A shot-noise-limited image profiling technique is developed whereby th
e images of special classes of objects that fall substantially below t
he conventionally defined diffraction limit can be characterized with
respect to their shape, size, orientation, and reflectance. This is ac
complished via the point-by-point subtraction of a point-source refere
nce profile from the unknown object profile, the resulting difference
profile providing a ''fingerprint'' of the unknown object. Computer si
mulations of this process have been Validated by experiments conducted
with 8-channel and 64-channel profiting instruments. The latter instr
ument performs hundreds of profile scans per second: autocentering, do
wnloading, and integrating the scans so as to provide shot-noise-limit
ed profiles in near-real time. Experiments conducted to date have demo
nstrated the characterization of objects subtending angles as small as
0.07 lambda/D, where lambda is the dominant wavelength of radiation a
nd D is the diametric aperture of the optical instrument used to achie
ve that resolution. It is predicted that an instrument similar to the
64-channel system, when used in conjunction with a l-m telescope in an
exoatmospheric environment, will make it possible to characterize obj
ects that are significantly less than a meter in size from distances i
n excess of 10,000 km. Meanwhile, its application to microscopy likely
will make possible the in vivo study of viruses. (C) 1997 Society of
Photo-Optical Instrumentation Engineers.