In March2014 I began writing various image processing modules using python. To generate test input for some of these code I have written a suite of Test Image (tim) codes. These are gfortran codes in the OTW environment. Unless otherwise noted, these test images have the output size of 256x256 pixels written as floating point (BITPIX=-32).
This generates an image where each column of pixels has a fixed value. Hence we are generating an image of "stripes". The arguments to tim_stripes are fed in a pars.in file. This file is generated by the script tim_stripes.sh. The arguments are the start and end stripe values, and the name of the output image to be generated.
% tim_stripes.sh 1 256 R.fitsThe image generated by this called, named R.fits, is shown in the figure below. The first column (x=1) has a values of I=1, and the last column (x=256) has a value of I=256. As an aside, I show below the commands I use to generate this image. The command named "jpeg256.py" is a python code that takes 3 (RGB) images, and generates an output jpeg image file. Actually, the whole genesis of the tim series of tasks lies in developing python codes like this. For now, we simply need to know that "jpeg256.py" takes the 3 images from tim_stripes runs, and generates a grey-scale image.
To create the figure below: # tim_stripes.sh 1 256 R.fits tim_stripes.sh 1 256 G.fits tim_stripes.sh 1 256 B.fits python jpeg256.py mv sco.jpeg test_stripes.jpeg It is useful to note here that I used the ImageMagick "display" command to view the test_stripes.jpeg file: # display test_stripes.jpegTo add the yellow test label seen in the figure below I just left-clicked in the display window. I then selected from the menu "Image Edit > Annotate". When doe with the labeleing I used the "File > Save" option to save the updated jpeg file. This is a rather klugy, but easy and diret way to make labeled images.
 |
| An example of an image generated with tim_stripes. As descibed above, I used the useful command "display tim_stripes.jpeg" to display this jpeg (from the command line). I then used a few simple cursor-based command to add the yellow text label in the center of the image. |
This generates an image of a circle. The brightness distribution with in the circle is defined radially using arguments 4 and 5. The position and size of the circle are also input values. I used this simple code to generate red, green and blue images that were then combined with the "jpeg256.py" code to create the color figure shown below.
# tim_circle.sh 128 100 90.0 256 200 R.fits tim_circle.sh 100 160 90.0 256 200 G.fits tim_circle.sh 160 160 90.0 256 200 B.fits python jpeg256.py convert sco.jpeg -resize 512x512 tim_circle_1.jpeg I use "convert" to make the image larger (512x512). Here are the arguments fed to tim_circle.sh: arg1 - Xo of circle arg2 - Yo of circle arg3 - Ro, radius of circle arg4 - I value at R=0 arg5 - I value at R=Ro arg6 - name of output test image
 |
| Color image generated with three (RGB) images from tim_circle.sh. For each circle (each a different color) the center of the circle has I=256 and the edge of the circle has I=200. The overlap regions are all white, since the color ratios from each image are equal for each color. |
One of the reasons for writing tim_circle was to illustrate not only how color images are generated, but how brightness is controlled in a color image. Below I show a script used to change the relative colors and brightnesses of the above figure. The resultant figure is displayed below.
% cat R2_2 # tim_circle.sh 128 100 90.0 256 100 R.fits tim_circle.sh 100 160 90.0 256 175 G.fits tim_circle.sh 160 160 90.0 256 40 B.fits jpeg256.py convert sco.jpeg -resize 512x512 tim_circle_2.jpeg display tim_circle_2.jpeg &
 |
| Color image generated with three (RGB) images from tim_circle.sh. The variation in pixel values in the constituent RGB images control not only the resultant color, but also the brightness in the output image. |