Data Processing I: Advancements in Machine Analysis of Multispectral Data

Research in multispectral data processing at LARS/Purdue is directed at supporting a substantial level of applications research as well as advancing the technology of remote sensing data processing. During the past year significant progress has bean made in both respects. Almost the entire multispectral data analysis process, from data editing to results evaluation, has been impacted, and the new level of technology has been vigorously tested by the data analysis operation associated with the 1971 Corn Blight Watch Experiment. The following discussion of these advancements is organized to follow generally the steps utilized in the multispectral data analysis procedure. In terms of Figure 1, we begin with the data display process used to accomplish data editing and proceed clockwise through clustering, statistics computation, etc. In the interest of brevity, each result will be treated here in a general way and references given to available sources where a more detailed treatment may be found