Fixed image zoom/scale option

diff --git a/.gitignore b/.gitignore
index 3820d63..40495b4 100644 (file)
--- a/.gitignore
+++ b/.gitignore
@@ -16,5 +16,6 @@ src/*.o
 stamp-h1
 *~
 .*.swp
+.*swp
 *.bak
 *.lock

diff --git a/ChangeLog.txt b/ChangeLog.txt
index e557841..4076135 100644 (file)
--- a/ChangeLog.txt
+++ b/ChangeLog.txt
@@ -3,5 +3,31 @@ CHANGELOG for CADS/UVS fits2jpeg software
 ------------------------------------------------------------------------
 
 NOTE: This file records the changes made after the release of the stable
-version fits2jpeg-1.0 
+version fits2jpeg-1.0
 ------------------------------------------------------------------------
+
+
+
+Updates in 2.0 Release
+----------------------
+
+  1. Merged scale & operations to single flag
+     v1.0 had normalize/equalize handled by "-e" flag and rest of
+     image operations by "-s" flag. Both are merged into "-s" in
+     this release.
+
+  2. New option to clip image min/max
+     You can limit the pixel-values in output image using the flag
+     "-r". For eg: -r 10:300 to generate jpegs with pixel values in
+     the range 10 to 300.
+
+  3. New option to negate image option
+     A negative image can be generated using "-n" flag.
+
+  4. New option to resize image
+     zoom or shrink output image using "-z <zoomfact>".
+
+  5. New option to choose jpeg quality factor
+     You can choose the jpeg quality factor using "-q <quality>"
+     Quality factor was fixed at 100 in previous release.
+

diff --git a/README.txt b/README.txt
index df79f4a..c5f5ca1 100644 (file)
--- a/README.txt
+++ b/README.txt
@@ -6,12 +6,11 @@ README file for CADS/UVS fits2jpeg software
 
 0. FEATURES
 -----------
-fits2jpeg is a tiny program to read FITS Images and convert them
-to jpeg, which is more popular. This software supports various
-type of intensity scaling and image enhancements. fits2jpeg do not
-support coordinate axes overlay. fits2jpeg supports batch mode
-operations - for instance, converting several fits files to jpeg
-at once.
+fits2jpeg is a tiny program to read FITS Images and convert them to jpeg, which
+is more popular. This software supports various type of intensity scaling and
+image enhancements. fits2jpeg do not support coordinate axes overlay. fits2jpeg
+supports batch mode operations - for instance, converting several fits files to
+jpeg at once.
 
 
 1. BUILD/INSTALL
@@ -26,9 +25,10 @@ Pre-requisites:
     http://heasarc.gsfc.nasa.gov/docs/software/fitsio/fitsio.html
 
  b) JPEGLIB
-    jpeg library and headers. libjpeg is installed by default on most
-    platforms, but we need the header files too. Once you have cfitsio
-    and jpeglib installed, you may proceed with compiling jpeg2fits.
+    jpeg library and headers. libjpeg is installed by default on most platforms,
+    but we need the header files too. Once you have cfitsio and jpeglib
+    installed, you may proceed with compiling jpeg2fits.
+
     Follow these 4 steps:
 
 
@@ -61,20 +61,19 @@ Pre-requisites:
   2. Type `make' to compile the package.
 
   3. Type `make install' to install the programs and any data files and
-     documentation. By default, the binary is copied to /usr/local/bin.
-     You may change this destination by passing on the following
-     argument to configure:
+     documentation. By default, the binary is copied to /usr/local/bin. You may
+     change this destination by passing on the following argument to configure:
 
              --prefix=/your/chosen/destination
 
      Which would result in the binary in /your/chosen/destination/bin
      You may need to ensure that in $PATH to run the program.
 
-  4. You can remove the program binaries and object files from the
-     source code directory by typing `make clean'.
+  4. You can remove the program binaries and object files from the source code
+     directory by typing `make clean'.
 
-You may read more about configure script and others in the accompanying
-file named `INSTALL.txt'.
+You may read more about configure script and others in the accompanying file
+named `INSTALL.txt'.
 
