Skip to content

I
imagej-elphel
Commit b66e802e authored by Andrey Filippov's avatar Andrey Filippov
Browse files
Options

Started LMA S/N measurements to compare contrast enhancement

parent 4b32667d
Expand all Show whitespace changes
Inline Side-by-side
Showing with 994 additions and 199 deletions
src/main/java/com/elphel/imagej/cameras/ThermalColor.java
View file @ b66e802e
......@@ -112,7 +112,121 @@ public class ThermalColor {
0xfff8c4, 0xfff9c7, 0xfff9ca, 0xfff9cd, 0xfffad1, 0xfffad4, 0xfffbd8, 0xfffcdb,
0xfffcdf, 0xfffde2, 0xfffde5, 0xfffde8, 0xfffeeb, 0xfffeee, 0xfffef1, 0xfffef4,
0xffffff};
int [][] palettes = {white_hot_palette, black_hot_palette, iron_palette};
int [] pal01_palette = {0x1f77b4,0xff7f0e,
0x2ca02c,
0xd62728,
0x9467bd,
0x8c564b,
0xe377c2,
0x7f7f7f,
0xbcbd22,
0x17becf};
int [] viridis_palette = {
0x46317e,
0x365b8c,
0x277e8e,
0x1fa187,
0x49c16d,
0x9fd938};
int [] viridis_mod_palette = {
0x401080,
0x46317e,
0x365b8c,
0x277e8e,
0x1fa187,
0x49c16d,
0x9fd938,
0xcc9000,
0xff6000
};
int [] coolwarm_palette = {
0x6788ed ,
0x99bafe ,
0xc8d7ef ,
0xedd0c1 ,
0xf6a789 ,
0xe16852};
int [] cubehelix_palette = {
0x1a2341 ,
0x1b6144 ,
0x687a30 ,
0xc77a7c ,
0xcda2e0 ,
0xc6e1f1};
int [] hsv_palette = {
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec};
int [] hsv2_palette = {
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec,
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec};
int [] hsv3_palette = {
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec,
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec ,
0xffd400 ,
0x4eff00 ,
0xff85 ,
0x9dff ,
0x3600ff ,
0xff00ec};
int [] spectral_palette = {
0xe1514a ,
0xfba55c ,
0xfee899 ,
0xecf7a2 ,
0xa1d9a4 ,
0x479fb3};
int [] spectral_r_palette = {
0x479fb3 ,
0xa1d9a4 ,
0xecf7a2 ,
0xfee899 ,
0xfba55c ,
0xe1514a};
int [][] palettes = {
white_hot_palette,
black_hot_palette,
iron_palette,
pal01_palette,
viridis_palette,
coolwarm_palette,
cubehelix_palette,
hsv_palette,
hsv2_palette,
hsv3_palette,
viridis_mod_palette,
spectral_palette,
spectral_r_palette
};
if (indx <0) indx = 0;
else if (indx >= palettes.length) indx = palettes.length - 1;
return palettes[indx];
......
