Skip to content

I
imagej-elphel
Commit 73b9d03b authored by Andrey Filippov's avatar Andrey Filippov
Browse files
Options

redesigning shell extraction

parent c5a866f4
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 1180 additions and 301 deletions
src/main/java/EyesisCorrectionParameters.java
View file @ 73b9d03b
......@@ -2155,6 +2155,13 @@ public class EyesisCorrectionParameters {
public boolean plPreferDisparity = false;// Always start with disparity-most axis (false - lowest eigenvalue)
public double plDispNorm = 3.0; // Normalize disparities to the average if above (now only for eigenvalue comparison)
public double plBlurBinVert = 1.2; // Blur disparity histograms for constant disparity clusters by this sigma (in bins)
public double plBlurBinHor = 0.8; // Blur disparity histograms for horizontal clusters by this sigma (in bins)
public double plMaxDiffVert = 0.4; // Maximal normalized disparity difference when initially assigning to vertical plane
public double plMaxDiffHor = 0.2; // Maximal normalized disparity difference when initially assigning to horizontal plane
public int plInitPasses = 3; // Number of initial passes to assign tiles to vert (const disparity) and hor planes
public int plMinPoints = 5; // Minimal number of points for plane detection
public double plTargetEigen = 0.1; // Remove outliers until main axis eigenvalue (possibly scaled by plDispNorm) gets below
public double plFractOutliers = 0.3; // Maximal fraction of outliers to remove
......@@ -2453,7 +2460,14 @@ public class EyesisCorrectionParameters {
properties.setProperty(prefix+"plPreferDisparity",this.plPreferDisparity+"");
properties.setProperty(prefix+"plDispNorm", this.plDispNorm +"");
properties.setProperty(prefix+"plMinPoints", this.plMinPoints+"");
properties.setProperty(prefix+"plBlurBinVert", this.plBlurBinVert +"");
properties.setProperty(prefix+"plBlurBinHor", this.plBlurBinHor +"");
properties.setProperty(prefix+"plMaxDiffVert", this.plMaxDiffVert +"");
properties.setProperty(prefix+"plMaxDiffHor", this.plMaxDiffHor +"");
properties.setProperty(prefix+"plInitPasses", this.plInitPasses+"");
properties.setProperty(prefix+"plMinPoints", this.plMinPoints+"");
properties.setProperty(prefix+"plTargetEigen", this.plTargetEigen +"");
properties.setProperty(prefix+"plFractOutliers", this.plFractOutliers +"");
properties.setProperty(prefix+"plMaxOutliers", this.plMaxOutliers+"");
......@@ -2463,10 +2477,8 @@ public class EyesisCorrectionParameters {
properties.setProperty(prefix+"plWorstWorsening", this.plWorstWorsening +"");
properties.setProperty(prefix+"plWeakWorsening", this.plWeakWorsening +"");
properties.setProperty(prefix+"plMutualOnly", this.plMutualOnly+"");
properties.setProperty(prefix+"plFillSquares", this.plFillSquares+"");
properties.setProperty(prefix+"plCutCorners", this.plCutCorners+"");
properties.setProperty(prefix+"plPull", this.plPull +"");
properties.setProperty(prefix+"plNormPow", this.plNormPow +"");
properties.setProperty(prefix+"plIterations", this.plIterations+"");
......@@ -2738,6 +2750,13 @@ public class EyesisCorrectionParameters {
if (properties.getProperty(prefix+"plPreferDisparity")!=null) this.plPreferDisparity=Boolean.parseBoolean(properties.getProperty(prefix+"plPreferDisparity"));
if (properties.getProperty(prefix+"plDispNorm")!=null) this.plDispNorm=Double.parseDouble(properties.getProperty(prefix+"plDispNorm"));
if (properties.getProperty(prefix+"plBlurBinVert")!=null) this.plBlurBinVert=Double.parseDouble(properties.getProperty(prefix+"plBlurBinVert"));
if (properties.getProperty(prefix+"plBlurBinHor")!=null) this.plBlurBinHor=Double.parseDouble(properties.getProperty(prefix+"plBlurBinHor"));
if (properties.getProperty(prefix+"plMaxDiffVert")!=null) this.plMaxDiffVert=Double.parseDouble(properties.getProperty(prefix+"plMaxDiffVert"));
