/*
* aptdec - A lightweight FOSS (NOAA) APT decoder
* Copyright (C) 2004-2009 Thierry Leconte (F4DWV) 2019-2023 Xerbo (xerbo@protonmail.com)
*
* This program is free software; you can redistribute it and/or modify
* 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.
*
* This program 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see .
*/
#include
#include
#include
#include
#include
#include
#include "algebra.h"
#include "util.h"
#include "calibration.h"
#define APT_COUNT_RATIO (1023.0f/255.0f)
apt_image_t apt_image_clone(apt_image_t img) {
apt_image_t _img = img;
_img.data = calloc(APT_IMG_WIDTH * img.rows, sizeof(uint8_t));
memcpy(_img.data, img.data, APT_IMG_WIDTH * img.rows);
return _img;
}
static void decode_telemetry(const float *data, size_t rows, size_t offset, float *wedges) {
// Calculate row average
float *telemetry_rows = calloc(rows, sizeof(float));
for (size_t y = 0; y < rows; y++) {
telemetry_rows[y] = meanf(&data[y*APT_IMG_WIDTH + offset + APT_CH_WIDTH], APT_TELEMETRY_WIDTH);
}
// Calculate relative telemetry offset (via step detection, i.e. wedge 8 to 9)
size_t telemetry_offset = 0;
float max_difference = 0.0f;
for (size_t y = APT_WEDGE_HEIGHT; y <= rows - APT_WEDGE_HEIGHT; y++) {
float difference = sumf(&telemetry_rows[y - APT_WEDGE_HEIGHT], APT_WEDGE_HEIGHT) - sumf(&telemetry_rows[y], APT_WEDGE_HEIGHT);
// Find the maximum difference
if (difference > max_difference) {
max_difference = difference;
telemetry_offset = (y + 64) % APT_FRAME_LEN;
}
}
// Find the least noisy frame (via standard deviation)
float best_noise = FLT_MAX;
size_t best_frame = 0;
for (size_t y = telemetry_offset; y < rows-APT_FRAME_LEN; y += APT_FRAME_LEN) {
float noise = 0.0f;
for (size_t i = 0; i < APT_FRAME_WEDGES; i++) {
noise += standard_deviation(&telemetry_rows[y + i*APT_WEDGE_HEIGHT], APT_WEDGE_HEIGHT);
}
if (noise < best_noise) {
best_noise = noise;
best_frame = y;
}
}
for (size_t i = 0; i < APT_FRAME_WEDGES; i++) {
wedges[i] = meanf(&telemetry_rows[best_frame + i*APT_WEDGE_HEIGHT], APT_WEDGE_HEIGHT);
}
free(telemetry_rows);
}
static float average_spc(apt_image_t *img, size_t offset) {
float *rows = calloc(img->rows, sizeof(float));
float average = 0.0f;
for (size_t y = 0; y < img->rows; y++) {
float row_average = 0.0f;
for (size_t x = 0; x < APT_SPC_WIDTH; x++) {
row_average += img->data[y*APT_IMG_WIDTH + offset - APT_SPC_WIDTH + x];
}
row_average /= (float)APT_SPC_WIDTH;
rows[y] = row_average;
average += row_average;
}
average /= (float)img->rows;
float weighted_average = 0.0f;
size_t n = 0;
for (size_t y = 0; y < img->rows; y++) {
if (fabsf(rows[y] - average) < 50.0f) {
weighted_average += rows[y];
n++;
}
}
free(rows);
return weighted_average / (float)n;
}
apt_image_t apt_normalize(const float *data, size_t rows, apt_satellite_t satellite, int *error) {
apt_image_t img;
img.rows = rows;
img.satellite = satellite;
*error = 0;
if (rows < APTDEC_NORMALIZE_ROWS) {
*error = -1;
return img;
}
// Decode and average wedges
float wedges[APT_FRAME_WEDGES];
float wedges_cha[APT_FRAME_WEDGES];
float wedges_chb[APT_FRAME_WEDGES];
decode_telemetry(data, rows, APT_CHA_OFFSET, wedges_cha);
decode_telemetry(data, rows, APT_CHB_OFFSET, wedges_chb);
for (size_t i = 0; i < APT_FRAME_WEDGES; i++) {
wedges[i] = (wedges_cha[i] + wedges_chb[i]) / 2.0f;
}
// Calculate normalization
const float reference[9] = { 31, 63, 95, 127, 159, 191, 223, 255, 0 };
linear_t normalization = linear_regression(wedges, reference, 9);
if (normalization.a < 0.0f) {
*error = -1;
return img;
