Discrimination of small targets in sea clutter using a hybrid CNN-LSTM network

Reliable and robust radar target detection is a challenging task in the maritime environment, particularly as bright returns or sea spikes share many similarities with true targets and can be difficult to distinguish. One approach to overcome this challenge is to exploit the persistence of the target and consider detection as a basic classification problem, i.e. either a target is present or absent. In this work, discrimination of small fluctuating targets in sea clutter is demonstrated using a hybrid convolutional neural network with long short-term memory (CNN-LSTM). The proposed approach is able to distinguish targets by exploiting amplitude fluctuations over time, with improved classification observed as longer sequences of data are processed.