Moving Object Detection by Track Analysis
Chenshan Wang · Calhoun: The Naval Postgraduate School Institutional Archive (Naval Postgraduate School) · 1990
The purpose of this thesis is to study the Hough transform method, as applied to the detection of tracks of underwater moving objects in Lofargrams. The subjects included are the Hough transformation, clustering study, and reconstruction. Two methods, LAS (Land Analysis System) cluster technique and Sorting, are used for cluster analysis. Encouraging results are obtained from the Sorting method. A further improvement of the Sorting is shown to yield better results in processing noisy track data. Experimental results dealing with noise free artificial data, noisy artificial data, and real noisy data are presented. The improved Sorting technique presented in the thesis has shown improvements compared to the straight forward Sorting when it is applied to spectral component tracking.