IDENTIFICATION OF CROPLAND AND AREA ESTIMATION FROM AERIAL PHOTOGRAPHY AND SATELLITE IMAGERY

Abstract

Listen

Translate article

The feasibility of using multiband aerial photographs and ERTS-1 imagery for manual field identification and measurements is presented for a Moderately Cold Subhumid Soil Climatic region of Saskatchewan. Using a simple six-step scale criterion, numeric values were assigned to the grey-level densities on black and white photographs (August 1972) for the infrared and red radiation bands. The grey-level densities were calibrated to known crops or other features, determined from data taken on the ground from training sites. A classification key was developed from the calibrated grey-scale values for the two spectral bands. Using the classification key, > 90% of the cereal crops — harvested, cereal crops not harvested, rapeseed, and fallow land — were identified on areas adjacent to the training site. It was not possible in this study to distinguish between wheat and barley from August photography. Nearby sites were used to verify the criteria. For nonnumeric analyses, infrared color was useful to substantiate the results. In general, the green spectral reflection appeared to be similar to the red for most of the fields except for a few of the fallow fields and grassland. However, further studies are required to evaluate the place of the green band in crop identification. The transference of known physiographic features from topographic survey maps and high-altitude aerial photographs onto ERTS imagery facilitated establishment of field boundaries and area measurements. By comparing the values of unknown fields to those calibrated with known features on the same photograph, variation in exposure, photographic development, and reproduction are eliminated.