src/main/java/com/elphel/imagej/correction/Eyesis_Correction.java
View file @ b66e802e
......@@ -7374,36 +7374,57 @@ private Panel panel1,
IJ.showMessage("Error","No images selected");
return false;
}
double disparity0 = 0.75;
double disparity_max = 75.0; // pix
String sky_mask = "/home/elphel/lwir16-proc/results-cuda/ERS-debug1/models3/1626032208_613623/v01/1626032208_613623-sky_mask.tiff";
double disp_inf = -0.2; // -1.0; // 0.0;
double disparity0 = 0.75; // 2; // 0.75;
double disparity_max = 23; // 75.0; // pix
boolean show_dialog2 = false;
boolean linear_disparity = false;
int palette = 4;
int legend_width = 0;
int legend_gap = 5;
int legend_width = 2; // 0;
int legend_gap = 2; // 5;
GenericJTabbedDialog gd0 = new GenericJTabbedDialog("Ln mode");
gd0.addNumericField("disparity0", disparity0, 5,8,"pix", "Disparity to swict from linear to log. ) - skip log mode");
gd0.addStringField ("File path of the sky mask", sky_mask, 80,"optional file with 0.0 - keep, >0.0 - replace with NaN");
gd0.addNumericField("Disparity at infinity", disp_inf, 5,8,"pix", "Disparity at infinity ");
gd0.addNumericField("disparity0", disparity0, 5,8,"pix", "Disparity to switch from linear to log. ) - skip log mode");
gd0.addNumericField("Maximal disparity", disparity_max, 5,8,"pix", "Fixed maximal disparity (0 - auto");
gd0.addCheckbox("Linear disparity legend", linear_disparity);
gd0.addNumericField("Legend width", legend_width, 0,3,"", "Optional disparity legend vertical bar width");
gd0.addNumericField("Legend gap", legend_gap, 0,3,"", "Optional disparity legend vertical bar gap");
gd0.addNumericField("palette", palette, 0,3,"", "Palette index");
gd0.addCheckbox("Show second dialog", show_dialog2);
gd0.showDialog();
if (gd0.wasCanceled()) return false;
sky_mask = gd0.getNextString();
disp_inf = gd0.getNextNumber();
disparity0= gd0.getNextNumber();
disparity_max= gd0.getNextNumber();
linear_disparity = gd0.getNextBoolean();
legend_width= (int) gd0.getNextNumber();
legend_gap= (int) gd0.getNextNumber();
palette= (int) gd0.getNextNumber();
show_dialog2 = gd0.getNextBoolean();
boolean log_mode = disparity0 > 0.0;
boolean log_mode = disparity0 > 0.00;
float [] fsky_mask = null;
if (sky_mask.length() > 0) {
ImagePlus imp_sky_mask = new ImagePlus(sky_mask);
fsky_mask = (float[]) imp_sky_mask.getProcessor().getPixels();
}
// float [] pixels=(float []) imp_srcs[srcChannel].getProcessor().getPixels();
int current_slice = imp_sel.getCurrentSlice();
ImageStack imageStack = imp_sel.getStack();
float [] fpixels = (float[]) imageStack.getPixels(current_slice);
float [] fpixels0 = (float[]) imageStack.getPixels(current_slice);
float [] fpixels = fpixels0.clone();
for (int i = 0; i <fpixels.length; i++) {
fpixels[i] -= (float) disp_inf;
if ((fsky_mask != null) && (fsky_mask[i] > 0.0f)) {
fpixels[i] = Float.NaN;
}
}
int width = imp_sel.getWidth();
int height = imp_sel.getHeight();
String title = imp_sel.getShortTitle(); // getTitle();
......@@ -7441,11 +7462,10 @@ private Panel panel1,
}
}
}
double mn = fpixels[0];
double mn = dpixels[0];
double mx = mn;
double pwr = 1.0;
int palette = 2;
for (int i = 0; i < fpixels.length; i++) {
for (int i = 0; i < dpixels.length; i++) {
double d = dpixels[i];
if (log_mode) {
if (!Double.isNaN(d)) {
......@@ -7458,7 +7478,8 @@ private Panel panel1,
}
}
}
if (!Double.isNaN(d)) {
int px = i % width;
if ((px < width0) && !Double.isNaN(d)) { // do not use legend
if (!(d <= mx)) mx = d;
if (!(d >= mn)) mn = d;
}
......@@ -7483,13 +7504,11 @@ private Panel panel1,
gd.addNumericField("min", mn, 5,8,"", "Minimal value to map");
gd.addNumericField("max", mx, 5,8,"", "Maximal value to map");
gd.addNumericField("pwr", pwr, 5,8,"", "Exponent power");
gd.addNumericField("palette", palette, 0,3,"", "Palette index");
gd.showDialog();
if (gd.wasCanceled()) return false;
mn= gd.getNextNumber();
mx= gd.getNextNumber();
pwr= gd.getNextNumber();
palette= (int)gd.getNextNumber();
}
if (pwr != 1.0) {
if (mn < 0) {
......@@ -7516,6 +7535,8 @@ private Panel panel1,
}
}
System.out.println("mn="+mn+", mx="+mx);
// (new ShowDoubleFloatArrays()) .showArrays(dpixels, width, height, "test_depth_mn="+mn+"_mx="+mx);
double [][] pseudo_pixels = new double [3] [dpixels.length];
ThermalColor tc = new ThermalColor(
......