}
// Normalize telemetry
for (size_t i = 0; i < APT_FRAME_WEDGES; i++) {
img.telemetry[0][i] = linear_calc(wedges_cha[i], normalization);
img.telemetry[1][i] = linear_calc(wedges_chb[i], normalization);
}
// Decode channel ID wedges
img.ch[0] = roundf(img.telemetry[0][15] / 32.0f);
img.ch[1] = roundf(img.telemetry[1][15] / 32.0f);
if (img.ch[0] < 1 || img.ch[0] > 6) img.ch[0] = AVHRR_CHANNEL_UNKNOWN;
if (img.ch[1] < 1 || img.ch[1] > 6) img.ch[1] = AVHRR_CHANNEL_UNKNOWN;
// Normalize and quantize image data
img.data = (uint8_t *)malloc(rows * APT_IMG_WIDTH);
for (size_t i = 0; i < rows * APT_IMG_WIDTH; i++) {
float count = linear_calc(data[i], normalization);
img.data[i] = clamp_int(roundf(count), 0, 255);
}
// Get space brightness
img.space_view[0] = average_spc(&img, APT_CHA_OFFSET);
img.space_view[1] = average_spc(&img, APT_CHB_OFFSET);
return img;
}
static void make_thermal_lut(apt_image_t *img, avhrr_channel_t ch, int satellite, float *lut) {
ch -= 4;
const calibration_t calibration = get_calibration(satellite);
const float Ns = calibration.cor[ch].Ns;
const float Vc = calibration.rad[ch].vc;
const float A = calibration.rad[ch].A;
const float B = calibration.rad[ch].B;
// Compute PRT temperature
float T[4] = { 0.0f };
for (size_t n = 0; n < 4; n++) {
T[n] = quadratic_calc(img->telemetry[1][n + 9] * APT_COUNT_RATIO, calibration.prt[n]);
}
float Tbb = meanf(T, 4); // Blackbody temperature
float Tbbstar = A + Tbb * B; // Effective blackbody temperature
float Nbb = C1 * pow(Vc, 3) / (expf(C2 * Vc / Tbbstar) - 1.0f); // Blackbody radiance
float Cs = img->space_view[1] * APT_COUNT_RATIO;
float Cb = img->telemetry[1][14] * APT_COUNT_RATIO;
for (size_t i = 0; i < 256; i++) {
float Nl = Ns + (Nbb - Ns) * (Cs - i * APT_COUNT_RATIO) / (Cs - Cb); // Linear radiance estimate
float Nc = quadratic_calc(Nl, calibration.cor[ch].quadratic); // Non-linear correction
float Ne = Nl + Nc; // Corrected radiance
float Testar = C2 * Vc / logf(C1 * powf(Vc, 3) / Ne + 1.0); // Equivalent black body temperature
float Te = (Testar - A) / B; // Temperature (kelvin)
// Convert to celsius
lut[i] = Te - 273.15;
}
}
int apt_calibrate_thermal(apt_image_t *img, apt_region_t region) {
if (img->ch[1] != AVHRR_CHANNEL_4 && img->ch[1] != AVHRR_CHANNEL_5 && img->ch[1] != AVHRR_CHANNEL_3B) {
return 1;
}
float lut[256] = { 0.0f };
make_thermal_lut(img, img->ch[1], img->satellite, lut);
for (size_t y = 0; y < img->rows; y++) {
for (size_t x = 0; x < region.width; x++) {
float temperature = lut[img->data[y * APT_IMG_WIDTH + region.offset + x]];
img->data[y * APT_IMG_WIDTH + region.offset + x] = clamp_int(roundf((temperature + 100.0) / 160.0 * 255.0), 0, 255);
}
}
return 0;
}
static float calibrate_pixel_visible(float value, int channel, calibration_t cal) {
if (value > cal.visible[channel].cutoff) {
return linear_calc(value * APT_COUNT_RATIO, cal.visible[channel].high) / 100.0f * 255.0f;
} else {
return linear_calc(value * APT_COUNT_RATIO, cal.visible[channel].low) / 100.0f * 255.0f;
}
}
int apt_calibrate_visible(apt_image_t *img, apt_region_t region) {
if (img->ch[0] != AVHRR_CHANNEL_1 && img->ch[0] != AVHRR_CHANNEL_2) {
return 1;
}
calibration_t calibration = get_calibration(img->satellite);
for (size_t y = 0; y < img->rows; y++) {
for (size_t x = 0; x < region.width; x++) {
float albedo = calibrate_pixel_visible(img->data[y * APT_IMG_WIDTH + region.offset + x], img->ch[0]-1, calibration);
img->data[y * APT_IMG_WIDTH + region.offset + x] = clamp_int(roundf(albedo), 0, 255);
}
}
return 0;
}