THE REMOTE IDENTIFICATION OF TERRAIN FEATURES AND MATERIALS AT A VIRGINIA TEST SITE: AN INVESTIGATIVE STUDY OF MULTISPECTRAL SENSING TECHNIQUES

This report documents the collection, processing, and analysis of multispectral remote sensor data from a 2- by 13-mile study site in Augusta County, Virginia. The objectives of this study were to develop and test, using analog processing facilities, techniques for the identification and automatic discrimination of selected terrain features and natural materials. Aircraft data were collected in up to 15 spectral bands at two times of the day during two different seasons. A ratio image-enhancement procedure and multi-object discrimination techniques were tested. Each of these techniques is described and the results illustrated. Results of this research demonstrate that level slicing (quantizing) selected ratios of two spectral bands from the visible and near-IR wavelength regions can be used to discriminate areas of surface water, bare soils, deciduaous vegetation, new crop vegetation, and roads and buildings. Natural soil drainage and soil texture information were obtained from a ratio of two thermal infrared bands. Likelihood ratio discrimination, based on representative training sets, separated the scene into nine different terrain categories-five of which included surface soil differences. Attempts to relate soil spectral differences to differences in the depth and composition and underlying geological materials were not successful.

  • Corporate Authors:

    Environmental Research Institute of Michigan

    P.O. Box 618
    Ann Arbor, MI  USA  48107
  • Authors:
    • Wagner, T W
    • Hasell Jr, P G
  • Publication Date: 1974-1

Media Info

  • Pagination: 77 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00263663
  • Record Type: Publication
  • Source Agency: Federal Highway Administration
  • Report/Paper Numbers: FHWA-RD-74- 10 Intrm Rpt., FCP 34E1-024
  • Contract Numbers: DOT-FH-11-7136
  • Files: TRIS, USDOT
  • Created Date: Dec 19 1976 12:00AM