Target Detection in Sea Clutter Based on Feature Re-Expression Using Spearman's Correlation

The target detection method based on multidimensional feature space is commonly used for detecting small maritime targets in sea clutter. However, existing methods do not fully explore and utilize the correlation information between features, limiting their detection performance. To address this limitation, this article proposes a target detection method in sea clutter based on feature re-expression using Spearman's correlation. First, a generalized linear grouping method based on Spearman's correlation is proposed to strengthen the linear relationships within groups while reducing intergroup feature correlations. Second, intragroup feature re-expression based on Bhattacharyya distance (BD) is introduced, yielding new features with superior classification capabilities and complementary informational relationships. Third, a tri-feature concave hull detection algorithm based on dual distance weights is designed to enhance detection performance while controlling the false alarm rate. Testing with measured data from three public datasets demonstrates that the proposed detection method exhibits significant improvements in detection performance compared to existing feature-based detection methods.