if (properties.getProperty(prefix+"plMaxDiffHor")!=null) this.plMaxDiffHor=Double.parseDouble(properties.getProperty(prefix+"plMaxDiffHor"));
if (properties.getProperty(prefix+"plInitPasses")!=null) this.plInitPasses=Integer.parseInt(properties.getProperty(prefix+"plInitPasses"));
if (properties.getProperty(prefix+"plMinPoints")!=null) this.plMinPoints=Integer.parseInt(properties.getProperty(prefix+"plMinPoints"));
if (properties.getProperty(prefix+"plTargetEigen")!=null) this.plTargetEigen=Double.parseDouble(properties.getProperty(prefix+"plTargetEigen"));
if (properties.getProperty(prefix+"plFractOutliers")!=null) this.plFractOutliers=Double.parseDouble(properties.getProperty(prefix+"plFractOutliers"));
......@@ -2746,7 +2765,7 @@ public class EyesisCorrectionParameters {
if (properties.getProperty(prefix+"plMaxEigen")!=null) this.plMaxEigen=Double.parseDouble(properties.getProperty(prefix+"plMaxEigen"));
if (properties.getProperty(prefix+"plDbgMerge")!=null) this.plDbgMerge=Boolean.parseBoolean(properties.getProperty(prefix+"plDbgMerge"));
if (properties.getProperty(prefix+"plWorstWorsening")!=null) this.plWorstWorsening=Double.parseDouble(properties.getProperty(prefix+"plWorstWorsening"));
if (properties.getProperty(prefix+"plWeakWorsening")!=null) this.plWeakWorsening=Double.parseDouble(properties.getProperty(prefix+"plWeakWorsening"));
if (properties.getProperty(prefix+"plWeakWorsening")!=null) this.plWeakWorsening=Double.parseDouble(properties.getProperty(prefix+"plWeakWorsening"));
if (properties.getProperty(prefix+"plMutualOnly")!=null) this.plMutualOnly=Boolean.parseBoolean(properties.getProperty(prefix+"plMutualOnly"));
if (properties.getProperty(prefix+"plFillSquares")!=null) this.plFillSquares=Boolean.parseBoolean(properties.getProperty(prefix+"plFillSquares"));
......@@ -3047,7 +3066,14 @@ public class EyesisCorrectionParameters {
gd.addMessage ("--- Planes detection ---");
gd.addCheckbox ("Always start with disparity-most axis (false - lowest eigenvalue)", this.plPreferDisparity);
gd.addNumericField("Normalize disparities to the average if above", this.plDispNorm, 6);
gd.addNumericField("Minimal number of points for plane detection", this.plMinPoints, 0);
gd.addNumericField("Blur disparity histograms for constant disparity clusters by this sigma (in bins)", this.plBlurBinVert, 6);
gd.addNumericField("Blur disparity histograms for horizontal clusters by this sigma (in bins)", this.plBlurBinHor, 6);
gd.addNumericField("Maximal normalized disparity difference when initially assigning to vertical plane", this.plMaxDiffVert, 6);
gd.addNumericField("Maximal normalized disparity difference when initially assigning to horizontal plane",this.plMaxDiffHor, 6);
gd.addNumericField("Number of initial passes to assign tiles to vert (const disparity) and hor planes", this.plInitPasses, 0);
gd.addNumericField("Minimal number of points for plane detection", this.plMinPoints, 0);
gd.addNumericField("Remove outliers until main axis eigenvalue (possibly scaled by plDispNorm) gets below", this.plTargetEigen, 6);
gd.addNumericField("Maximal fraction of outliers to remove", this.plFractOutliers, 6);
gd.addNumericField("Maximal number of outliers to remove", this.plMaxOutliers, 0);
......@@ -3340,6 +3366,13 @@ public class EyesisCorrectionParameters {
this.plPreferDisparity= gd.getNextBoolean();
this.plDispNorm= gd.getNextNumber();
this.plBlurBinVert= gd.getNextNumber();
this.plBlurBinHor= gd.getNextNumber();
this.plMaxDiffVert= gd.getNextNumber();
this.plMaxDiffHor= gd.getNextNumber();
this.plInitPasses= (int) gd.getNextNumber();
this.plMinPoints= (int) gd.getNextNumber();
this.plTargetEigen= gd.getNextNumber();
this.plFractOutliers= gd.getNextNumber();
......
src/main/java/SuperTiles.java
View file @ 73b9d03b
This diff is collapsed.
