Automatic Clustering for Precision Reconnaissance and Surveillance

Abstract

In this paper, we propose a coarse-to-fine approach-based automatic clustering algorithm for precision reconnaissance and surveillance. In order to improve the strengths and compensate for the weaknesses of the basic sequential algorithm and K-means clustering algorithm (the first one is simple to implement but may generate significant noise, whereas the second one returns precise results although the number of clusters needs to be known in advance), the proposed algorithm was developed by fusing them. It was validated and verified in a MATLAB simulation and embedded processing. As a result, it can be used to shrink the data effectively in a large dataset.