 
 2.USAGE
@@ -116,6 +115,9 @@ file named `INSTALL.txt'.
          0.5            Shrink output to half of input
          2.0            Magnify output to double the size
 
+       fits2jpeg uses a bilinear interpolation based algorithm
+       to scale the image. Allowed range: 0.01 to 4.0
+       NOTE: Anything outside the allowed range will be clipped
 
   Output will be written to <fits_file_root>.jpg. For eg.,
 
@@ -123,8 +125,8 @@ file named `INSTALL.txt'.
 
   writes output to jpeg file 30dor.jpg
 
-  Wild card entries allowed in <fits_file>. For eg: *.fits,
-  m31*.fits ngc???.fits etc.
+  Wild card entries allowed in <fits_file>. For eg: *.fits, m31*.fits
+  ngc???.fits etc.
 
   More examples:
 
@@ -145,11 +147,22 @@ iv.  fits2jpeg -s square -n -r 10:2000 m31.fits
      units.
 
 
+3.1. Sequence of operations
+---------------------------
+Here is the sequence of operations, after reading the fits file:
+
+     i. clip the image to user specified pixel limits, if it was requested
+    ii. Squeeze image pixel values to jpeg limits (0 - 255)
+   iii. image scaling function, default being linear pixel scale
+    iv. Negate the image, if requested
+     v. Image zoom - magnify or shrink the image
+    vi. Write out jpeg file
+
 3. DOCUMENTATION
 ----------------
-An extensive coverage of how this program works along with line
-by line account of the happenings within the code, are dealt with
-in the user manual, which is yet to be written :(
+An extensive coverage of how this program works along with line by line account
+of the happenings within the code, are dealt with in the user manual, which is
+yet to be written :(
 
 
 4. LICENSE: GPL [See the file COPYING.txt for details]
@@ -158,10 +171,9 @@ in the user manual, which is yet to be written :(
 
 5. DISCLAIMER
 -------------
-You may encounter bugs in this software. If you do, please
-report them. Your bug reports are valuable contributions,
-since they allow us to notice and fix problems on
-machines/platforms we don't have, and/or remained un-noticed.
+You may encounter bugs in this software. If you do, please report them. Your bug
+reports are valuable contributions, since they allow us to notice and fix
+problems on machines/platforms we don't have, and/or remained un-noticed.
 
 
 6. REPORTING BUGS
@@ -174,6 +186,7 @@ If you are too lazy, drop in an email to: cads_AT_iiap.res.in
 
 Either way, please include as many details as possible.
 
+
 -----------------------------------------------------------
 Reks, 28 June 2012 <reks_at_iiap.res.in>
       Last modified: 10 July 2012

diff --git a/configure.ac b/configure.ac
index dd6b85e..770550d 100644 (file)
--- a/configure.ac
+++ b/configure.ac
@@ -13,15 +13,16 @@ AC_PROG_MAKE_SET
 # Checks for libraries.
 
 # Checks for header files.
-AC_CHECK_HEADERS([stdio.h stdlib.h string.h math.h float.h ctype.h libgen.h dirent.h])
+AC_CHECK_HEADERS([string.h math.h float.h ctype.h libgen.h dirent.h])
 
 # Checks for typedefs, structures, and compiler characteristics.
 
+
 # Checks for library functions.
 AC_FUNC_MALLOC
 
 LIBS="$LIBS -L/usr/lib64 -L/usr/lib -lm -ljpeg -lcfitsio"
-CFLAGS="-ansi $CFLAGS -I/usr/include -I/usr/include/cfitsio"
+CFLAGS="-ansi -Wall $CFLAGS -I/usr/include -I/usr/include/cfitsio"
 #---------------------------------------------------------------------#
 #           Now we check for cfitsio headers and library              #
 #---------------------------------------------------------------------#

diff --git a/src/fits2jpeg.c b/src/fits2jpeg.c
index 60420e2..cc97e08 100644 (file)
--- a/src/fits2jpeg.c
+++ b/src/fits2jpeg.c
@@ -77,11 +77,10 @@
   4. negate image option
   5. jpeg quality factor
   6. pixel scaling operations moved to image.c
-  7. Image resizing works for reducing size. segfaults for zoom > 1 :(
+  7. Image resizing
   TODO
       1. move fits read to image.c
       2. specify output directory
-      3. Image resizing options
  *---------------------------------------------------------------------------*/
 