src/main/java/TilePlanes.java
View file @ 73b9d03b
......@@ -165,6 +165,10 @@ public class TilePlanes {
}
return this.measuredSelection[nl];
}
public boolean [][] getMeasSelection(){
return this.measuredSelection;
}
public MeasuredLayers getMeasuredLayers()
......@@ -1006,10 +1010,8 @@ public class TilePlanes {
debugLevel = 20;
}
// if (eig.getD().get(0, 0) == 0.0){
// debugLevel = 10;
// }
if (debugLevel > 0){
if (debugLevel > 3){
// if (debugLevel > 0){
System.out.println("getCovar(): sw = "+sw +", swz = "+swz +", swx = "+swx +", swy = "+swy +", covar.det() = "+covar.det());
System.out.println("getCovar(): covarianvce matrix, number of used points:"+num_tiles);
covar.print(10, 6); // w,d
......@@ -1022,15 +1024,6 @@ public class TilePlanes {
return null; // testing with zero eigenvalue
// Problem with zero eigenvalue is with derivatives and coordinate conversion
}
/*
double [][][] rslt = {
eig.getD().getArray(),
eig.getV().getArray(),
{
{sw,kz,numPoints},
{swz, swx, swy}}};
return rslt;
*/
double [][] eig_val = eig.getD().getArray(); // rslt[0];
double [][] eig_vect = eig.getV().getArray(); // rslt[1];
......@@ -1363,13 +1356,13 @@ public class TilePlanes {
double y = tileSize * (sy + 0.5) + 0.5 - zxy[2];
for (int sx = -3 * superTileSize/2; sx < 3 * superTileSize / 2; sx++){
double x = tileSize * (sx + 0.5) + 0.5 - zxy[1];
disparities[indx] = zxy[0] - (normal[1] * x + normal[2] * y)/normal[0];
disparities[indx] = zxy[0] - (normal[1] * x + normal[2] * y)/normal[0];
indx++;
}
}
return disparities;
}
public double[] getTriplePlaneDisparity(
int dir)
{
......@@ -1389,13 +1382,79 @@ public class TilePlanes {
}
return plane;
}
// double px = tileSize*(superTileSize * sTileXY[0] + superTileSize/2) + zxy[1]; // [3] - plane point {disparity, x, y), x=0, y=0 is a 4,4 point of an 8x8 supertile
// double py = tileSize*(superTileSize * sTileXY[1] + superTileSize/2) + zxy[2];
/**
* Get disparity values for the tiles of this overlapping supertile as [2*superTileSize * 2*superTileSize] array
* and weights combined from provided window function, optional selection and using ellipsoid projection on the
* px, py plane (constant disparity
* @param window null or window function as [2*superTileSize * 2*superTileSize] array
* @param use_sel use plane selection (this.sel_mask) to select only some part of the plane
* @param scale_projection use plane ellipsoid projection for weight: 0 - do not use, > 0 linearly scale ellipsoid
* @return a pair of ar5rays {disparity, strength}, each [2*superTileSize * 2*superTileSize]
*/
public double[][] getDoublePlaneDisparityStrength(
double [] window,
boolean use_sel,
double scale_projection,
int debugLevel)
{
double [][] disp_strength = new double[2][4*superTileSize*superTileSize];
int indx = 0;
double [] normal = getVector();
double [] zxy = getZxy(); // {disparity, x center in pixels, y center in pixels (relative to a supertile center)
double weight = getWeight();
double k_gauss = 0;
Matrix val2d = null, vect2d = null;
if (scale_projection > 0.0){
double [] vals3d = getValues();
double [][] vectors3d = getVectors();
double [][] acovar = new double [2][2];
for (int i = 0; i < 2; i++){
for (int j = i; j < 2; j++){
acovar[i][j] = 0.0;
for (int k = 0; k < 3; k++){
acovar[i][j] += vals3d[k] * vectors3d[k][i+1] * vectors3d[k][j+1]; // 0 - z, disparity == 0
}
if (i != j) {
acovar[j][i] =acovar[i][j];
}
}
}
Matrix covar = new Matrix(acovar); // 2d, x y only
EigenvalueDecomposition eig = covar.eig();
val2d = eig.getD();
vect2d = eig.getV().transpose();
k_gauss = 0.5/(scale_projection*scale_projection);
}
for (int sy = -superTileSize; sy < superTileSize; sy++){
// adding half-tile and half-pixel to match the center of the pixel. Supertile center is between
// pixel 31 and pixel 32 (counting from 0) in both directions
double y = tileSize * (sy + 0.5) + 0.5 - zxy[2];
for (int sx = -superTileSize; sx < superTileSize; sx++){
double x = tileSize * (sx + 0.5) + 0.5 - zxy[1];
disp_strength[0][indx] = zxy[0] - (normal[1] * x + normal[2] * y)/normal[0];
double w = weight;
if (window != null) w *= window[indx];
if (use_sel && (sel_mask != null) && !(sel_mask[indx])) w = 0.0;
if ((w > 0.0) && (scale_projection > 0.0)){
double [] xy = {x,y};
Matrix vxy = vect2d.times(new Matrix(xy,2)); // verify if it is correct
double r2 = 0;
for (int i = 0; i <2; i++){
double d = vxy.get(i,0);
r2 += d * d / val2d.get(i, i);
}
w *= Math.exp(-k_gauss*r2); // verify it is correct size - maybe it should be -0.5*r2 ?
