Abstract: In unknown underwater target detection and bearing estimation, the weak targets are covered by strong targets, which cannot be detected and estimated at the same time. A weak target detection method based on the subspace bearing stability is proposed to solve the problem. Firstly, this method makes eigen decomposition for covariance matrix of array element receiving data. Then, it utilizes the characteristic that the beam-forming bearing estimation values of targets subspace are stable and close to the true values, and the beam-forming bearing estimation values of background noise subspace are randomly distributed in ［0°,180°］. Lastly, the method statistically weights the normalized spatial spectra of each subspace by the bearing variance, so it can further restrain the spatial spectrum disturbances of background noise subspace, and can decrease the spatial spectrum difference between strong targets and weak targets subspace. The theoretical analysis, numerical simulation and experimental results show that, at the same frequency band and the same beam, compared with sub-band conventional beam-forming (SBCBF) detection method, our method can well enhance the spatial spectrum amplitude of weak target subspace, and decrease the spatial spectrum difference between strong targets and weak targets subspace, and detect strong targets and weak targets at the same time.

Key words: signal processing technology, weak target detection, eigen decomposition, bearing stability