An iterative clustering procedure

In many remote sensing applications, millions of measurements can be made from a satellite at one time, and many times the data is of marginal value. In these situations, clustering techniques might save much data transmission without loss of information since cluster codes may be transmitted instead of multi-dimensional data points. Data points within a cluster are highly similar so that interpretation of the cluster code can be meaningfully made on the basis of knowing what sort of data point is typical of those in the cluster. In this paper we introduce an iterative clustering technique. The procedure suboptimally minimizes the probability of differences between the binary reconstructions from the clusters codes and the original binary data. The iterative clustering technique was programmed for the GE-635, KANDIDATS (Kansas Digital Image Data System) system and tested on a multi-image data set. Twelve images of the northern part of Yellowstone Park were taken by the Michigan scanner system. The images were reduced and run with the program. Clustered into four clusters were 30,000 data points, each consisting of a binary vector of 25 components. The percentage difference between the components of the reconstructed binary data and the original binary data was 20.7 percent.