[GRASS-SVN] r51647 - grass/branches/develbranch_6/imagery/i.pca

svn_grass at osgeo.org svn_grass at osgeo.org
Fri May 18 04:16:08 EDT 2012


Author: mmetz
Date: 2012-05-18 01:16:07 -0700 (Fri, 18 May 2012)
New Revision: 51647

Modified:
   grass/branches/develbranch_6/imagery/i.pca/description.html
   grass/branches/develbranch_6/imagery/i.pca/local_proto.h
   grass/branches/develbranch_6/imagery/i.pca/main.c
   grass/branches/develbranch_6/imagery/i.pca/support.c
Log:
i.pca backport from trunk

Modified: grass/branches/develbranch_6/imagery/i.pca/description.html
===================================================================
--- grass/branches/develbranch_6/imagery/i.pca/description.html	2012-05-18 02:16:57 UTC (rev 51646)
+++ grass/branches/develbranch_6/imagery/i.pca/description.html	2012-05-18 08:16:07 UTC (rev 51647)
@@ -1,6 +1,5 @@
 <h2>DESCRIPTION</h2>
 
-
 <em>i.pca</em> is an image processing program based on the
 algorithm provided by Vali (1990), that processes n
 (n&nbsp;&gt;=&nbsp;2) input raster map layers and produces n output
@@ -33,12 +32,12 @@
 <dd>If output is rescaled, the output raster will be of type CELL. If 
 the output is not rescaled, the output raster will be of type DCELL.
 
-<dt><B>-n</B> <EM>normalize input raster maps</EM>
-<dd>By default, the values of the input raster maps are centered for each
-map separately with <EM>x - mean</EM>. With <EM>-n</EM>, the input raster
-maps are normalized for each map separately with <EM>(x - mean) / stddev</EM>.
+<dt><b>-n</b> <em>normalize input raster maps</em> 
+<dd>By default, the values of the input raster maps are centered for each 
+map separately with <em>x - mean</em>. With <em>-n</em>, the input raster 
+maps are normalized for each map separately with <em>(x - mean) / stddev</em>. 
 Normalizing is highly recommended when the input raster maps have different
-units, e.g. represent different environmental parameters.
+units, e.g. represent different environmental parameters. 
 </dl>
 
 
@@ -47,9 +46,8 @@
 Richards (1986) gives a good example of the application of principal
 components analysis (pca) to a time series of LANDSAT images of a burned
 region in Australia.
-<p>
-Eigenvalue and eigenvector information is stored in the output maps'
-history files. View with <em>r.info -h</em>.
+<p>Eigenvalue and eigenvector information is stored in the output maps'
+history files. View with <em>r.info</em>.
 
 
 <h2>EXAMPLE</h2>
@@ -74,14 +72,12 @@
 Springer-Verlag, 1986.
 
 <p>
-
 Vali, Ali R.,
 Personal communication,
 Space Research Center, 
 University of Texas, Austin, 1990.
 
 <p>
-
 <em>
 <a href="i.cca.html">i.cca</a><br>
 <a href="i.class.html">i.class</a><br>
@@ -96,15 +92,12 @@
 <h2>AUTHOR</h2>
 
 David Satnik, GIS Laboratory
-<p>
-Major modifications for GRASS 4.1 were made by <br>
+<p>Major modifications for GRASS 4.1 were made by <br>
 Olga Waupotitsch and
 Michael Shapiro,
 U.S.Army Construction Engineering 
 Research Laboratory
-<p>
-Rewritten for GRASS 6.x and major modifications by <br>
+<p>Rewritten for GRASS 6.x and major modifications by <br>
 Brad Douglas
 
-<p>
-<i>Last changed: $Date$</i>
+<p><i>Last changed: $Date$</i>

Modified: grass/branches/develbranch_6/imagery/i.pca/local_proto.h
===================================================================
--- grass/branches/develbranch_6/imagery/i.pca/local_proto.h	2012-05-18 02:16:57 UTC (rev 51646)
+++ grass/branches/develbranch_6/imagery/i.pca/local_proto.h	2012-05-18 08:16:07 UTC (rev 51647)
@@ -2,6 +2,6 @@
 #define __LOCAL_PROTO_H__
 