 #include "fits2jpeg.h"
@@ -94,7 +93,7 @@ int main(int argc, char *argv[], char *envp[])
 {
     fitsfile *fptr;
 
-    int scale = 0, process = 0, status = 0, jpgqual = 100, nfound, anynull;
+    int scale = 0, status = 0, jpgqual = 100, nfound, anynull;
     unsigned int i = 0, j = 0, usrclip = 0, usrnegt = 0, usrzoom = 0;
     long xdim = 0, ydim = 0, naxes[2], row_stride;
     unsigned long npixels = 0;
@@ -133,6 +132,7 @@ int main(int argc, char *argv[], char *envp[])
             else if (strcmp(optarg, "normalize")==0) scale = 5;
             else if (strcmp(optarg, "equalize") ==0) scale = 6;
             else
+                printwarn(strcat(optarg, " -- Unrecognized option"));
                 scale = 0;
             break;
 
@@ -165,7 +165,6 @@ int main(int argc, char *argv[], char *envp[])
                 sptr = strtok(tmpstr, ":");
                 datamin = atof(sptr);
             }
-            printinfo("User defined pixel range");
 
             /* Remember.. now we have user specified range */
             usrclip = 1;
@@ -187,7 +186,9 @@ int main(int argc, char *argv[], char *envp[])
             /* Do something to scale image
              */
             zoomfact = atof(optarg);
-            if (zoomfact < 0.0) zoomfact = 1.0;
+            zoomfact = abs(zoomfact);
+            if (zoomfact < 0.01) zoomfact = 0.01;
+            if (zoomfact > 4.0) zoomfact = 4.0;
             usrzoom = 1;
             break;
 
@@ -225,7 +226,7 @@ int main(int argc, char *argv[], char *envp[])
 
         /* Read in data */
         npixels = naxes[0] * naxes[1];
-        data = (float *) malloc(sizeof(float) * npixels);
+        data = malloc(sizeof(float) * npixels);
 
         nullval = 0;
         if (fits_read_img(fptr, TFLOAT, 1, npixels, &nullval, data, &anynull,
@@ -243,6 +244,7 @@ int main(int argc, char *argv[], char *envp[])
         /* IF user has provided a range, fix them as min & max in image */
         if (usrclip == 1)
         {
+            printinfo("user defined pixel range");
             for (i = 0; i < npixels; ++i)
             {
                 if (data[i] > datamax) data[i] = datamax;
@@ -250,9 +252,7 @@ int main(int argc, char *argv[], char *envp[])
             } /*endfor*/
         }
 
-        /* Allocate image buffer */
-        image_buffer = (unsigned char *) malloc(sizeof(char) * npixels);
-        scale_pixels(scale, npixels, data, image_buffer);
+        scale_pixels(scale, npixels, data, &image_buffer);
 
         /* Before we write jpeg, check if there is any requirement to negate
          * the image (usrnegt = 1)                                           */
@@ -264,7 +264,11 @@ int main(int argc, char *argv[], char *envp[])
 
         xdim = naxes[0];
         ydim = naxes[1];
-        if (usrzoom == 1) resize_image(&xdim, &ydim, zoomfact, image_buffer);
+        if (usrzoom == 1)
+        {
+            printf("INFO   : Image zoom factor = %3.2f\n", zoomfact);
+            resize_image(&xdim, &ydim, zoomfact, &image_buffer);
+        }
 
         /*Write out data into JPEG file*/
         jpeg_file = fopen(jpeg_file_name, "wb");/* Open JPEG file for writing*/

diff --git a/src/fits2jpeg.h b/src/fits2jpeg.h
index a96f9fb..3920d82 100644 (file)
--- a/src/fits2jpeg.h
+++ b/src/fits2jpeg.h
@@ -39,13 +39,15 @@
 
