Coastal Remote Sensing Investigations Volume 2: Beach Environment

Imagerv obtained from the tour channel ERIM radar system was evaluated tor soil survey and field type identification purposes. The area imaged was snow covered to a maximum depth of approximately 1 meter. An lev snow surface, high moisture content, or ice layers within the snowpack may have been the cause of attenuation of the X-HH signal, creating a lack of tonal variations. The other three channels exhibited tonality patterns which varied between and within channels and agricultural fields. These tonality patterns were not related to geologic or topographic Influences, but low correlations with the soils map of ehe area were observed. Vegetation at these locations appeared to be the cause of these correlations. Grey scale values of the seven dominant field types encountered in the study area indicated that discrimination is possible using tonal signatures from all four channels of this SAR system. The cross-comparison of each channel's response value may separate fields with very similar averaged reflectance values. More ground truth and more effective use of ancillary data are needed to improve the interpretation procedure. A more quantitative method of the tonal measurement is also needed and combined with a measure of texture would greatly enhance the discriminative ability of radar systems. The combination of these improvements would make possible the digital classification of SAR data by computer. Thus multi-temporal and multispectral radar Imagery, In combination with the effective use of ancillary data, have the potential to discriminate crops and soils for agricultural purposes. RESUME Des images de terrain ennelgi, obtenuea avec un svstarne radar de quaere bandes, ont ete evaluees en vues d'utilisation pour 1'lnventalre des sols alnsl que pour la reconnaissance des caractirlstiques das surfaces agrlcoles. Un« surface glacic, une teneur en eau ilevie, ou des couches de glace i 1'Interieur de la neige ont pu causer 1'attenuation du signal Presented to the 3th Canadian Symposium on Remote Sensing, Victoria, August 1978. X-HH creant alnsl un manque de variations dans la tonallte. Dans Les crois autres canaux les tonalites des couleurs variaient entre les bandes et suivant le type de champs agricole. La tonallte des couleurs n'apparalt pas Stre fonctlon de facteurs giolcglques ou copographiques et les correlations avec le relev* des sols pour la region sont pauvres. Ceci pourralt Stre du aux Influences du couvert vegetal. Les valeurs obtenues dans I'echelle des gris pour les sept types de champs indiquent qu'une distinction est possible suivant Les oaracteristiques tonalss de chacune des quatres bandes. £r comparant les valeurs obtenues pour cheque bände, des champs avec dec reflectances "moyennes" cres semblable peuvent etre separis . Ce procedä de discrinination pourralt etre amcliore par une tiesure plus quantitative du ton et une meilleure reconnaissance sur le terrain. Ce procede pourralt amener une classification automatlque par ordlnateur des donnses obtenues avec un systems SAR. Une teile imagerle au radar multl-cemporelle et multi-spectrale, utilisee effectivement avec des donnees auxlliaires peut permettre d'identifier les recoltes et les sols agrlcoles. INTRODUCTION Tonal variations observed on radar images of the earth's surface depict the ability of the ground targets to reflect electromagnetic radiation incident on the imaged targets. The proportion of Incident radiation returned to the radar sensor, i.e., backscatter, depends on dielectrical and geometrical properties of the ground elements. Various research efforts in the past have attributed the tone differences observed on radar imagery to characteristics c£ snow cover (Waits and MacDonald, 1970; Vickers and Rosa, 1972), surface soil conditions (Morain and Campbell, 1974; Cihlar et aL, 1975), and crop type (Harallck et al., 1970; Bush and Ulaby, 1975). Radar is a sensor of great potential to many kinds of studies in agronomy, but more research is needed on the interactions of the earth's surface and the radar waves to develop effective interpretation procedures. With the advent of synthetic aperture radar systems (SAR), finer resolution Imagery has provided more detailed representations of the earth's surface. Dual polarization, multi-frequency radar systems are the microwave equivalents of mltispectral scanners, which have proven useful for many remote sensing applications (Harallck et al., 1970). The purpose of this study was to evaluate four channel SAR imagery, of snow covered " terrain, for Information pertaining to soil survey and cover type discrimination.