Closeout Report, NAG5-6650

March 24, 1999

Center of Environmental Remote Sensing Education for Engineers and Scientists at Michigan Tech

PRINCIPLE INVESTIGATOR: Gregg J.S. Bluth, located at Michigan Technological University

Project Overview

Justification and Objectives. This proposal concept, like the Remote Sensing Institute (RSI) itself, grew from our conclusion that traditional programs do not lead students into the interdisciplinary training they need for remote sensing careers. Michigan Tech is a large engineering and science university with many talented undergraduates, and this initiative is aimed at preparing them for environmental remote sensing careers. Powerful remote sensing techniques have been developed in the past decade which facilitate the study of environmental processes, and allow data from multiple instruments to be processed and merged. However, this type of research is computer intensive, and the ability to analyze and disseminate large data sets as derived from remote sensing instruments depends on computing ability. The Remote Sensing Institute is dedicated to education and research in environmental remote sensing. Thus, the acquisition of computer equipment is central to the development of the RSI, which in turn is the catalyst for integration of diverse research projects hosted by Michigan Tech.

This proposal was to provide the computer infrastructure to educate undergraduates in the basic skills needed for data fusion and visualization, and to prepare graduate students for remote sensing research aimed at important environmental problems. The proposal asked for support to equip new computer laboratories, primarily located in a new building dedicated to environmental monitoring, with equipment specifically chosen for remote sensing.

This proposal was granted approximately $400K from NASA, which was surpassed with $430K matching funds from Michigan Tech, to help us toward the objectives of building a campus-wide institute in remote sensing education and research. The following sections will detail: (1) development of the computer laboratories; (2) usage of the non-computing field equipment; (3) development of the RSI; (4) creation of a Minor program in Remote Sensing. The summary section comprises a "before and after" description pertaining to the COE grant.

(1) Creation of Computer Laboratories/Classrooms.

The actual implementation of the research equipment proceeded as scheduled in the original proposal, with installation and full usage commencing in January 1999. There were a number of modifications to the original proposal, which reflect the following: a $70K reduction in the original requested amount from NASA; physical constraints placed by the new building infrastructure; our ability to strike bargains with Sun Microsystems. One of the original PC-based classrooms in the Dow building was converted to a Sun workstation lab, but all workstations are equipped with PC-emulation software. Overall, our original goals were exceeded, primarily due to the competive nature of the computer business and the working relationships established by our computer systems administrator (co-PI Mike Dolan).

The computer hardware funds were used to populate the Dow building with high-end graphics workstations, as well as powerful data and application servers. Almost all of the equipment purchased was from Sun Microsystems. MTU has a long outstanding history with this company, and by leveraging a larger campus purchase we were able to obtain incredible one-time discounts on all the hardware purchased.

PC applications are run through the Insignia SoftWindows emulator, and we are expected to outfit most of the lab seats with a Sun manufactured PC-coprocessor card capable of running all PC applications in their native format on the Ultra10 and Ultra60 workstations.

Table 1. Computer Classroom/Laboratories (708 and 709 Dow)

These two student labs are designed for instruction in image analysis and processing, innovative remote sensing software packages, GIS, computer modeling techniques, and data analysis and processing. Outside of scheduled class labs, the rooms remain open for undergraduate computing.

Table 2. Server Room (709B Dow)

Table 3. Imaging and Analysis Lab (312 Dow)

This high end lab is dedicated to complex computational tasks such as manipulation of large data sets, advanced image processing, and data fusion tasks.

Table 4. System/Research Administrator (302 Dow)

Table 5. Flex Graduate/Research machines (currently 308 and 309 Dow)

These machines and their placements are designed to accomodate non-permanent graduate or research needs and are not dedicated to any particular area.

Table 6. Printing (including 708, 709, 312 Dow, as well as dedicated alcoves)

These machines are strategically positioned through the Dow building's computing facilities.

Table 7. Multimedia laboratory (513 Dow)

Table 8. Forestry: The Remote Sensing/GIS Teaching Lab

The School of Forestry and Wood Products provided funds for the renovation of the room housing the lab. This included the installation of air conditioning, painting, installation of a white board, computer furniture for 16 workstations and network connections. The School also provided funds for the purchase of software which includes ERDAS Imagine and the complete suite of ArcInfo and ArcView GIS. The School plans to add 6 more workstations to the lab this summer to bring the lab to capacity seating (15 student and 1 teaching workstation).

