Machine Processing of Remotely Sensed Data for Change Detection

Remote sensing can furnish a quantitative method for detecting change in earth surface features. LANDSAT-1 MSS data was temporally overlayed and geometrically corrected. Data from one date was then classified and a LARSYS results tape produced. Data from the second date was then similarly processed. Classifications for the same geographical area were then available for two (or more) dates. These classifications were then compared on a point by point basis by the computer to determine the type and location Of changes occurring for the intervening period between data sets. Two applications of this post-classification change detection technique were made. The first was used to study fluctuations of the snowpack in the Rocky Mountains of Colorado. The purpose of this application was to accurately determine the rate of diminution of the snow pack and thus provide information for predicting flow rates. The second application was to study land use changes in Marion County, Indiana to provide information to land use planners