src/main/java/com/elphel/imagej/tileprocessor/Correlation2d.java
View file @ b66e802e
......@@ -209,6 +209,12 @@ public class Correlation2d {
private static int SUB_SAMPLE = 16; // subsample source pixel in each direction when generating
public static int THREADS_MAX = 100;
public int getPairLength(int pair) {
return pair_length[pair];
}
public int [] getPairLengths() {
return pair_length;
}
// All used pairs (but the diameters) are clockwise (end is clockwise of start)
// Orientation calculations are valid for clockwise only
......@@ -429,7 +435,7 @@ public class Correlation2d {
* Add 2D correlation for a pair to the combined correlation tile, applying rotation/scaling
* @param accum_tile tile for accumulation in line-scan order, same dimension as during generateResample()
* @param corr_tile correlation tile (currently 15x15)
* @param num_pair number of correlation pair for which resamplind data exists
* @param num_pair number of correlation pair for which resampling data exists
* @param weight multiply added tile data by this coefficient before accumulation
*/
public void accummulatePair(
......@@ -472,6 +478,26 @@ public class Correlation2d {
return sumw;
}
public double accummulatePairs(
double [] accum_tile,
double [][] corr_tiles,
double [] weights) {
double sumw = 0.0;
for (int num_pair = 0; num_pair < weights.length; num_pair++) {
if ((weights[num_pair]>0.0) && (corr_tiles[num_pair] != null)) {
accummulatePair(
accum_tile,
corr_tiles[num_pair],
num_pair,
weights[num_pair]);
sumw+=weights[num_pair];
}
}
return sumw;
}
public void normalizeAccumulatedPairs(
double [] accum_tile,
double sumw) {
......@@ -483,6 +509,36 @@ public class Correlation2d {
}
}
/**
* Calculate correlation pair width in the disparity direction to boost weights of the most
* informative (for disparity) pairs
* @param corr_tile for selected pair
* @param num_pair selected pair number
* @return pair width (in pixels) in the disparity direction after applying rotation and scaling
*/
public double getPairWidth(
double [] corr_tile,
int num_pair){ // Center line only
if ((resample_indices == null) || (resample_indices[num_pair] == null)) {
throw new IllegalArgumentException ("getPairWidth(): No resample data for num_pair = "+num_pair);
}
double s0 = 0.0, s1 = 0.0, s2 = 0.0;
int indx = mcorr_comb_width * (mcorr_comb_offset + ((mcorr_comb_height - 1) / 2));
for (int ix = 0; ix < mcorr_comb_width; ix ++ ) if (resample_indices[num_pair][indx] != null){
for (int j = 0; j < resample_indices[num_pair][indx].length; j++) {
double d = corr_tile[resample_indices[num_pair][indx][j]] * resample_weights[num_pair][indx][j];
if (d < 0) {
d = 0.0; // was get5ting negative (s0*s2- s1*s1) !
}
s0 += d;
s1 += d * ix;
s2 += d * ix * ix;
}
indx++;
}
return Math.sqrt(s0*s2- s1*s1)/s0;
}
public double [] accumulateInit() {return new double [mcorr_comb_width * mcorr_comb_height]; }
public int getCombWidth() {return mcorr_comb_width;}
......
src/main/java/com/elphel/imagej/tileprocessor/ErsCorrection.java
View file @ b66e802e
......@@ -917,7 +917,7 @@ public class ErsCorrection extends GeometryCorrection {
* @param camera_xyz camera lens position during centerline acquisition in world coordinates, null - use instance global
* @param camera_atr camera orientation during centerline acquisition in world frame, null - use instance global
* @param line_err iterate until the line (pY) correction is below this value
* @return {px, py, disparity } (right, down) or null if behind the camera
* @return {px, py, disparity } (right, down) or null if behind the camera of the other camera
*/
public double [] getImageCoordinatesERS(
......@@ -1064,7 +1064,8 @@ public class ErsCorrection extends GeometryCorrection {
* @param line_err iterate until the line (pY) correction is below this value
* @return {px, py, disparity } (right, down)
*/
public double [] getImageCoordinatesERS( // USED in lwir
@Deprecated
public double [] getImageCoordinatesERS( // not used?