 /* support.c */
-int write_support(int, char *, double **, double *);
+void write_support(int, char *, double **, double *);
 
 #endif

Modified: grass/branches/develbranch_6/imagery/i.pca/main.c
===================================================================
--- grass/branches/develbranch_6/imagery/i.pca/main.c	2012-05-18 02:16:57 UTC (rev 51646)
+++ grass/branches/develbranch_6/imagery/i.pca/main.c	2012-05-18 08:16:07 UTC (rev 51647)
@@ -5,10 +5,11 @@
  *
  * AUTHOR(S):    Original author Center for Space Research (Uni. of TX)
  *               Rewritten by Brad Douglas <rez touchofmadness com>
+ *               NULL value/MASK handling and speed up by Markus Metz
  *
- * PURPOSE:      Principal Component Analysis transform of satellite data.
+ * PURPOSE:      Principal Component Analysis transform of raster data.
  *
- * COPYRIGHT:    (C) 2004-2008 by the GRASS Development Team
+ * COPYRIGHT:    (C) 2004-2011 by the GRASS Development Team
  *
  *               This program is free software under the GNU General Public
  *               License (>=v2). Read the file COPYING that comes with GRASS
@@ -33,8 +34,7 @@
 /* function prototypes */
 static CELL round_c(double);
 static int set_output_scale(struct Option *, int *, int *, int *);
-static int calc_mu(int *, double *, double *, int);
-static int calc_covariance(int *, double **, double *, double *, int);
+static int calc_mu_cov(int *, double **, double *, double *, int);
 static int write_pca(double **, double *, double *, int *, char *, int,
                      int, int, int);
 
@@ -142,12 +142,9 @@
 			  opt_in->answers[i]);
     }
 
-    G_verbose_message(_("Calculating covariance matrix..."));
-    if (!calc_mu(inp_fd, mu, stddev, bands))
+    if (!calc_mu_cov(inp_fd, covar, mu, stddev, bands))
 	G_fatal_error(_("No non-null values"));
 
-    calc_covariance(inp_fd, covar, mu, stddev, bands);
-
     G_math_d_copy(covar[0], eigmat[0], bands*bands);
     G_debug(1, "Calculating eigenvalues and eigenvectors...");
     G_math_eigen(eigmat, eigval, bands);
@@ -237,86 +234,32 @@
 }
 
 
-static int calc_mu(int *fds, double *mu, double *stddev, int bands)
+static int calc_mu_cov(int *fds, double **covar, double *mu, 
+                           double *stddev, int bands)
 {
-    int i;
+    int i, j;
     int row, col;
     int rows = G_window_rows();
     int cols = G_window_cols();
     off_t count = 0;
-    double *sumsq;
     DCELL **rowbuf = (DCELL **) G_malloc(bands * sizeof(DCELL *));
+    double **sum2 = (double **)G_calloc(bands, sizeof(double *));
+    double *sumsq, *sd, *sum;
 
-    for (i = 0; i < bands; i++) {
-	rowbuf[i] = G_allocate_d_raster_buf();
-    }
-
     if (stddev) {
-	G_message(_("Computing means and standard deviations..."));
 	sumsq = (double *)G_calloc(bands, sizeof(double));
+	sd = (double *)G_calloc(bands, sizeof(double));
     }
     else {
-	G_message(_("Computing means..."));
 	sumsq = NULL;
+	sd = NULL;
     }
 
