Advanced Image Enhancement Techniques for Small Object Detection in Synthetic Aperture Radar (SAR) Imagery

Abstract

The product-based image enhancement model enhances the detection of small objects in unstable environments, with the seminal line 'Hermitian Product' delivering promising results. The proposed work utilizes Sub-Aperture Cross-Correlation Magnitude (SCM) to identify small ship targets in synthetic aperture radar (SAR) images of sea water. The “SCM” method employs various techniques such as Azimuth bandwidth reduction, sub-band extraction, up-sampling, Hermitian Inner product, low pass filter, down-sampling, and Local Averaging. I-SCM is analyzed using Coefficient of Variations (CV) and Spatial Resolution (SV), assessing target and background heterogeneity and sub aperture bandwidth influence. Experiments show that constant mean value and decreasing standard deviation improve target spotting in terms of 'CV', while partially over-lapped sub apertures provide best contrast in 'SV' relationships.