🚀 go-pugleaf

An invitation for input to a GeoWorld review of image processing
software

In a forthcoming issue of GeoWorld we will be printing a comparative
review of the six leading commercial remote-sensing image-processing
packages: ERDAS Imagine, ENVI, Intergraph Imager, PCI, ER-Mapper and
TNT-MIPS. The overall approach for the review is provided in the
following note "Evaluating image processing software." 

I would like to invite your input, as users, to the review. While I am
interested in any comment you may have, I am particularly interested in
(1) information on characteristics of the software that have only become
apparent as you have used it more extensively and (2) your opinion as to
the best "niche" for the software - to what area(s) does it most provide
value/capabilities? Any other comments are also appreciated. 

It would be very helpful if you could also provide a thumbnail
describing how you use the system(s). Are you an occasional user or
production oriented? Does your work involve vegetation mapping, geology
or something else? What types of data are you using - etc. 

I may use some of your comments in the article so please let me know if
you DO or DO NOT want your comments printed. If you don't say I WILL NOT
use them.

Thanks in advance for your help.


BTW - If you are aware of other list serves or groups that should be
included please forward that information to me. I have been unable to
identify a specific list serve or group for ENVI or TNT-MIPS.
------

Evaluating image processing software 

Many software evaluations involve creation of a list of features to
which each package is compared. The package with the longest list is the
winner.  That approach doesn't consider how the real world works. If we
applied the same logic to automobiles we would all be driving James
Bond's latest reincarnation of the Aston Martin because it has the most
cool gizmos. In the real world, some of us want pick-up trucks, some
SUVs, some six passenger sedans and a few, a very few, need Aston
Martins - with a Bond girl in the passenger seat. The giant review list
also leads to bloated software - as developers respond to "gotcha lists"
with their, often, unrealistic demands. Outside of the clean pages of a
magazine review all software doesn't fall into a single neat ordering
from best to worst. There can be turkeys out there as well as stars but
most software is designed with a particular target in mind. It probably
does that thing really well. This is particularly true in the arena of
remotely-sensed image processing software. Image processing has moved
from being a skill practiced by a few specialists at NASA, some
universities and large corporations, to something that can, potentially,
be found on any desktop. There are many different types of sensor
systems, from (now) routine Landsat to hyperspectral systems, radar and
others. There are many different domains where image processing is used.
From enhancing black and white photography, to mapping infrastructure,
to forest inventory, to geo-chemical analyses. Finally there are many
different levels of use, from someone who uses a system once in a while
to prepare imagery to serve as a backdrop to a GIS analysis, to a
researcher who is trying to find a better algorithm to identify a
specific crop, to a company that processes thousands of images in a
production setting. Another key factor is the size and complexity of the
data that will be processed and the requirements for the image
processing system to inter-operate with other systems. The real problem
today is not to identify the "best" image processing system but to find
the best fit between your needs and the available systems. Helping the
readers of GeoWorld do that is the goal of this review. 

-- 
W. Fredrick Limp,   Director and Professor       FAX: (501) 575-5218
CAST, Center for Advanced Spatial Technologies   TEL: (501) 575-6159     
12 Ozark Hall, U of Ark., Fayetteville AR 72701   fred@cast.uark.edu 
Opinions expressed here are mine, at least I think they are.