-    for (row = 0; row < rows; row++) {
-	G_percent(row, rows, 2);
-	for (i = 0; i < bands; i++)
-	    G_get_d_raster_row(fds[i], rowbuf[i], row);
-
-	for (col = 0; col < cols; col++) {
-	    /* ignore cells where any of the maps has null value */
-	    for (i = 0; i < bands; i++)
-		if (G_is_d_null_value(&rowbuf[i][col]))
-		    break;
-	    if (i != bands)
-		continue;
-	    count++;
-	    for (i = 0; i < bands; i++) {
-		mu[i] += rowbuf[i][col];
-		if (stddev)
-		    sumsq[i] += rowbuf[i][col] * rowbuf[i][col];
-	    }
-	}
-    }
-    G_percent(1, 1, 1);
-    
-    if (count < 2)
-	return 0;
-
     for (i = 0; i < bands; i++) {
-	if (stddev)
-	    /* mu[i] is here still the sum */
-	    stddev[i] = sqrt((sumsq[i] - mu[i] * mu[i] / count) /
-	                     (count - 1));
-
-	mu[i] = mu[i] / count;
-	G_free(rowbuf[i]);
-    }
-
-    if (rowbuf)
-	G_free(rowbuf);
-    if (sumsq)
-	G_free(sumsq);
-
-    return 1;
-}
-
-
-static int calc_covariance(int *fds, double **covar, double *mu, 
-                           double *stddev, int bands)
-{
-    int i, j;
-    int row, col;
-    int rows = G_window_rows();
-    int cols = G_window_cols();
-    off_t count = 0;
-    DCELL **rowbuf = (DCELL **) G_malloc(bands * sizeof(DCELL *));
-
-    for (i = 0; i < bands; i++) {
 	rowbuf[i] = G_allocate_d_raster_buf();
+	sum2[i] = (double *)G_calloc(bands, sizeof(double));
     }
+    sum = mu;
 
     G_message(_("Computing covariance matrix..."));
 
@@ -334,38 +277,54 @@
 		continue;
 	    count++;
 	    for (i = 0; i < bands; i++) {
-		DCELL dval1 = rowbuf[i][col];
 
-		for (j = i; j < bands; j++) {
-		    DCELL dval2 = rowbuf[j][col];
+		sum[i] += rowbuf[i][col];
+		if (stddev)
+		    sumsq[i] += rowbuf[i][col] * rowbuf[i][col];
 
-		    if (stddev) {
-			covar[i][j] += (dval1 - mu[i]) * (dval2 - mu[j]) /
-			               (stddev[i] * stddev[j]);
-		    }
-		    else {
-			covar[i][j] += (dval1 - mu[i]) * (dval2 - mu[j]);
-		    }
-
-		}
+		for (j = 0; j <= i; j++)
+		    sum2[i][j] += rowbuf[i][col] * rowbuf[j][col];
 	    }
 	}
     }
     G_percent(1, 1, 1);
     
+    if (count < 2)
+	return 0;
+
     for (i = 0; i < bands; i++) {
-	for (j = i; j < bands; j++) {
-	    covar[i][j] = covar[i][j] / (count - 1);
+	if (stddev) {
+	    sd[i] = sqrt(count * sumsq[i] - sum[i] * sum[i]);
+	    stddev[i] = sqrt((sumsq[i] - sum[i] * sum[i] / count) /
+	                     (count - 1));
+	}
+	for (j = 0; j <= i; j++) {
+	    if (stddev)
+		covar[i][j] = (count * sum2[i][j] - sum[i] * sum[j]) /
+		              (sd[i] * sd[j]);
+	    else
+		covar[i][j] = (sum2[i][j] - sum[i] * sum[j] / count) /
+		              (count - 1);
 	    G_debug(3, "covar[%d][%d] = %f", i, j, covar[i][j]);
 	    if (j != i)
 		covar[j][i] = covar[i][j];
 	}
 
+	G_free(sum2[i]);
 	G_free(rowbuf[i]);
     }
+    for (i = 0; i < bands; i++)
+	mu[i] = sum[i] / count;
+
     G_free(rowbuf);
+    
+    G_free(sum2);
+    if (sd)
+	G_free(sd);
+    if (sumsq)
+	G_free(sumsq);
 