 #define PROGRAM "fits2jpeg"
 #define r2d (90./atan2(1,0))
-#define MAX_TEXT 150
+#define MAX_TEXT 1024
 
 int my_getopt(int, char * const *, const char *);
+void PRINT_BANNER(void);
+void usage(void);
 void signals_handler(int);
 void set_signals(void);
 void printinfo(const char *);
 void printwarn(const char *);
 void printerro(const char *);
-void scale_pixels(int, unsigned int, float *, JSAMPLE *);
-void resize_image(long *, long *, float, JSAMPLE *);
+void scale_pixels(int, unsigned int, float *, JSAMPLE **);
+void resize_image(long *, long *, float, JSAMPLE **);

diff --git a/src/image.c b/src/image.c
index d1c22f0..f5f6c46 100644 (file)
--- a/src/image.c
+++ b/src/image.c
@@ -27,8 +27,8 @@
 /*---------------------------------------------------------------------------*
  * SCALE_PIXELS: Changes the pixel scale to linear/log/sqroot/etc..
  *---------------------------------------------------------------------------*/
-void scale_pixels(int scale, unsigned int npixels,
-                 float *data, JSAMPLE *image_buffer)
+void scale_pixels(int scale, unsigned int npixels, float *data,
+                 JSAMPLE ** image_buffer)
 {
     unsigned int i = 0;
     int JMAXVAL = 255;
@@ -74,10 +74,14 @@ void scale_pixels(int scale, unsigned int npixels,
     for (i = 1; i <= JMAXVAL; ++i)
         cumhist[i] += cumhist[i - 1] + hist[i];
 
+    /* Allocate image buffer */
+    (*image_buffer) = malloc(sizeof(unsigned char) * npixels);
+
+
     /* Linear scale (min-max) : This is the default scaling
      * histo-eq will fail if we dont generate image_buffer here              */
     for (i = 0; i < npixels; ++i)
-        image_buffer[i] = (int)(data[i]);
+        (*image_buffer)[i] = (int)(data[i]);
 
     /*-----------------------------------------------------------------------*/
 
@@ -88,28 +92,28 @@ void scale_pixels(int scale, unsigned int npixels,
             printinfo("Using square-root scale");
             scl_data = sqrt((float)JMAXVAL)/(float)JMAXVAL;
             for (i = 0; i < npixels; ++i)
-                image_buffer[i] = (int)(sqrt(data[i])/scl_data);
+                (*image_buffer)[i] = (int)(sqrt(data[i])/scl_data);
             break;
 
         case 2 :                                                   /* Square */
             printinfo("Using quadratic scale");
             scl_data = pow((float)JMAXVAL,2)/(float)JMAXVAL;
             for (i = 0; i < npixels; ++i)
-                image_buffer[i] = (int)abs((pow(data[i],2) - 1.0)/scl_data);
+                (*image_buffer)[i] = (int)abs((pow(data[i],2) - 1.0)/scl_data);
             break;
 
         case 3 :                                                    /* Cubic */
             printinfo("Using cubic scale");
             scl_data = pow((float)JMAXVAL,3)/(float)JMAXVAL;
             for (i = 0; i < npixels; ++i)
-                image_buffer[i] = (int)abs((pow(data[i],3) - 1.0)/scl_data);
+                (*image_buffer)[i] = (int)abs((pow(data[i],3) - 1.0)/scl_data);
             break;
 
         case 4 :                                                      /* log */
             printinfo("Using log scale");
             scl_data = log(1.0 + (float)JMAXVAL)/(float)JMAXVAL;
             for (i = 0; i < npixels; ++i)
-                image_buffer[i] = (int)((log(abs(data[i]) + 1.0))/scl_data);
+                (*image_buffer)[i] = (int)((log(abs(data[i]) + 1.0))/scl_data);
             break;
 
         case 5 :
@@ -141,12 +145,12 @@ void scale_pixels(int scale, unsigned int npixels,
             scl_data = (datamax - datamin)/(float)JMAXVAL;
             for (i = 0; i < npixels; ++i)
             {
-                if (image_buffer[i] >= datamax)
-                    image_buffer[i] = JMAXVAL;
-                else if (image_buffer[i] <= datamin)
-                    image_buffer[i] = 0;
+                if ((*image_buffer)[i] >= datamax)
+                    (*image_buffer)[i] = JMAXVAL;
+                else if ((*image_buffer)[i] <= datamin)
+                    (*image_buffer)[i] = 0;
                 else
-                    image_buffer[i] = (int) abs((image_buffer[i]
+                    (*image_buffer)[i] = (int) abs(((*image_buffer)[i]
                                     - datamin)/scl_data);
             }
             break;
@@ -155,19 +159,21 @@ void scale_pixels(int scale, unsigned int npixels,
             /* histogram equalization */
             printinfo("Performing Histogram Equalization");
             for (i = 0; i <  npixels; ++i)
-                image_buffer[i] = cumhist[image_buffer[i]] * JMAXVAL;
+                (*image_buffer)[i] = cumhist[(*image_buffer)[i]] * JMAXVAL;
             break;
         default :
             printinfo("Using linear scale");
             break;
     }
-
 }
 