(2) Non-Computer Equipment: Inventory and Usage

Software. The establishment of Unix workstation classrooms has opened the door to many exceptional software deals. Some were obained with the promise of using the classrooms for software training seminars, other major deals involved campus-wide licenses. Below is a listing of the major successes in this area, with dollar amounts if available. Some have a enormous impact on our research, for example, many of our data processing and merging routines, and image processing algorithms are IDL-based.

Table 9. Software Deals for RSI Laboratories

Atmospheric Studies. As proposed, the two radiometers and associated data acquisition system are mounted on a rooftop on the MTU campus. In addition, each is connected to the world wide web via a Unix workstation. This radiometer observatory has supplied data from August 5th, 1998 until the present. Although there have been setbacks (the data acquisition system was shorted out by rain in August, 1998 and had to be repaired), the radiometers continue to function as expected. The radiometer web site is still evolving, and work to modify a Celestron 8 telescope mount for use as a solar tracker is nearly complete.

In addition to the work which is underway and has already been accomplished, plans are being made to use at least one of these radiometers for research on the Greenland icecap in the summers of 1999 and 2000. The transmission of visible and UV light through snow will be measured using the SPUV-6 radiometer. This information is needed to study the photolysis of chemicals in the snow, and thus to understand how atmospheric trace gases are transferred to the Greenland icecap.

Table 10. Equipment Purchased: Atmospheric Studies

Aquatic Studies of Lake Superior. The Satlantic optical profiler and reference radiometers were deployed throughout the first field season of the Keweenaw Interdisciplinary Transport Experiment in Superior (KITES). These instruments will be employed during all 5 years of KITES to determine the underwater light field in the lake, and follow spatial, seasonal, and interannual variability in optical properties of the lake.

The radiometer data are being used for three primary applications: (i) direct measurement of the in-water light field to track photochemical transformations of light-absorbing material; (ii) comparison to satellite data for ground-truthing of chlorophyll and suspended particle algorithms

in this coastal fresh water system; (iii) comparison to data from an airborne hyperspectral sensor for high resolution (spatial and spectral) tracking of water masses and currents. Our first year's data demonstrate that changes in light absorption in the lake are coupled to the onset of stratification, apparently due to photoreactions in the surface water.

Table 11. Equipment Purchased: Aquatic Studies

Hyperspectral Studies. An enormous advantage was made of the non-computing equipment, in supporting the hyperspectral flight missions. We could not have done them, or arrange to pre-process all the 100's of GB without the COE equipment. Some of the equipment is also being utilized to design a new, tomographic hyperspectral sensor which simultaneous records both spatial dimensions, and spectral data on a single 2-D CCD. The equipment was used in conjunction with two visits by Kestrel corporation with their Cessna 206 and visible hyperspectral sensor whose optical system co-PI Rafert helped design, to perform the first two ever hyperspectral surveys of Michigan's Upper Peninsula.

Table 12. Equipment Purchased: Hyperspectral Studies

(3) Growth of the Remote Sensing Institute

The original director of RSI, Bruce Rafert (Physics, co-PI) served the initial year (1997-1998) and was followed by Charles Kerfoot (Biology) for the current year (1998-1999). The plan is to have one more interim director before performing a national search for a permanent head of the Institute in the year 2000. Faculty membership in RSI has grown from the initial dozen to 32 current members, spread amongst the eight member departments. The number of non-faculty researchers actively involved with the RSI laboratories varies (see Table 13), but has roughly doubled over the past year.

Research. Below is a listing of faculty (department) and number of research students actively involved with the RSI laboratories, both in Forestry and Dow buildings.

Table 13. Student Research Use of RSI laboratories (1998-1999 Academic Year)

Education and Outreach. The classrooms were still under construction until October 1998, but once completed they have come under heavy use as undergraduate facilities; likewise, use of these as classrooms is growing. Below is a list of courses which have made use of these computer classrooms.