-    return 0;
+    return 1;
 }
 
 
@@ -375,165 +334,147 @@
 	  int scale, int scale_min, int scale_max)
 {
     int i, j;
-    void *outbuf, *outptr;
-    double min = 0.;
-    double max = 0.;
-    double old_range = 0.;
+    void **outbuf = (void **) G_malloc(bands * sizeof(void *));
+    void **outptr = (void **) G_malloc(bands * sizeof(void *));
+    double *min = (double *) G_malloc(bands * sizeof(double));
+    double *max = (double *) G_malloc(bands * sizeof(double));
+    double *old_range = (double *) G_calloc(bands, sizeof(double));
     double new_range = 0.;
+    int pass;
     int rows = G_window_rows();
     int cols = G_window_cols();
     /* why CELL_TYPE when scaling output ? */
-    int outmap_type = (scale) ? CELL_TYPE : DCELL_TYPE;
+    int outmap_type = (scale) ? CELL_TYPE : FCELL_TYPE;
     int outcell_mapsiz = G_raster_size(outmap_type);
-    DCELL *d_buf;
+    int *out_fd = (int *) G_malloc(bands * sizeof(int));
+    DCELL **inbuf = (DCELL **) G_malloc(bands * sizeof(DCELL *));
 
     /* 2 passes for rescale.  1 pass for no rescale */
     int PASSES = (scale) ? 2 : 1;
 
-    /* temporary row storage */
-    d_buf = G_allocate_d_raster_buf();
-
-    /* allocate memory for output row buffer */
-    outbuf = G_allocate_raster_buf(outmap_type);
-
-    if (!outbuf)
-	G_fatal_error(_("Unable to allocate memory for raster row"));
-
+    /* allocate memory for row buffers */
     for (i = 0; i < bands; i++) {
-	char name[100];
-	int out_fd;
-	int pass;
+	char name[GNAME_MAX];
 
+	/* open output raster maps */
 	sprintf(name, "%s.%d", out_basename, i + 1);
+	out_fd[i] = G_open_raster_new(name, outmap_type);
 
-	G_message(_("Transforming <%s>..."), name);
+	inbuf[i] = G_allocate_d_raster_buf();
+	outbuf[i] = G_allocate_raster_buf(outmap_type);
+	min[i] = max[i] = old_range[i] = 0;
+    }
 
-	/* open a new file for output */
-	out_fd = G_open_raster_new(name, outmap_type);
+    for (pass = 1; pass <= PASSES; pass++) {
+	int row, col;
+	int first = 1;
 
-	for (pass = 1; pass <= PASSES; pass++) {
-	    void *rowbuf = NULL;
-	    int row, col;
+	if (scale && (pass == PASSES)) {
+	    G_message(_("Rescaling to range %d,%d..."),
+		      scale_min, scale_max);
 
-	    if (scale && (pass == PASSES)) {
-		G_message(_("Rescaling <%s> to range %d,%d..."),
-			  name, scale_min, scale_max);
+	    for (i = 0; i < bands; i++)
+		old_range[i] = max[i] - min[i];
+	    new_range = (double)(scale_max - scale_min);
+	}
+	else {
+	    G_message(_("Calculating principal components..."));
+	}
 
-		old_range = max - min;
-		new_range = (double)(scale_max - scale_min);
-	    }
+	for (row = 0; row < rows; row++) {
 
-	    for (row = 0; row < rows; row++) {
-		void *rowptr;
+	    G_percent(row, rows, 2);
 
-		G_percent(row, rows, 2);
+	    for (i = 0; i < bands; i++) {
+		G_get_d_raster_row(inp_fd[i], inbuf[i], row);
+		outptr[i] = outbuf[i];
+	    }
+	    for (col = 0; col < cols; col++) {
+		/* ignore cells where any of the maps has null value */
+		for (i = 0; i < bands; i++)
+		    if (G_is_d_null_value(&inbuf[i][col]))
+			break;
+			
+		if (i != bands) {
+		    for (i = 0; i < bands; i++) {
+			G_set_null_value(outptr[i], 1, outmap_type);
+			outptr[i] =
+			    G_incr_void_ptr(outptr[i], outcell_mapsiz);
+		    }
+		    continue;
+		}
 
-		/* reset d_buf */
-		for (col = 0; col < cols; col++)
-		    d_buf[col] = 0.;
+		for (i = 0; i < bands; i++) {
+		    DCELL dval = 0.;
 