double [] xyzw,
boolean correctDistortions, // correct distortion (will need corrected background too !)
double line_err) // threshold error in scan lines (1.0)
......
src/main/java/com/elphel/imagej/tileprocessor/ImageDttCPU.java
View file @ b66e802e
......@@ -16193,7 +16193,11 @@ public class ImageDttCPU {
}
}
}
boolean debugTile0 =(tileX == debug_tileX) && (tileY == debug_tileY) && (globalDebugLevel > -3);
boolean debugTile0 =(tileX == debug_tileX) && (tileY == debug_tileY) && (globalDebugLevel > -6);
if (debugTile0) {
System.out.println("clt_process_tl_correlations(): tileX="+tileX+", tileY="+tileY);
}
// TODO: move port coordinates out of color channel loop
// double [][] centersXY = tp_tasks[iTile].getDoubleXY(); // isAux());
double [][] disp_dist = tp_tasks[iTile].getDoubleDispDist();
......@@ -16225,11 +16229,55 @@ public class ImageDttCPU {
double [] disp_str = {0.0, 0.0}; // diaprity = 0 will be initial approximation for LMA if no averaging
if (combine_corrs) {
double [] corr_combo_tile = correlation2d.accumulateInit(); // combine all available pairs
double sumw = correlation2d.accummulatePairs(
double sumw = 0.0;
if (imgdtt_params.mcorr_static_weights || imgdtt_params.mcorr_dynamic_weights) {
double [] weights = new double [correlation2d.getNumPairs()];
if (imgdtt_params.mcorr_static_weights) {
int [] pair_lengths = correlation2d.getPairLengths();
for (int npair = 0; npair< weights.length; npair++) if (corrs[npair] != null) {
weights[npair] = imgdtt_params.mcorr_weights[pair_lengths[npair] - 1];
}
} else {
for (int npair = 0; npair< weights.length; npair++) if (corrs[npair] != null) {
weights[npair] = 1.0;
}
// Arrays.fill(weights, 1.0);
}
if (imgdtt_params.mcorr_dynamic_weights) {
for (int npair = 0; npair< weights.length; npair++) if (corrs[npair] != null) {
double pair_width = correlation2d. getPairWidth(
corrs[npair] , //double [] corr_tile,
npair); // int num_pair)
if (pair_width < 0.1) {
System.out.println("pair_width["+npair+"]="+pair_width);
} else {
weights[npair] /= Math.pow(pair_width, imgdtt_params.mcorr_weights_power);
}
}
}
if (debugTile0) {
System.out.println("clt_process_tl_correlations(): per-pair weights:");
for (int npair = 0; npair< weights.length; npair++) if (weights[npair] > 0.0) {
int [] pair = correlation2d.getPair(npair);
int pair_length = correlation2d.getPairLength(npair);
System.out.println(String.format("%3d: %2d -> %2d [%2d]: %8.5f",npair,pair[0],pair[1],pair_length,weights[npair]));
}
}
sumw = correlation2d.accummulatePairs(
corr_combo_tile, // double [] accum_tile,
corrs, // double [][] corr_tiles, may be longer than selection, extra will be ignored
weights); // double [] weights);
} else { // old way, same weight for all pairs
sumw = correlation2d.accummulatePairs(
corr_combo_tile, // double [] accum_tile,
corrs, // double [][] corr_tiles, may be longer than selection, extra will be ignored
correlation2d.selectAll(), // boolean [] selection,
1.0); // double weight);
}
correlation2d.normalizeAccumulatedPairs(
corr_combo_tile,
sumw);
......
src/main/java/com/elphel/imagej/tileprocessor/ImageDttParameters.java
View file @ b66e802e
......@@ -116,9 +116,10 @@ public class ImageDttParameters {
public boolean mcorr_cons_hor = true; // consolidate all horizontal pairs
public boolean mcorr_cons_vert = true; // consolidate all vertical pairs
public double [] mcorr_weights = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 16.0}; // static weights of pairs of length 1...8
// public double [] mcorr_weights = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 16.0}; // static weights of pairs of length 1...8
public double [] mcorr_weights = {1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0}; // static weights of pairs of length 1...8
public boolean mcorr_static_weights = true; // when mixing , apply static weights to pairs depending on their lengths
public double mcorr_weights_power = 1.0; // divide pair by horizontal (disparity) width after rotation/scaling (skip negative when measuring width)
public double mcorr_weights_power = 2.0; // divide pair by horizontal (disparity) width after rotation/scaling (skip negative when measuring width)
public boolean mcorr_dynamic_weights = true; // Apply weights to pairs dependent on the width in disparity direction
/// these are just for testing, actual will be generated specifically for different applications (LY will use closest pairs)
......
