Abstract:

In extended target detection, the feature square matching detector utilizes the prior envelope modulus of the target's one-dimensional range profile to enhance the signal energy accumulation effect. This approach alleviates, to some extent, the drop-off loss problem that traditional energy detectors experience when the scatterer distribution is sparse. However, obtaining the prior envelope modulus of the target’s one-dimensional range profile under all attitude angles is challenging. Additionally, if the target's attitude angle is unknown or there is estimation error, leading to inaccurate prior information, the detection performance may sharply decline. Therefore, there is a need for prior templates with strong angular adaptability. This paper proposes a prior knowledge-assisted extended target detection method. The Clean algorithm is used to extract the location and amplitude information of strong scatterers in the prior one-dimensional range profile. A prior auxiliary template is constructed using a rectangular window combination and then squared matching is performed with the observed data. The use of prior information improves detection performance and enhances angular adaptability. Simulation experiments and measured data indicate that under low signal-to-noise ratio conditions, when achieving a detection probability of 0.8, the proposed method can adapt to a noise power increase of approximately 4.5 dB and exhibits good angular adaptability compared to energy detectors. This method can be used to extract prior information to assist detection and improve detection performance.

Key words: prior information, one-dimensional range profile, Clean algorithm, extended target detection