An Efficient Method for Moving Multi-target Detection and Tracking

摘要

The Complete Orthogonal Decomposition (COD) is an efficient subspace-based algorithm, which is numerically stable and accurate. For the detection and tracking of moving multi-target with blind signal processing, a new efficient method of URV decomposition is afforded, using primary matrix transform. And based on it, a new way of ranking-revealing is provided. At the last part of the paper, the tank detection experiment is carried out. The result shows that the COD can reveal the subspace of sound signal clearly and timely, and the COD-ICA can detect and track moving multi-target effectively.