 /*---------------------------------------------------------------------------*
- * RESIZE_IMAGE: Scales down/up the image_buffer
+ * RESIZE_IMAGE: Scales down/up the image_buffer using bilinear scaling
+ * Based on an article by "John" at
+ * http://tech-algorithm.com/articles/bilinear-image-scaling/
  *---------------------------------------------------------------------------*/
-void resize_image(long *xdim, long *ydim, float zoomfact, JSAMPLE *image_buffer)
+void resize_image(long *xdim, long *ydim, float zoomfact,
+                  JSAMPLE ** image_buffer)
 {
     int offset = 0, index = 0;
     int A, B, C, D, x, y, gray;
@@ -186,8 +192,8 @@ void resize_image(long *xdim, long *ydim, float zoomfact, JSAMPLE *image_buffer)
     xratio = ((float)(w - 1))/zxdim;
     yratio = ((float)(h - 1))/zydim;
 
-                                     /* allocate space for *buff */
-    buff   = (unsigned char *) malloc(sizeof(char) * zxdim * zydim);
+                            /* allocate space for *buff */
+    buff   = malloc(sizeof(unsigned char) * zxdim * zydim);
 
     index  = 0;
     offset = 0;
@@ -203,10 +209,10 @@ void resize_image(long *xdim, long *ydim, float zoomfact, JSAMPLE *image_buffer)
             xdiff = (xratio * j) - x;
             index = y * w + x;
 
-            A = image_buffer[index]         & 0xff;
-            B = image_buffer[index + 1]     & 0xff;
-            C = image_buffer[index + w]     & 0xff;
-            D = image_buffer[index + w + 1] & 0xff;
+            A = (*image_buffer)[index]         & 0xff;
+            B = (*image_buffer)[index + 1]     & 0xff;
+            C = (*image_buffer)[index + w]     & 0xff;
+            D = (*image_buffer)[index + w + 1] & 0xff;
 
             gray = (int)(A * (1 - xdiff) * (1 - ydiff)
                  +       B * (xdiff)     * (1 - ydiff)
@@ -218,10 +224,8 @@ void resize_image(long *xdim, long *ydim, float zoomfact, JSAMPLE *image_buffer)
     }
     *xdim = zxdim;
     *ydim = zydim;
-    image_buffer = realloc(image_buffer, sizeof(char) * npixels);
-    if (!image_buffer)
-        printerro("Failed to allocate memory");
-
+    (*image_buffer) = realloc((*image_buffer), sizeof(unsigned char) * npixels);
     for (i = 0; i <  npixels; ++i)
-        image_buffer[i] = buff[i];
+        (*image_buffer)[i] = buff[i];
+    free(buff);
 }

diff --git a/src/messages.c b/src/messages.c
index edcdc68..75f890c 100644 (file)
--- a/src/messages.c
+++ b/src/messages.c
@@ -62,6 +62,9 @@ void usage()
     printf ("       Resize/Scale output image by <zoomfactor>. Eg:      \n");
     printf ("         0.5            Shrink output to half of input     \n");
     printf ("         2.0            Magnify output to double the size  \n");
+    printf ("                        Allowed range: 0.01 to 4.0         \n");
+    printf ("       NOTE: Zooming will be carried out after all other   \n");
+    printf ("             operations, before writing out jpeg image     \n");
     printf ("  Output will be written to <fits_file_root>.jpg. Wild card\n");
     printf ("  entries allowed in <fits_file>; For eg: *.fits, m31*.fits\n");
     printf ("  ngc???.fits etc.                                         \n");