Landsat Digital Enhancements for Lineament Detection STEPHEN J. WALSH Oklahoma State University Department of Geography Stillwater, Oklahoma 74078 and F. MYNAR, II Lockheed Engineering, Inc Las Vegas, Nevada 89109 ABSTRACT / Identification of lineaments has important ramifi- cations in geology because lineaments can signify zones of both hazardous potential and economically valuable environ- ments The synoptic view of Iocal and regional lineament patterns by Landsat is a useful mapping technique in areas considered to be well mapped as well as in poorly mapped areas Image enhancement of digitally constructed Landsat images increases contrast and sharpness between geologic features and improves the recognition of subtle differences Five enhancement techniques are applied to Landsat digital data for lineament detection (1) mean value of all four bands, (2) principal components, (3) band ratio, (4) histo- gram equalization, and (5) high-pass digital filtering Of the five enhancement techniques evaluated, the principal com- ponents analysis identified the greatest number of lineaments and the greatest total length of the lineaments. All five tech- niques identified a significant amount of unique lineaments, which were not found by any other technique. Unique linea- ments identified by each technique are combined through a composite process yielding a lineament interpretation which exceeds the detection capability of the principal components interpretation Introduction Remote sensing techniques have become invaluable tools for use in the investigation of a wide variety of problems affecting the physical and cultural environ- ments. One application of remote sensing technology is the identification, via satellite data collection and analysis, of geologic/geomorphic features known as lineaments. The term lineament, as used in this article, conforms to the definition by O'Leary and others (1976) as "a mappable, simple or composite linear feature of a sur- face, whose parts are aligned in a rectangular or slightly curvilinear relationship and which differs dis- tinctly from the patterns of adjacent features and pre- sumably reflects a subsurface phenomenon." The detection of lineaments through remote sensing techniques has been a problem of considerable interest to geologists and geographers since the incep- tion of aerial photography. Not until the launch of the first Landsat satellite in 1972 was the true worth of remote sensing technology for lineament detection fully realized. Upon interpretation of the first Landsat images, lineaments were readily visible, a condition that can be attributed to the oblique illumination, sup- pression of distracting details, and regional coverage characteristic of Landsat imagery. Goetz and Rowan (1981) remarked that one of the most striking results of Landsat image analysis is the discovery of nu- merous, previously unmapped regional linear features in both well-mapped and poorly mapped areas. Rowan and Wetlaufer (1973)agreed that, while Landsat images provide useful information to several geologic applications, lineament analysis is benefited particularly, because such features can be efficiently delineated by the satellite's synoptic view. Environ Geol Water Sci Vot 8, No 3, 123-128 The identification of lineaments including faults, joints, fractures, topographic crests, tithologic contacts, and bedding terraces are important in environmental geology for a number of reasons. Lineaments can rep- resent faults and fracture zones of both potential hazards and economically important environments. Lineaments can, for example, identify zones of seismic risk and areas of subsidence caused by the rapid with- drawal of groundwater migration and concentration and identify settings of ore deposits. Information about certain types of oil and gas deposits are com- monly expressed at the surface through lineaments. Specific arrangements of lineaments indicate direct ev- idence of folds and faults. Lineaments are also indi- cators of gross lithologic changes used in the detection of hydrocarbon reservoirs. Indirect clues to under- lying hydrocarbon structures are revealed by morpho- logical anomalies directed by lineaments indicating such phenomena as deflected drainage or offset hills. Finally, identification of lineaments can also provide insight into landforms, land use, and geobotanical pa- rameters useful for structural and tectonic investiga- tions (Saunders and Hicks 1976). This article assesses the detection and discrimina- tion of lineaments through the application of com- puter-enhanced Landsat-satellite digital data. The Landsat multispectral scanner (MSS) detects radiation in four different regions of the electromagnetic spec- trum, through the visible and into the near infrared wavelengths. The four spectral regions sensed by the MSS are 0.5-0.6 ~x (band 4), 0.6-0.7 tx (band 5), 0.7-0.8 tx (band 6), and 0.8-1.1 ~x (band 7). Each band senses distinct properties or attributes of phe- nomena because of its distinct spectral sensitivity. The objectives of this research are twofold: (1) to evaluate the unique and redundant lineaments identi- @ 1986 Springer-Verlag New York Inc