Difference between revisions of "First-Hand:Lab for Agricultural Remote Sensing"
Revision as of 12:47, 10 July 2012
In 1966 as a grad student at Purdue I worked for Dr. David Landgrebe in the Laboratory for Agricultural Remote Sensing (LARS) Lab. The main project I recall involved flying a multispectral scanner over crops and using reflected spectral information to distinguish types of vegetation and particularly diseased vs. healthy crops. The LARS team included Ag subject matter people as well as the spectrum analysis and transform math skills of Dr. Landgrebe and co-workers. Algorithms were developed for detection of plant types as well as diseased plant classifications. The scanner data was recorded on board a light plane and returned to the LARS lab where it was digitized and fed to an IBM 360/44 mainframe. The data was on analog instrumentation tape in FM format, fed from the recorder electronics to an A/D converter, and then to 9 track digital tape for post-flight processing. In those days good plotters were not available to us and computer-driven graphics CRTs were in the development stages. Thus, impact printers were used as plotting devices. The long dimension of the paper was the aircraft path over the field and the width of the paper was the scan width provided by the scanner. Different letters were used for crop types (e.g. C for corn or similar) and other characters were used for diseased corn. The end result was a visual representation of the field showing patches of disease among healthy areas, etc. Viewed from several feet away it was much like a gray-scale image.
My role was to build a pattern recognition circuit to detect the start of a lateral scan and a counter to allow selection of how many detected scans to use. Because of the processing limitations of the 360/44, it was not practical to process all the scanner data. The front panel selection on my box was to decimate the scans by a selected fraction, such as “every 10th scan” or “every 8th scan” being passed to the digital tape. This process was still in use when I returned on recruiting trips for the Naval Air Test Center several years later. I suppose I should note that another grad student at a bench behind me was building a mini-computer, PDP-8 vintage, out of Digital Equipment Corp (DEC) discrete component cards. Another student in our peer group was able to achieve lasing on an optical workbench and was able to demonstrate a laser which took up about 20 feet of total bench length.
I have to wonder if grad students today are having the kind of fun we had in those days.
Bill Rymer, 50 year member, IEEE.