src/main/java/com/elphel/imagej/tileprocessor/InterIntraLMA.java 0 → 100644
View file @ b66e802e
package com.elphel.imagej.tileprocessor;
import java.util.Arrays;
/**
**
** InterIntraLMA - Comparing contrast for different interscene and intrascene
** connfifurations
**
** Copyright (C) 2021 Elphel, Inc.
**
** -----------------------------------------------------------------------------**
**
** InterIntraLMA.java 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 3 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 <http://www.gnu.org/licenses/>.
** -----------------------------------------------------------------------------**
**
*/
public class InterIntraLMA {
/**
* Estimate noise threshold for each combination of inter/intra and number of
* sensors.
* @param noise_file synthetic noise level per file (either rnd or fpn)
* @param senosor_mode_file sensor mode (0 - 16, 1 - 8, 2 - 4, 3 - 2 sensors) per file
* @param inter_file true - interscene , false - intrascene only
* @param good_file_tile true if tile of this file is considered "good"
* @param min_modes minimal number of modes the tile should be defined for (undefine if less than)
* @return per tile per mode average between highest noise of the good tile and lowest noise
* of the bad tile. Double.NaN if noise threshold can not be determined. null if the tile is
* undefined for all modes
*/
public static double [][] getNoiseThreshold(
double [] noise_file, // = new double [noise_files.length];
int [] sensor_mode_file,
boolean [] inter_file,
boolean [][] good_file_tile,
int min_modes)
{
int num_sensor_modes = 0;
int num_tiles = good_file_tile[0].length;
for (int i = 0; i < sensor_mode_file.length; i++) {
if (sensor_mode_file[i] > num_sensor_modes) {
num_sensor_modes = sensor_mode_file[i];
}
}
num_sensor_modes ++;
int num_modes = 2 * num_sensor_modes;
double [][] rslt = new double [num_tiles][]; // number of tiles
double [][][] noise_interval = new double[num_modes][num_tiles][2]; // [modes][tiles]
for (int i = 0; i < num_modes; i++) {
for (int j= 0; j < num_tiles; j++) {
noise_interval[i][j][0] =Double.NaN;
noise_interval[i][j][1] =Double.NaN;
}
}
for (int nf = 0; nf < noise_file.length; nf++) {
double noise = noise_file[nf];
int mode = sensor_mode_file[nf] + (inter_file[nf] ? 0: num_sensor_modes);
for (int ntile = 0; ntile < num_tiles; ntile++) {
if (good_file_tile[nf][ntile]) { // good tile
if (!(noise <= noise_interval[mode][ntile][0])){ // including Double.isNaN(noise_interval[mode][ntile][0]
noise_interval[mode][ntile][0] = noise;
}
} else { // bad tile
if (!(noise >= noise_interval[mode][ntile][1])){ // including Double.isNaN(noise_interval[mode][ntile][1]
noise_interval[mode][ntile][1] = noise;
}
}
}
}
for (int ntile = 0; ntile < num_tiles; ntile++){
int num_defined = 0;
for (int mode = 0; mode < num_modes; mode++) {
if (!Double.isNaN(noise_interval[mode][ntile][0]) && !Double.isNaN(noise_interval[mode][ntile][0])) {
num_defined++;
}
}
if (num_defined >= min_modes) {
rslt[ntile] = new double [num_modes];
for (int mode = 0; mode < num_modes; mode++) {
if (!Double.isNaN(noise_interval[mode][ntile][0]) && !Double.isNaN(noise_interval[mode][ntile][0])) {
rslt[ntile][mode] = 0.5 * (noise_interval[mode][ntile][0] + noise_interval[mode][ntile][1]);
} else {
rslt[ntile][mode] = Double.NaN;
}
}
}
}
return rslt;
}
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment