
+            } \
+        } \
+        break
+
+            CASE_RGB(8, dst8, uint8_t, *);
+            CASE_RGB(16, dst16, uint16_t, AV_RB16);
+        }
+    } else {
+        switch (header.bitpix) {
+#define CASE_GRAY(cas, dst, type, t, rd) \
+    case cas: \
+        for (i = 0; i < avctx->height; i++) { \
+            dst = (type *) (p->data[0] + (avctx->height-i-1)* p->linesize[0]); \
+            for (j = 0; j < avctx->width; j++) { \
+                t = rd; \
+                if (!header.blank_found || t != header.blank) { \
+                    t = ((t - header.data_min) * ((1 << (sizeof(type) * 8)) - 1)) / (header.data_max - header.data_min); \
+                } else { \
+                    t = fitsctx->blank_val; \
+                } \
+                *dst++ = (type) t; \
+                ptr8 += abs(cas) >> 3; \
+            } \
+        } \
+        break
+
+            CASE_GRAY(-64, dst16, uint16_t, tdbl, av_int2double(AV_RB64(ptr8)));
+            CASE_GRAY(-32, dst16, uint16_t, tflt, av_int2float(AV_RB32(ptr8)));
+            CASE_GRAY(8, dst8, uint8_t, t8, ptr8[0]);
+            CASE_GRAY(16, dst16, uint16_t, t16, AV_RB16(ptr8));
+            CASE_GRAY(32, dst16, uint16_t, t32, AV_RB32(ptr8));
+            CASE_GRAY(64, dst16, uint16_t, t64, AV_RB64(ptr8));
+            default:
+                av_log(avctx, AV_LOG_ERROR, "invalid BITPIX, %d\n", header.bitpix);
+                return AVERROR_INVALIDDATA;
+        }
+    }
+
+    p->key_frame = 1;
+    p->pict_type = AV_PICTURE_TYPE_I;
+
+    *got_frame = 1;
+
+    return avpkt->size;
+}
+
+static const AVOption fits_options[] = {
+    { "blank_value", "value that is used to replace BLANK pixels in data array", offsetof(FITSContext, blank_val), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 65535, AV_OPT_FLAG_DECODING_PARAM | AV_OPT_FLAG_VIDEO_PARAM},
+    { NULL },
+};
+
+static const AVClass fits_decoder_class = {
+    .class_name = "FITS decoder",
+    .item_name  = av_default_item_name,
+    .option     = fits_options,
+    .version    = LIBAVUTIL_VERSION_INT,
+};
+
+AVCodec ff_fits_decoder = {
+    .name           = "fits",
+    .type           = AVMEDIA_TYPE_VIDEO,
+    .id             = AV_CODEC_ID_FITS,
+    .priv_data_size = sizeof(FITSContext),
+    .decode         = fits_decode_frame,
+    .capabilities   = AV_CODEC_CAP_DR1,
+    .long_name      = NULL_IF_CONFIG_SMALL("Flexible Image Transport System"),
+    .priv_class     = &fits_decoder_class
+};