Professional: DIRSIG

In 1983 I was still trying to make connections and break RIT into the remote sensing research business. To that end I took a summer fellowship with the Naval Research Lab in Washington DC. They asked me to model the contrast between ships and the sea as it would be observed from space with some future thermal infrared sensor. I spent the summer assembling and running radiation propagation models for what the radiated and reflected energy from a ship propagated to space over a range of atmospheres and view angles would look like. After a similar effort to model what water would look like the difference was the contrast or signature the Navy wanted. I wrapped things up in August and the Navy was quite happy. I went back to RIT and started talking to my research group about the “right way” to study the problem including all the ship to ship and ship to water interactions that I had neglected in the summer study. This would take extensive geometric computer models coupled to much more involved radiation propagation models to study properly. I convinced the Navy this was important and that we had oversimplified the problem in the summer study. To prove it we built an “infrared theater” in the laboratory with toy boats in a fish tank surrounded by an antifreeze chilled “sky” to simulate the cold infrared sky. The result, while primitive, compared favorably to early versions of our computer models. They demonstrated that the ship-water interactions were indeed very important and often produced a much stronger signature than what the simple modeling I had done over the summer suggested. As a result, I convinced the Navy and later a number of sponsors to support a program to build much better geometric models coupled to radiation propagation models that would simulate in the computer what an airborne or spaced based sensor would “see”. The first student to work this project called his very simple code The Digital Imaging and Remote Sensing (DIRS) Image Generation (DIRSIG) model (DIRS being the name of my research group). Over the last 35+ years the model has grown dramatically in sophistication, power and the range of sensors it can simulate. RIT has become a world resource for image simulation and DIRSIG is now used by both the defense/intelligence community and the civil remote sensing community to model new sensors and to support development of new analytical tools. My last few years at RIT were spent adding sensor models into DIRSIG to demonstrate to NASA the power of modeling to support new sensor design and then working with them to support design trade-off studies for the next generation of Landsat sensor. This section provides just a few fun looks at the DIRSIG story. DIRSIG continues to be a major tool for the government and aerospace industry and a major component of ongoing work at DIRS. You can find out more at the DIRSIG website by clicking here.