Year: 2026
Type: article
Abstract: The purpose of this study was to address the persistent challenge of clutter in marine radar systems to improve target detection and classification accuracy. A Convolutional Neural Network (CNN), was developed and trained to process radar signals, enabling it to effectively distinguish real targets from unwanted noise such as sea clutter. The methodology involved using a 3Ghz S-Band Marine Radar Data samples containing both Clutter and Target Echo Signal samples from vessels as Radar Signal represented by a sinusoidal waveform, introducing random noise to simulate realistic clutter conditions, and then processing the resulting signal through multiple CNN Convolution layers. The Model was optimized using a Mean Squared Error (MSE)-based loss function to Suppress clutter while preserving meaningful target information. The performance of the proposed system was evaluated using Clutter Suppression Ratio (CSR), Signal-to-Clutter Ratio (SCR), f1-Score, and Classification probability.
Institution: Rivers State University
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7236707494
