Engineering PapersSearch

Engineering topics

Register, D. T.

Publications and source records attributed to Register, D. T..

Crop identification studies using Landsat data Separation of barley from other spring small grains and corn and soybean decision logic

Two labeling procedures were developed which identify various agricultural crops through the use of Landsat data. One procedure separates barley from other spring small grains, and the other identifies corn and soybeans. For both procedures, a minimum data set (critical acquisition time) has been designated. Landsat data in both image format and various graphic displays were used along with ancillary data to obtain information which aided in labeling the spectral signatures. The corn and soybean procedure also employed a structured decision logic. Test results for the barley separation procedure emphasized the importance of obtaining a critical acquisition and showed some success especially in areas where spring crops followed the expected growth patterns. Two tests of the corn and soybean procedure produced good labeling accuracies. Problems with the procedure were easy to identify, and some solutions were implemented for the second test. Automation of various parts of the procedure and extension to other crops and regions were recommended.

Dailey, C. L.

On evaluating clustering procedures for use in classification

The problem of evaluating clustering algorithms and their respective computer programs for use in a preprocessing step for classification is addressed. In clustering for classification the probability of correct classification is suggested as the ultimate measure of accuracy on training data. A means of implementing this criterion and a measure of cluster purity are discussed. Examples are given. A procedure for cluster labeling that is based on cluster purity and sample size is presented.

Pore, M. D.

Evaluation of classification procedures for estimating wheat acreage in Kansas

This report presents the results of experiments which were performed to evaluate procedures for estimating wheat acreage in intensive test sites (ITS's) in Kansas. An analyst/interpreter (AI) selected and labeled fields from Landsat-1 satellite imagery. Statistics were generated for each selected ITS, and the imagery was classified using a maximum likelihood classifier. Various components of the classification process were tested.

Flores, L. M.