%T 3-D Crater Analysis of LDEF Impact Features from Stereo Imagery
%A Clyde A. Sapp
%A Thomas H. See
%A Michael E. Zolensky
%B 69 Months in Space - Second LDEF Post-Retrieval Symposium
%R NASA CP-3194, Part 2
%D April, 1993
%P 339-345
%E Arlene S. Levine
%I Langley Research Center Hampton, VA 23681-0001
%U http://setas-www.larc.nasa.gov/setas/PUBS/LDEF/cp3194-93-p339.ps.Z
%X <H5>SUMMARY</H5>
We report here preliminary results from attempts to derive depth and diameter 
information from digitized stereo images of impact features on 
<A HREF="http://setas-www.larc.nasa.gov/setas/ldef.html">LDEF</A>. Contrary to our prior 
assumption, we find that impact craters in the T6 Al alloy are not paraboloid in cross section, 
but rather are better described by a 6th-order polynomial curve.  We explore the implications of 
this discovery.<P>
<H5>INTRODUCTION</H5>
In expectation of the <A HREF="http://setas-www.larc.nasa.gov/setas/ldef.html">LDEF</A> return, 
the requirement for a system to analyze the hypervelocity impact craters on the space-exposed 
surfaces of the spacecraft was determined.  Ideally, 
this analysis system would be able to define in three dimensions the surface structure of each 
crater to a high degree of precision.  As a minimum, the system should be able to determine the 
true depth and diameter of each crater.  The 'true' depth is defined as the deepest point in the 
crater as measured from the level of the ambient surface, and the 'true' diameter is the inside 
diameter of the crater when measured at the level of the ambient surface 
(see Figure 1).<P>
A number of constraints were placed upon this system design.  The budgetary limitations 
were fairly severe, and the time frame for technique investigations was short.  It was essential 
that the analysis system use a technique that was non-destructive and remote (i.e., no contact 
with the material surface permitted).  In addition, the system must use a technique that could be 
incorporated into a portable system to be used at 
<A HREF="http://www.ksc.nasa.gov/ksc.html">Kennedy Space Center</A> during the deintegration 
of the <A HREF="http://setas-www.larc.nasa.gov/setas/ldef.html">LDEF spacecraft</A>. <P>
It was decided to use binocular imagery to analyze the crater morphologies.  It was fairly 
inexpensive to achieve, and made use of some existing hardware to collect the information.  A 
portable system configuration consisted of a portable PC equipped with a color video digitizing 
board and a color video multiplexer, a binocular microscope, a pair of video cameras, and a pair of 
optical disk drives with removable media.  This system configuration would collect pairs of color 
digital images and store them to the optical media for later analysis.  It was also decided to write 
software that would automatically register the image pairs on a pixel by pixel basis using a 
traditional cross-correlation technique.   The parallax information in each pixel registration would 
provide depth data for each pixel, and thereby provide a full three-dimensional representation of 
the crater surface. <P>
During the three month deintegration of 
<A HREF="http://setas-www.larc.nasa.gov/setas/ldef.html">LDEF</A>, the
<A HREF="http://www-sn.jsc.nasa.gov/curator/seh/ldef/m&d_fact.htm">Meteoroid and Debris Special 
Investigation Group (M&D SIG)</A> generated approximately 5000 digital color stereo image pairs 
of impact-related features from all space-exposed surfaces.  An earlier paper (1) describes the 
theory and practice of determining this 3-dimensional feature information 
from stereo imagery.<P>