/*
* Copyright (C) 2014 Dr. John Lindsay <jlindsay@uoguelph.ca>
*
* 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 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/>.
*/
package whitebox.algorithms;
import java.util.Date;
import whitebox.geospatialfiles.WhiteboxRaster;
/**
* This algorithm takes an input raster of categorical data (e.g. a land-use
* image) and assigns a unique identifier value to each contiguous group of
* same-valued grid cells (i.e. region).
*
* @author johnlindsay
*/
public class Clump {
double noData = -32768;
WhiteboxRaster image;
WhiteboxRaster output;
double currentPatchNumber = 0;
double currentImageValue = 0;
int maxDepth = 1000;
int depth = 0;
int[] dX = new int[]{1, 1, 1, 0, -1, -1, -1, 0};
int[] dY = new int[]{-1, 0, 1, 1, 1, 0, -1, -1};
int numScanCells = 8;
boolean blnIncludeDiagNeighbour = true;
boolean blnTreatZerosAsBackground = false;
String outputHeader = "";
public Clump(String inputHeaderFile) {
this.image = new WhiteboxRaster(inputHeaderFile, "r");
}
public Clump(WhiteboxRaster inputImage) {
this.image = inputImage;
}
public Clump(WhiteboxRaster inputImage, boolean includeDiagonalNeighbours) {
this.image = inputImage;
this.blnIncludeDiagNeighbour = includeDiagonalNeighbours;
if (!blnIncludeDiagNeighbour) {
dX = new int[]{0, 1, 0, -1};
dY = new int[]{-1, 0, 1, 0};
}
}
public Clump(WhiteboxRaster inputImage, boolean includeDiagonalNeighbours,
boolean treatZerosAsBackground) {
this.image = inputImage;
this.blnIncludeDiagNeighbour = includeDiagonalNeighbours;
if (!blnIncludeDiagNeighbour) {
dX = new int[]{0, 1, 0, -1};
dY = new int[]{-1, 0, 1, 0};
}
this.blnTreatZerosAsBackground = treatZerosAsBackground;
}
public String getOutputHeader() {
return outputHeader;
}
public void setOutputHeader(String outputHeader) {
this.outputHeader = outputHeader;
}
public void setIncludeDiagonalNeighbours(boolean value) {
this.blnIncludeDiagNeighbour = value;
if (!blnIncludeDiagNeighbour) {
dX = new int[]{0, 1, 0, -1};
dY = new int[]{-1, 0, 1, 0};
}
}
public WhiteboxRaster run() throws Exception {
int row, col, x, y, i;
boolean blnFoundNeighbour;
double maxPatchValue = 1;
numScanCells = dY.length;
int rows = image.getNumberRows();
int cols = image.getNumberColumns();
noData = image.getNoDataValue();
double initialValue = -1;
if (outputHeader.isEmpty()) {
outputHeader = image.getHeaderFile().replace(".dep", "_clumped.dep");
}
output = new WhiteboxRaster(outputHeader, "rw", image.getHeaderFile(), WhiteboxRaster.DataType.FLOAT, initialValue);
output.setDataScale(WhiteboxRaster.DataScale.CATEGORICAL);
output.setPreferredPalette("qual.pal");
if (blnTreatZerosAsBackground) {
for (row = 0; row < rows; row++) {
for (col = 0; col < cols; col++) {
if (image.getValue(row, col) == 0) {
output.setValue(row, col, 0);
}
}
}
if (output.getValue(0, 0) == -1) {
output.setValue(0, 0, maxPatchValue);
// recursively scan all connected cells of equal value in Image
depth = 0;
ScanConnectedCells(0, 0, image.getValue(0, 0), initialValue, maxPatchValue);
}
} else {
output.setValue(0, 0, maxPatchValue);
// recursively scan all connected cells of equal value in Image
depth = 0;
ScanConnectedCells(0, 0, image.getValue(0, 0), initialValue, maxPatchValue);
}
double patchValue = 0;
double neighbourPatchValue = 0;
double newPatchValue = 0;
double imageValue = 0;
for (row = 0; row < rows; row++) {
for (col = 0; col < cols; col++) {
imageValue = image.getValue(row, col);
if (imageValue != noData) {
patchValue = output.getValue(row, col);
if (patchValue == initialValue) {
// see if any neighbour has the same value in the input image
blnFoundNeighbour = false;
for (i = 0; i < numScanCells; i++) {
x = col + dX[i];
y = row + dY[i];
neighbourPatchValue = output.getValue(y, x);
if (neighbourPatchValue != initialValue
&& image.getValue(y, x) == imageValue) {
// cell is neighbouring a cell with the same value in image that
// has already been assigned a patch value
output.setValue(row, col, neighbourPatchValue);
newPatchValue = neighbourPatchValue;
blnFoundNeighbour = true;
break;
}
}
if (!blnFoundNeighbour) {
// no neighbouring cell has the same value in Image and has
// already been assigned a value. A new one is needed.
