1978 · 47 citations · 0 references
AgricultureAgronomyGeoinformaticsCrop ManagementCrop SensingRemote SensingAgricultural Land ManagementCrop ClassificationLand Cover ClassificationLand Cover Pattern AnalysisRemote Sensing For Crop ManagementZone DelineationArea EstimationClassification AccuracyClassification Accuracy AnalysisClassification Accuracy AssessmentImage Classification For CropsRemote Sensing Crop ClassificationSystematic SamplingSystematic Spatial Sampling
The study for which the results are presented had these objectives: (1) to use Landsat data and computer-implemented pattern recognition to classify the major crops from regions encompassing different climates, soils, and crops; (2) to estimate crop areas for counties and states by using crop identification data obtained from the Landsat identifications; and (3) to evaluate the accuracy, precision, and timeliness of crop area estimates obtained from Landsat data. The paper describes the method of developing the training statistics and evaluating the classification accuracy. Landsat MSS data were adequate to accurately identify wheat in Kansas; corn and soybean estimates for Indiana were less accurate. Systematic sampling of entire counties made possible by computer classification methods resulted in very precise area estimates at county, district, and state levels.