Water column detection method based on Doppler modeling

This paper proposes a method for the automatic detection of water columns in the presence of sea clutter, with the objective of facilitating the identification of water column targets. Firstly, a motion model of the water column target is established in accordance with the characteristics of the water column target. Subsequently, simulations were conducted to assess the efficacy of water column target detection at varying distances within the context of sea clutter. The simulated water column target was subjected to echo signal modulation and pulse compression, enabling the extraction of target Doppler frequency shift information. Subsequently, the Fast Fourier Transform is employed to analyse the Doppler information of the water column target in terms of amplitude-frequency, thereby constructing amplitude-frequency diagrams. Ultimately, the amplitude-frequency diagrams are inputted into the convolutional neural network for training, with the objective of obtaining the feature information of the water column target and optimising the network parameters through iteration, thus completing the water column target detection process. The network parameters are optimized in order to complete the detection of water column targets.