-		for (j = 0; j < bands; j++) {
-		    RASTER_MAP_TYPE maptype =
-			G_get_raster_map_type(inp_fd[j]);
-
-		    /* don't assume each image is of the same type */
-		    if (rowbuf)
-			G_free(rowbuf);
-		    if (!(rowbuf = G_allocate_raster_buf(maptype)))
-			G_fatal_error(_("Unable allocate memory for row buffer"));
-
-		    if (G_get_raster_row(inp_fd[j], rowbuf, row, maptype) < 0)
-			G_fatal_error(_("Unable to read raster map row %d"), row);
-
-		    rowptr = rowbuf;
-		    outptr = outbuf;
-
-		    /* add into the output cell eigmat[i][j] * corresp cell 
-		     * of j-th band for current j */
-		    for (col = 0; col < cols; col++) {
-			DCELL dval;
-
-			/* handle null cells */
-			if (G_is_null_value(rowptr, maptype)) {
-			    G_set_null_value(outptr, 1, outmap_type);
-			    outptr =
-				G_incr_void_ptr(outptr, outcell_mapsiz);
-
-			    rowptr =
-				G_incr_void_ptr(rowptr,
-						G_raster_size(maptype));
-			    continue;
-			}
-
+		    for (j = 0; j < bands; j++) {
 			/* corresp. cell of j-th band */
-			dval = G_get_raster_value_d(rowptr, maptype);
 			if (stddev)
-			    d_buf[col] += eigmat[i][j] * ((dval - mu[j]) / stddev[j]);
+			    dval += eigmat[i][j] * 
+			            ((inbuf[j][col] - mu[j]) / stddev[j]);
 			else
-			    d_buf[col] += eigmat[i][j] * (dval - mu[j]);
+			    dval += eigmat[i][j] * (inbuf[j][col] - mu[j]);
+		    }
 
-			/* the cell entry is complete */
-			if (j == (bands - 1)) {
-			    if (scale && (pass == 1)) {
-				if ((row == 0) && (col == 0))
-				    min = max = d_buf[0];
 
-				if (d_buf[col] < min)
-				    min = d_buf[col];
+		    /* the cell entry is complete */
+		    if (scale && (pass == 1)) {
+			if (first)
+			    min[i] = max[i] = dval;
+			if (dval < min[i])
+			    min[i] = dval;
 
-				if (d_buf[col] > max)
-				    max = d_buf[col];
-			    }
-			    else if (scale) {
+			if (dval > max[i])
+			    max[i] = dval;
+		    }
+		    else if (scale) {
 
-				if (min == max) {
-				    G_set_raster_value_c(outptr, 1,
-							 CELL_TYPE);
-				}
-				else {
-				    /* map data to 0, (new_range-1) and then adding new_min */
-				    CELL tmpcell =
-					round_c((new_range *
-						 (d_buf[col] -
-						  min) / old_range) +
-						scale_min);
-
-				    G_set_raster_value_c(outptr, tmpcell,
-							 outmap_type);
-				}
-			    }
-			    else {	/* (!scale) */
-
-				G_set_raster_value_d(outptr, d_buf[col],
-						     outmap_type);
-			    }
+			if (min[i] == max[i]) {
+			    G_set_raster_value_c(outptr[i], 1,
+						 CELL_TYPE);
 			}
+			else {
+			    /* map data to 0, (new_range-1) and then adding new_min */
+			    CELL tmpcell =
+				round_c((new_range * (dval - min[i]) /
+				         old_range[i]) + scale_min);
 
-			outptr = 
-			    G_incr_void_ptr(outptr, outcell_mapsiz);
-
-			rowptr =
-			    G_incr_void_ptr(rowptr, G_raster_size(maptype));
+			    G_set_raster_value_c(outptr[i], tmpcell,
+						 outmap_type);
+			}
 		    }
-		}		/* for j = 0 to bands */
+		    else {	/* (!scale) */
 
-		if (pass == PASSES) {
-		    G_put_raster_row(out_fd, outbuf, outmap_type);
+			G_set_raster_value_d(outptr[i], dval,
+					     outmap_type);
+		    }
+		    outptr[i] = G_incr_void_ptr(outptr[i], outcell_mapsiz);
 		}
+		first = 0;
 	    }
+	    if (pass == PASSES) {
+		for (i = 0; i < bands; i++)
+		    G_put_raster_row(out_fd[i], outbuf[i], outmap_type);
+	    }
+	}
+	G_percent(1, 1, 1);
 