}
disp_strength[1][indx] = w;
indx++;
}
}
return disp_strength;
}
/**
* Cross product of 2 3-d vectors as column matrices
* @param v1
......@@ -1412,8 +1471,8 @@ public class TilePlanes {
{av1[0][0]*av2[1][0] - av1[1][0]*av2[0][0]}};
return new Matrix (ar);
}
/**
* Combine 2 Plane instances using centers, eigenvalues eihenvectors and total weights of this and other PlaneData objects
* other plane should already be transformed to the same supertile coordinate system (with getPlaneToThis() method)
......@@ -1476,7 +1535,7 @@ public class TilePlanes {
other_offset.print(8, 6);
System.out.println("other_fraction="+other_fraction);
}
double [][] acovar = { // covariance matrix of center masses (not yet scaled by weight)
{other_offset.get(0,0)*other_offset.get(0,0), other_offset.get(0,0)*other_offset.get(1,0), other_offset.get(0,0)*other_offset.get(2,0)},
{other_offset.get(1,0)*other_offset.get(0,0), other_offset.get(1,0)*other_offset.get(1,0), other_offset.get(1,0)*other_offset.get(2,0)},
......@@ -1493,7 +1552,7 @@ public class TilePlanes {
System.out.println("covar");
covar.print(8, 6);
}
covar.plusEquals(other_covar.times(other_fraction));
if (debugLevel > 0) {
System.out.println("covar with other_covar");
......@@ -1510,8 +1569,8 @@ public class TilePlanes {
}
// extract new eigenvalues, eigenvectors
EigenvalueDecomposition eig = covar.eig(); // verify NaN - it gets stuck
// eig.getD().getArray(),
// eig.getV().getArray(),
// eig.getD().getArray(),
// eig.getV().getArray(),
if (debugLevel > 0) {
System.out.println("eig.getV()");
eig.getV().print(8, 6);
......@@ -1519,19 +1578,19 @@ public class TilePlanes {
eig.getD().print(8, 6);
}
double [][] eig_vect = eig.getV().getArray();
double [][] eig_val = eig.getD().getArray();
// make towards camera, left coordinate system
/*
/*
int oindx = 0;
for (int i = 1; i <3; i++){
if (Math.abs(eig_vect[0][i]) > Math.abs(eig_vect[0][oindx])){
oindx = i;
}
}
*/
*/
int oindx = 0;
if (preferDisparity) {
for (int i = 1; i <3; i++){
......@@ -1547,7 +1606,7 @@ public class TilePlanes {
}
}
// select 2 other axes for increasing eigenvalues (so v is short axis, h is the long one)
int vindx = (oindx == 0)? 1 : 0;
int hindx = (oindx == 0)? 2 : ((oindx == 1) ? 2 : 1);
......@@ -1565,18 +1624,18 @@ public class TilePlanes {
for (int v = 0; v < 2; v++) {
if (plane[v][v] < 0.0) for (int i = 0; i < 3; i ++) plane[v][i] = -plane[v][i];
}
// make direction last vector so px (x) py (.) disp < 0 (left-hand coordinate system)
if (new Matrix(plane).det() > 0){
for (int i = 0; i < 3; i ++) plane[2][i] = -plane[2][i];
}
PlaneData pd = this.clone(); // will copy selections too
pd.invalidateCalculated(); // real world vectors
pd.setValues(eig_val[oindx][oindx],eig_val[vindx][vindx],eig_val[hindx][hindx]); // eigenvalues [0] - thickness, 2 other to detect skinny (poles)
pd.setVectors(plane);
pd.setZxy(common_center.getColumnPackedCopy()); // set new center
// what weight to use? cloned is original weight for this supertile
// or use weighted average like below?
