Remote Sensing for Agricultural Purposes

Abstract

The potential applications of remote sensing from aircraft in agriculture are many. Suggested uses include determination of crop acreages, surveys of land use, soil surveys, rangeland surveys, and determination of crop condition. The latter has been of special interest particularly with respect to crop damage from drought, salinity and plant pests. Quantitative measures of incidence and severity of plant diseases and other crop pests require high expenditures of time by trained personnel. Consequently, most estimates of crop losses from plant pests in major production regions of the United States and other countries are derived from inadequate surveys. In the late 1950's, U.S. industry members of the Agricultural Research Institute asked the Agricultural Board of the National Academy of SciencesNational Research Council to study methods by which the accuracy of such information could be improved. This request was guided by the reports of earlier work by R. N. Colwell (1956) on the detection of wheat stem rust disease in plots of graded infection severity by use of aerial infrared photography. In 1960, a committee of the NAS-NRC was appointed. In selecting committee members, the Academy took cognizance of the advances in technology that had been recently made in radiation detectors and optical mechanical scanners as well as photographic films. Members were selected from the disciplines of physics, engineering, agriculture, and forestry. From the outset, their studies dealt with not only the photographic spectral regions but also the ultra-violet, reflective infrared, emissive infrared, and radar regions. The committee was named Remote Sensing for Agricultural Purposes and was given permission to plan the research approach and to seek support for it. Because of the complexities of assembling equipment and obtaining agricultural test areas, the committee was later given permission to participate in the early feasibility studies. The first experiment was conducted in 1964 using University of Michigan aircraft and sensing equipment over Purdue University Agricultural Experiment Station farms (Holter et al., 1964). The major objective was to determine the capability of recognizing crop species by comparing energy returns from the different crops in 18 rather narrow spectral bands ranging from 0.35 to 16 /. The sensing instruments were a combination of optical mechanical scanners with appropriate filters and detectors and multiple and single lens camera systems. This combination of instruments was the best that could be put together at the time, but of course, there was no way to calibrate the energies measured in each channel. Data were recorded on film. Analyses were limited largely to comparing photographic tones of different fields with each other and with a selected standard area such as a road, pond, or grass strip common to each scene. Despite these and many other limitations, the power of such a multispectral system to differentiate crop species and soil types and to detect weedy and damaged areas within fields was indicated (Hoffer et al., 1966; Holmes & Hoffer, 1966). In 1965, the University of Michigan improved the airborne sensing system by mounting all equipment in a single plane, and devising a single-aperture, 12-channel scanner with internal calibration. The channels were selected