214 
as the best to discriminate soybeans, com and 
sugarcane in the area. 
- Separability of J-M distance indicated that the 
statistical structure for crop discrimination was 
three-or four-dimensional. .Comparison of the 
classification matrices of the training areas using 
band combinations 2-4-5 and 2-4-5-7 showed similar 
results and a small difference in computer 
processing time. However, the smaller data storage 
volume of the three-band classification makes it the 
most appropriate for crop discrimination using 
digital analysis in this study area. 
REFERENCES 
Anuta, P.E., Bartolucci L.A., Dean, M.E., Lazano, 
D.F., Malaret E., McGillem C.D., 1984 Landsat-4 
MSS and Thematic Mapper data quality and 
information content analysis. TREE Trans, on 
Geosci. and Remote Sens. GE-22(3): 222-235. 
Atkinson P., Cushnie J.L. and Townshend J.R.G. 1985. 
Inproving Thematic Mapper land cover 
classification using filtered data. Int. J. Remote 
Sens. 6(6): 955-961. 
Salcmonson, V.V., Smith P.L., Pink A.B., Webb, W.C., 
and Lynch T.J. 1980. An overview of progress in 
the design and implementation of Landsat D systems 
I.E.E.E. Trans. Geosci. Remote Sensing 18:137- 
146. 
Sheffield C., 1985. Selecting band combinations 
frcm multispectral data. Photogram. Eng. Remote 
Sens. 51(6): 681-687. 
Swain, P.M., King, R.C. 1973. Two effective 
feature selection criterion for multispectral 
remote sensing. LARS information note 402673 West 
Lafayette, Purdue University. 
Townshend, J.R.G., Gayler, J.R., Hardy, J.R., 
Jackson M.J., and Baker, J.R. 1983. Preliminary 
analysis of Landsat-4 Thematic Mapper products. 
Int. J. Remote Sens. 4 (4): 817-828. 
Townshend J.R.G. 1984 Agricultural land-cover 
discrimination using thematic mapper spectral 
bands. Int. J. Remote Sens. 5(4): 681-698. 
Young, T.Y., Calvert, T.W. 1974. Classification 
estimation and pattern recognition. New York, 
Elsevier.