......@@ -1588,7 +1647,7 @@ public class TilePlanes {
return pd;
}
/**
* Convert plane data from other supertile to this one (disparity, px, py) for the center of this supertile
......@@ -1956,11 +2015,7 @@ public class TilePlanes {
System.out.println("getCovar(): Double.isNaN(eig.getV().get(0, 0))");
debugLevel = 20;
}
// if (eig.getD().get(0, 0) == 0.0){
// debugLevel = 10;
// }
if (debugLevel > 0){
if (debugLevel > 3){
System.out.println("getCovar(): sw = "+sw +", swz = "+swz +", swx = "+swx +", swy = "+swy +", covar.det() = "+covar.det());
System.out.println("getCovar(): covarianvce matrix, number of used points:"+numPoints);
covar.print(10, 6); // w,d
......
src/main/java/TileProcessor.java
View file @ 73b9d03b
This diff is collapsed.
src/main/java/TileSurface.java 0 → 100644
View file @ 73b9d03b
/**
**
** TileSurface - hadle tile surfaces
**
** Copyright (C) 2017 Elphel, Inc.
**
** -----------------------------------------------------------------------------**
**
** TileSurface.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 TileSurface {
// public
private int tileSize;
private int stSize;
private int tilesX;
private int tilesY;
private int [] st_dirs8;
// private int nsTilesstSize = 0; // 8;
GeometryCorrection geometryCorrection = null;
public TileSurface(
int tileSize,
int stSize,
int tilesX,
int tilesY,
GeometryCorrection geometryCorrection){
this.tileSize = tileSize;
this.stSize = stSize;
this.geometryCorrection =geometryCorrection;
this.tilesX = tilesX;
this.tilesY = tilesY;
int [] dirs = {-tilesX, -tilesX + 1, 1, tilesX + 1, tilesX, tilesX - 1, -1, -tilesX - 1};
this.st_dirs8 = dirs;
}
public class TileData{
double disparity;
double strength;
double enable;
int [] neighbors = null;
}
/**
* Get tile surface number from supertile number, direction (-1 same) and the supertile plane index
* @param nsTile number of the supertile
* @param dir direction -1 - same supertile, 0 - N (up), 1 - NE, .. 7 - NW
* @param np number of the supertile plane
* @param planes array of the per-supertile, per plane plane data (each level can be null)
* @return unique tile plane index
*/
public int getTileSurfaceNumber (
int nsTile,
int dir, // direction, or -1 (same)
int np,
TilePlanes.PlaneData [][] planes
){
if ((planes[nsTile] == null) || (planes[nsTile][np] == null)){
return -1; // empty supertile or supertile plane
}
if (dir < 0) {
return np;
}
int tsn = planes[nsTile].length;
for (int d = 0; d < dir; d ++){
int nsTile1 = nsTile + st_dirs8[d];
if (planes[nsTile1] != null){
tsn += planes[nsTile1].length;
}
}
return tsn + np;
}
/**
* Get supertile direction and the plane number that contributeted to a specific tile surface
* @param nsTile supertile index
* @param tp tile surface index (generated by getTileSurfaceNumber)
* @param planes array of the per-supertile, per plane plane data (each level can be null)
* @return a pair of {dir, plane index}. dir is -1 for the plane in the same supertile, 0..7 for neighbors
*/
public int [] getSuperTileDirPlane (
int nsTile,
int tp,
TilePlanes.PlaneData [][] planes
)
{
int num_planes = (planes[nsTile] == null)? 0: planes[nsTile].length;
int [] rslt = {-1, tp};
if (tp < num_planes) return rslt;
tp -= num_planes;
for (int d = 0; d < st_dirs8.length; d ++){
int nsTile1 = nsTile + st_dirs8[d];
num_planes = (planes[nsTile1] == null)? 0: planes[nsTile1].length;
if (tp < num_planes){
rslt[0] = d;
rslt[1] = tp;
return rslt;
}
tp -= num_planes;
}
return null; // error - invalid input
}
}
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