[Table of Contents]


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[PADG:1609] RE: Microfilming Connecticut aerial photographs



Jane,

I've been out of the preservation microfilming business here at the
University of Florida now for three years and can't point you to a specific
article or document on the topic.  However, I can address the following:

***
Consider inviting a GIS Aerial Image specialist, perhaps from the University
of Connecticut, to consult.

***
Consider applying for an IMLS grant to evaluate the microfilm product for
subsequent digitization, rectification, and use in GIS applications.
The product should satisfy future use as well as preservation.

1 -- The vendor shall ...
***
Continuous tone is (or at least used to) have no direct duplicating negative
(that type of film just wasn't made).  The second generation and third
generation films were both positive.  
Preservation Resources can confirm this or correct my recollection.

2 -- For the master negative (first-generation), the resolving power ...
***
"Bigger is better" particularly with photographic materials and more so for
aerials.  You might consider revising your specification to, at least,
*prefer* greater line pairs.  Many of the current microfilming
specifications were based on hardware limitations since exceeded.  
For aerial photographs, I'd want to shoot for somewhere between 150 and 175
line pairs **at minimum**.  Herrmann and Kraemer cameras have the capacity
to capture slightly more that 190 line pairs.
***
In order to plan for use as well as preservation, determine the altitude at
which the source photograph was imaged then see USGS Map Accuracy Standards
(http://mac.usgs.gov/mac/isb/pubs/factsheets/fs17199.html) -- the document
does not translate its ground accuracy requirement to resolving power, but
it you'll see that the margin of error is pretty small.
Cf, also the **slightly* more relaxed Canadian/British Columbia standard:
http://www.for.gov.bc.ca/ric/Pubs/Aquatic/AerialVideo/AerialVid-05.htm#P432_
47173
I much prefer the British "good practice"
http://ads.ahds.ac.uk/project/goodguides/apandrs/, which does a very good
overall job of discussing a given use (digitizing for archaeology uses).

2 -- ... The resolution will be determined by finding the lowest pattern
resolved on any of the 5 test charts in the frame at the beginning and at
the end of the reel.
***
Consider, for aerial photographs, that resolution should be even (or within
minimal tolerance) across the surface of a frame comparable in size and
shape to the source aerials at each corner and center.
***
THIS IS CRUCIAL for aerial photography particularly if the images will later
be digitized for use in GIS applications.
USGS accuracy standards state that 90% of all points must be tested.
Users of aerial photographs highly value photogrammetric correctness.
Inability to adequately resolve these five areas consistently (i.e., each to
the same level within a frame) is detrimental.  The resulting images will
not be photogrammetrically correct.  Using these images, it would then be
difficult to geo-rectify.
***
For purposes of subsequent use (digitized for GIS applications), consider
requiring that the image frame be, at least, slightly larger than the size
of the source.
* Most photographic curvature (see below) occurs toward the edges of the
film.
* If this can be achieved by means other than increasing reduction ratio,
prefer it.
* Require for each microfilmed aerial that -- I don't know if I can do this
without an illustration :: see
http://web.uflib.ufl.edu/digital/temporary/aerial.jpg (size of elements
included in frame exaggerated for illustration of placement) -- that the
test targets be included in frame with the image.  Require consistent
placement of within the frame to facilitate selection of the aerial
photograph when it is later digitized.
* Require that a percentage (preferably consistent with USGS standard, i.e.,
90%) of these in frame targets we resolved. Keep documentation for later
use.  The variation between targets will assist in future correction and
subsequent photogrammetric evaluation.
***
Consider other factors that might make photogrammetry difficult:
* Curvature of the source photograph :: require that all photographs be held
flat on a level plane perpendicular to the camera head while being filmed,
preferably with the use of vacuum table rather than by glass which may
introduce distortion.
* Emulsion :: address the placement of lights are to produce even
illumination ... you should find this in the RLG guidelines as well.  RLG
requires placement of a solid white target ... the density of this target on
the master negative (or whatever you may later use to digitize) is measured
for density.  All corners and the center should be measured (and for aerials
the density readings should all be the same or within minimal tolerance of
one another).  RLG recommends, I believe, a non-reflective target.  For
aerials, however, I would recommend a solid white target with similar
reflectance characteristics/gloss as that of the aerials.
* Position of the film :: prefer use of a camera (e.g., Zeutschel) that
sucks the film flat against the lens to reduce distortion resulting from
curvature of the film.
* Ensure that introductory target sequence includes information about every
aspect of filming: camera, reduction, film stock, etc...  this information
will be useful when digitizing for GIS applications.
*** 
You should consider aerials to be more similar to works of art than to
photographs.  
A reproduction that represents the source aerial as faithfully and as
exactly as possible should be required.

3 -- Background density ...
***
see http://www.asprs.org/asprs/resources/standards/photography.htm section
3.5.3 for required of newly born aerial photographs and attempt compliance.

3 -- ... No less than five readings shall be taken per roll, at the
beginning, middle and end of the roll...
***
Consider more readings to better ensure accuracy.



Erich J. Kesse
Director, Digital Library Center - George A. Smathers Libraries
University of Florida, Gainesville, FL   32611-7007
email: kesse@xxxxxxx    tel: (1)352-846-0129   fax: (1)352.846-3702





[Subject index] [Index for current month] [Table of Contents]