maxPatchValue++;
newPatchValue = maxPatchValue;
output.setValue(row, col, newPatchValue);
}
// recursively scan all connected cells of equal value in Image
depth = 0;
ScanConnectedCells(row, col, imageValue, initialValue, newPatchValue);
}
} else {
output.setValue(row, col, noData);
}
}
}
// find all cells with neighbouring cells that have the same value in
// the input image but different patch values in the output image.
// Recursively scan them to change the larger patch ID to the lower value.
// Iterate this process until there are no further changes to the image.
boolean somethingDone;
double[] reclass = new double[(int) maxPatchValue + 1];
// this array is used to keep track of the eliminated patches.
do {
somethingDone = false;
for (row = 0; row < rows; row++) {
for (col = 0; col < cols; col++) {
imageValue = image.getValue(row, col);
if (imageValue != noData) {
patchValue = output.getValue(row, col);
for (i = 0; i < numScanCells; i++) {
x = col + dX[i];
y = row + dY[i];
neighbourPatchValue = output.getValue(y, x);
if (neighbourPatchValue != patchValue
&& image.getValue(y, x) == imageValue) {
// The two patches are equivalent. Find the
// lower valued cell and initiate a recursive
// scan from there.
somethingDone = true;
if (patchValue < neighbourPatchValue) {
reclass[(int) neighbourPatchValue] = -1;
output.setValue(y, x, patchValue);
ScanConnectedCells(y, x, imageValue, neighbourPatchValue, patchValue);
} else {
reclass[(int) patchValue] = -1;
output.setValue(row, col, neighbourPatchValue);
ScanConnectedCells(row, col, imageValue, patchValue, neighbourPatchValue);
patchValue = neighbourPatchValue;
}
}
}
}
}
}
} while (somethingDone);
i = 0;
for (int a = 0; a < maxPatchValue + 1; a++) {
if (reclass[a] != -1) {
reclass[a] = i;
i++;
}
}
for (row = 0; row < rows; row++) {
for (col = 0; col < cols; col++) {
patchValue = output.getValue(row, col);
if (patchValue != noData) {
output.setValue(row, col, reclass[(int) patchValue]);
}
}
}
output.addMetadataEntry("Created by the "
+ "Clump algorithm.");
output.addMetadataEntry("Created on " + new Date());
output.flush();
output.writeHeaderFile();
return output;
}
private void ScanConnectedCells(int row, int col, double imageValue, double currentPatchValue, double newPatchValue) {
depth++;
int x, y;
if (depth < maxDepth) {
for (int c = 0; c < numScanCells; c++) {
x = col + dX[c];
y = row + dY[c];
if ((output.getValue(y, x) == currentPatchValue)
&& (image.getValue(y, x) == imageValue)) {
// cell should be assigned the new patch value and has the same value in Image
output.setValue(y, x, newPatchValue);
ScanConnectedCells(y, x, imageValue, currentPatchValue, newPatchValue);
}
}
}
depth--;
}
}