-	    G_percent(1, 1, 1);
-
-	    /* close output file */
-	    if (pass == PASSES)
-		G_close_cell(out_fd);
+	/* close output file */
+	if (pass == PASSES) {
+	    for (i = 0; i < bands; i++) {
+		G_close_cell(out_fd[i]);
+		G_free(inbuf[i]);
+		G_free(outbuf[i]);
+	    }
 	}
     }
 
-    if (d_buf)
-	G_free(d_buf);
-    if (outbuf)
-	G_free(outbuf);
+    G_free(inbuf);
+    G_free(outbuf);
+    G_free(outptr);
+    G_free(min);
+    G_free(max);
+    G_free(old_range);
 
     return 0;
 }

Modified: grass/branches/develbranch_6/imagery/i.pca/support.c
===================================================================
--- grass/branches/develbranch_6/imagery/i.pca/support.c	2012-05-18 02:16:57 UTC (rev 51646)
+++ grass/branches/develbranch_6/imagery/i.pca/support.c	2012-05-18 08:16:07 UTC (rev 51647)
@@ -1,15 +1,15 @@
 #include <string.h>
 #include <grass/gis.h>
 #include <grass/glocale.h>
+#include "local_proto.h"
 
-
 /* function prototypes */
-static int write_history(int, char *, double **, double *);
+static void write_history(int, char *, double **, double *);
 
 
-int write_support(int bands, char *outname, double **eigmat, double *eigval)
+void write_support(int bands, char *outname, double **eigmat, double *eigval)
 {
-    char *mapset = G_mapset();
+    const char *mapset = G_mapset();
     struct Colors colors;
     struct FPRange range;
     DCELL min, max;
@@ -23,14 +23,13 @@
     if (G_raster_map_is_fp(outname, mapset))
 	G_mark_colors_as_fp(&colors);
 
-    if (G_write_colors(outname, mapset, &colors) < 0)
-	G_message(_("Unable to write color table for raster map <%s>"), outname);
+    G_write_colors(outname, mapset, &colors);
 
-    return write_history(bands, outname, eigmat, eigval);
+    write_history(bands, outname, eigmat, eigval);
 }
 
 
-static int write_history(int bands, char *outname, double **eigmat, double *eigval)
+static void write_history(int bands, char *outname, double **eigmat, double *eigval)
 {
     int i, j;
     static int first_map = TRUE;     /* write to stderr? */
@@ -52,29 +51,28 @@
 	sprintf(tmpeigen, "PC%d %9.2f (", i+1, eigval[i]);
 	for (j = 0; j < bands; j++) {
 	    sprintf(tmpa, "%7.4f", eigmat[i][j]);
-	    G_strcat(tmpeigen, tmpa);
+	    strcat(tmpeigen, tmpa);
 	    if (j < (bands - 1) ){
 		sprintf(tmpa, ",");
-		G_strcat(tmpeigen, tmpa);
+		strcat(tmpeigen, tmpa);
 	    }
 	}
-	G_strcat(tmpeigen, ") ");
-
+	strcat(tmpeigen, ") ");
+	
 	sprintf(tmpa, "[%5.2f%%]", eigval[i] * 100/eigval_total);
-	G_strcat(tmpeigen, tmpa);
+	strcat(tmpeigen, tmpa);
 
 	sprintf(hist.edhist[i + 1], tmpeigen);
 
 	/* write eigen values to screen */
 	if(first_map)
-	    G_message("%s", tmpeigen);
+	    fprintf(stdout, "%s\n", tmpeigen);
     }
 
-    hist.edlinecnt = i + 1;
     G_command_history(&hist);
 
     /* only write to stderr the first time (this fn runs for every output map) */
     first_map = FALSE;
 
-    return G_write_history(outname, &hist);
+    G_write_history(outname, &hist);
 }



More information about the grass-commit mailing list