Imaging Experiments

Reproducing The Effective Field Of The Microscopic Imager - Part 2

BACK

   Converting the original images into a monochrome image was simple.  A picture was taken with the digital camera and uploaded into the computer.

   Here is a representative image as it was produced by the camera.  This is not meant as a reproduction of the image field; it is a test image made by photographing through a common 10x eye loupe, commonly used for inspection of jewelry and small components.

   This showed the sand granules as the proper general size, but it also showed that they do not have the cohesion of Martian soil.
   Here the image has been converted to monochrome by using the "hue and saturation" image control and setting the saturation to 0%.

   The resulting image is a fairly good recreation of the image field and its look.  The lighting was intentionally left harsh, using a bare bulb with a simple foil back reflector to emulate the ambient illumination present from light reflecting from nearby rock outcrops.

   It was rather ironic that one of the limestone pieces turned out to be a terrestrial sand dollar that measured about 7 millimeters in diameter.  That is featured in the center of this image, nearly buried in the red "Mars sand."
   As a first effort, this is actually quite good.  It closely represents the sort of images we have been receiving from Opportunity.  It took almost no effort or equipment to reproduce the image quality, content, and look of the original NASA data.

   This means that it should be a simple matter to effectively duplicate the microscopic imager and how it records its images.

   This also means that if we apply the same image processing methods that I have used on the actual data, we should be able to determine their effectiveness and whether we are seeing real data or artifacts of some sort.
   The first step in my image processing methods would be to produce a sharpness enhancement using what is known as a difference method.

   This is done by creating a negative of the image and overlaying on the original image with a slight offset of a small number of pixels.  The result is that fine features are "doubled" in scale, yielding a sharpening or enhancing of the visual image.

   Note that this does not add more data to the image, but instead relies on how the human visual system interprets the data.  If this step is not performed carefully, it will end up adding more noise to the image and actually degrading the information content.

   Even now, image processing tends to be a "black art" that requires skill and experience to yield useful results.  Here is the slightly enhanced image that it yielded.

   What this illustrates is that it is a very simple task to create images that might pass as valid data.  This means that unscrupulous persons might use this method to create fake imagery or perpetrate a hoax, if they should desire.

   Fortunately, this is not an issue as all the image data comes directly from NASA/JPL and is therefore in no danger of being compromised.  Still, this is a cautionary result that should be borne in mind- not all images that might appear in publications are automatically to be accepted as valid, unless they can be verified by directly downloading them from NASA/JPL or Cornell themselves.  Images from other sources should be considered as suspect.

NEXT

BACK