Electrical utility companies spend millions each year to keep their transmission corridors free of encroachments. Major threats to transmission structures and wires include vegetation growth (especially fast-growing pines in the South) and unlicensed construction projects (such as swimming pools, garages, and other annexing structures). Traditional methods for encroachment detection can be summarized by the practices of one electrical utility serving a rural area in the South. This utility dispatches a small plane to patrol about 2000 miles of transmission corridors varying from 30 to 200 feet in width. The patrol is repeated every 2 months and cost about $500 per coverage for the aircraft alone. No aerial photographs are taken; data are captured by video or visual observation. Additionally, the utility hires a helicopter every six to eight years to patrol the lines at a cost of $500 per hour for as long as five full days. The expense, limited temporal resolution, and lack of a permanent record of the information captured using these traditional methods have led some utilities to explore remote sensing technologies for detecting encroachments.
      Mississippi Power Company, an investor-owned utility operating in 23 counties in southern Mississippi, recently invested in AM/FM/GIS software to keep track of its transmission and distribution systems, surrounding land use/land cover, and encroachments on easements. The company is currently incorporating data layers by deriving information about its 13,000-square-mile operating area from a variety of sources, including field surveys, company maps, U.S. Geological Survey (USGS) quad sheets, and National Aerial Photography Program aerial photography. In addition, Project Engineer Lou Occhi decided to investigate the usefulness of satellite imagery as a source of land use/land cover information and, more significantly, as a means of identifying encroachments on transmission corridors in a reliable, timely, and repeatable fashion.
      To learn more about satellite imagery and the incorporation of raster data sets into a vector-based GIS, Mississippi Power contacted the NASA Commercial Remote Sensing Program (CRSP) Office at nearby Stennis Space Center. The CRSP offers a Visiting Investigator Program (VIP) for American companies like Mississippi Power that are interested in remote sensing but are not certain of its possible benefits and do not know how to explore its applications. The VIP offers these companies a low-cost opportunity to explore remote sensing/GIS applications so they can make informed decisions on this technology's usefulness toward their profitability. Typically a VIP project is a demonstration set up by CRSP personnel working with commercial partners in an informal, over-the-shoulder manner.
      For Mississippi Power, project participants set up a prototypical remote sensing/GIS database for a 2.5-square-mile transmission corridor near Laurel, Miss. Arc/Info 7.0 served as the GIS software for the data, and ERDAS Imagine 8.1 was used as the image processing tool. Imagery was collected by NASA's Airborne Terrestrial Applications Sensor (ATLAS) sensor and Zeiss camera mounted on a LearJet and flown on Sept. 13, 1994 (Figures 3a and 3b). The 15-channel (visible to far infrared) ATLAS imagery was collected at 2.5-meter resolution, and the color infrared (CIR) aerial photography was captured at 1:8000 scale (1-meter resolution after scanning).
      Both data sets exhibit finer spatial resolution than any currently available U.S. commercial satellite imagery (a situation that will change in the late 90s). Through resampling, the data sets provided a general idea on how upcoming space-based commercial remote sensors may serve to detect tree and structure encroachments on transmission corridors as narrow as 100 feet. Table 1 summarizes the findings of the resolution comparison when identifying easement encroachments, such as tree clumps over 20 feet high or construction within a 200-foot-wide corridor.
      To continue demonstrating how remotely sensed imagery can be integrated into a vector GIS, CRSP and Mississippi Power personnel conducted a series of supervised and unsupervised classifications on the ATLAS imagery to discern between various land covers. The thematic raster information was then vectorized into polygons and transferred to Arc/Info, where it was displayed and shaded in ArcPlot through menu systems built with Arc Macro Language. This automated analysis was compared to a visual analysis (through heads-up digitizing) of the scanned and georeferenced aerial photos. These comparisons demonstrated to Mississippi Power personnel the speed and effectiveness of automated information extraction versus visual/manual analysis and illustrated the time-consuming process of scanning, georeferencing, and mosaicking aerial photos.
      Other spatial data techniques were also demonstrated through the VIP project to help Mississippi Power personnel learn to perform basic image processing steps, GIS integration, and interpretation of results on their newly acquired data:
¥ Computing band ratios and red-green-blue displays to reveal certain characteristics in the landscape. One such technique involved the display of ATLAS channel 14 (10.2 to 11.2µm thermal channel) in red, a Normalized Difference Vegetation Index (ratio of the near infrared to visible bands) in green, and channel 2 (.52 to .6µm) in blue. This display provided an interesting juxtaposition of thermal properties against vegetative reflectance and helped differentiate among vegetation growth in the transmission corridor (Figure 2).
• Creating custom menu systems to call up coverages, images, and attribute data (Figure 1).
• Creating Triangulated Irregular Network (TIN) models from vector data sets.
• Obtaining digital line graph and digital elevation models (DEM) from the USGS using the Internet and integrating them into the GIS.
• Obtaining soil information, cadastral (land ownership) information, and other data layers from federal, state, and local governments.
• Moving spatial data sets from platform to platform (such as Unix-based Arc/Info to PC-based ArcView), from software to software (ERDAS Imagine to Arc/Info), from media to media (hard copy to digitized, tape to hard disk), and from format to format (raster to vector).
      Cost comparisons between traditional methods and simulated satellite imagery for this application are difficult because of the different data attributes of the two methods, as summarized in Table 2. Companies should consider these issues and weigh them according to company needs before conducting a cost/benefit analysis.
      The use of satellite imagery to detect easement encroachments on narrow transmission corridors will be possible within the next few years when a host of satellite-based high-resolution sensors comes on line for commercial use. Some important criteria for the data used in this application include the need for spatial resolution of at least 3 meters, multispectral (visible and near infrared) resolution, monthly temporal availability, and the ability to extract DEMs from the imagery. Utilities should be able to purchase the data in a size that conforms to the strip/path nature of transmissions systems instead of in a rectangular block of data more applicable to land-use studies. Users should also be able to select, inspect, and download the imagery with a minimum of hassle and phone calls.

About the Authors:
Richard Campanella is an RS/GIS analyst with Lockheed Stennis Operations in support of NASA's Commercial Remote Sensing Program at Stennis Space Center, Miss. Bruce Davis is a remote sensing and GIS applications specialist for NASA at John C. Stennis Space Center. He is currently program manager for the Visiting Investigator Program and for Applications under the Space Act Program. Louis Occhi is a senior engineer in the Distribution Engineering Department and project manager for the AM/FM program at Mississippi Power Company.

Back