Table 14. Courses using the RSI Labs (1998-1999 Academic Year)

Two courses of note include the RSI seminar series, and an outreach program to local high schools students:

UN401 Remote Sensing seminar Rose/Kerfoot 42 students

This unique seminar series began in Fall 1998, and included speakers from the full spectrum of the RSI membership: graduate students, post-doctoral resreachers, and Michigan Tech Faculty. Many fine outside speakers also participated in the seminar series, most notably Nobel laureate Dr. Paul Crutzen in Fall 1998.

Outreach: "Remote Sensing Introduction (for high school students)", reaching approximately 200 regional high school students per year; taught by co-PI Budd.

(4) Development of the Remote Sensing Minor

Minors have previously been unknown at Michigan Tech, but all majors have at least some elective choices. We have proposed an interdisciplinary minor in remote sensing, compatible with all member departments in the Remote Sensing Institute. The cornerstone class is the junior-level introductory course in remote sensing (PH330), with lectures and labs taught by four RSI members. The minor consists of eight quarter courses (which would convert to six semester courses during the upcoming curricular changes) in addition to the RSI seminar series.

The Senate has approved minors at Michigan Tech for departments and institutes, as a direct result of the RSI efforts. This required many revisions and negotiations, but the inititiation of the minor with its ability to attract students from many major disciplines represents one of the main educational goals of the Institute. However, the remote sensing minor itself has not been approved, and this process continues with the goal of attaining the minor program for the University conversion to semesters in Fall 2000.

Summary

Before COE: Although Michigan Tech's efforts in remote sensing was strong (approximately 10-15 faculty), facilities and resources were limited and no infrastructure existed to facilitate colaboration outside of an individual department. The four main centers of remote sensing research were housed in Biology, Forestry, Geology and Physics, with labs composed of 2-6 workstations apiece, purchased and maintained individually. Likewise, software licenses were purchased and maintained individually, thus data and image transfer between researchers was nearly impossible. Instruction in remote sensing techniques was difficult because there were no computer labs large enough to accomodate more than a few students at a time, being limited both in seats and in software licenses. Fundamentals of remote sensing education largely hinged on the Forestry department, which took the responsibility of basic remote sensing theory through applications. Lack of support equipment precluded the use of field measurement studies to related to remotely sensed data. Seminars in remote sensing were not uncommon, but occurred sporadically and were rarely advertised outside a single department.

After COE: During the period of this proposal, the Remote Sensing Institute is experienceing rapid growth both in personnel (now 31 faculty strong) and in resources. The remote sensing facilities are outstanding: we have two 25-seat classrooms and a 6-seat advanced research lab, all staffed with high end Sun workstations and HP printers, a 10-seat PC lab in Forestry; a multimedia laboratory for plotting, printing and scanning, and and 4 roving Sun workstations for short-term stations. These facilities have allowed use to generate software deals, providing campus-wide licenses for IDL, Imagine, Arc products. The field radiometers and camera equipment enhance our aquatic and atmospheric spectral studies using satellite sensors, and our partnership with the Kestrel Corporation in continuing to develop and apply hyperspectral technology to environmental research.

Since their construction in the Fall of 1998, the RSI laboratories are in use by approximately 40 students from six departments, ranging from undergraduate to post-doctoral reserarchers. The classrooms have been host to over 20 separate courses this year, reaching over 450 students. Outreach has gone to the high school community at a pace to bring 200 students annually into the Michigan Tech classrooms.

Students in remote sensing students now have a clear educational pathway, beginning with our fundamental remote sensing course, taught in Physics and the RSI laboratories, which provides solid grounding in optics, electromagnetics, and fundamentals of data processing and imaging. Regular interaction amongst faculty and students is now facilitated by the remote sensing seminar series, which in this year was highlighted by the visit of Nobel laureate Paul Crutzen. Through RSI efforts minor programs are now allowed at Michigan Tech, and our objective to establish a Remote Sensing Minor is being proposed to the University senate.

The Institute is far from complete. The main issues to be settled involve equitable usage of resources by all members, long term maintenance and upgrades, and development of RSI, through both University and external grants, as a viable research and educational unit. However, a great deal of progress has been made in the past 2 years, most notably of course in terms of computing and field resources. The core of the Institute is composed of highly active and dedicated faculty who have worked very hard to make the Remote Sensing Institute a reality - the resources provided by NASA and the Dow Corporation provide us with a solid foundation upon which to continue to grow and collaborate.