This paper describes a multiresolution Landsat TM and AWiFS sensor assessment conducted by the USDA's National Agricultural Statistics Service (NASS). NASS’ Remote Sensing Acreage Estimation Program has used Landsat TM and ETM+ data successfully for years for crop acreage estimation at the state and county levels. However, NASS is currently faced with a reduction in the supply of available imagery due to a malfunction in the Landsat 7 sensor. Consequently, NASS has begun evaluating alternative sources of imagery. The goal of this investigation was to assess the suitability of AWiFS data for crop acreage estimation. Three crop land classifications of Nebraska (2004) were produced including a multitemporal TM classification, a unitemporal TM classification and a unitemporal AWiFS classification. Once the classifications were completed, classifier statistics for corn, soybean and overall crops were evaluated for accuracy. Strata regression (r-square) statistics were derived by crop type for all classifications. State level estimates for corn and soybeans using ground survey data alone were compared to estimates from the remote sensing classifications. The AWiFS classification results were less accurate than the results from either Landsat classification. Reductions in classification accuracy can be attributed to the moderate spatial resolution (56 meters) and the reduced number of spectral bands of AWiFS relative to the TM sensor. However, it is anticipated that the increased temporal frequency of AWiFS (5 day) versus the TM sensor (16 day) should improve classification accuracy. Furthermore, AWiFS data appears acceptable for acreage estimation over large, spectrally homogenous, crop areas.
[1]
Charles Day,et al.
A Compilation of PEDITOR Estimation Formulas
,
2002
.
[2]
Russell G. Congalton,et al.
Assessing the accuracy of remotely sensed data : principles and practices
,
1998
.
[3]
J. Stakenborg,et al.
The Effects Of The Application Of Smoothing And Orthogonal Transforms To Spot And Tm Data On Regression Based Crop Acreage Estimates
,
1990,
10th Annual International Symposium on Geoscience and Remote Sensing.
[4]
Martin Ozga,et al.
EVALUATION OF CLUSTERING TECHNIQUES FOR CROP AREA ESTIMATIONUSING REMOTELY SENSED DATA
,
1991
.
[5]
M. Craig,et al.
PEDITOR - STATISTICALIMAGEANALYSISFOR AGRICULTURE
,
1995
.
[6]
J. R. Jensen.
Remote Sensing of the Environment: An Earth Resource Perspective
,
2000
.