Akademik Veri Yönetim Sistemi
Advances in Space Research, cilt.77, sa.9, ss.8956-8981, 2026 (SCI-Expanded, Scopus)
Pansharpening, a super-resolution imaging technique, involves the fusion of high-resolution panchromatic images with lower-resolution multispectral images to generate a high-resolution pansharpened image. This technique offers significant advantages, including simplified sensor design, efficient utilization of communication bandwidth, and conservation of onboard energy in unmanned aerial vehicles. Pansharpening is particularly crucial in aerial and aerospace imaging applications. It mitigates payload capacity and energy constraints in multi-sensor imaging systems, facilitates efficient processing of large-scale datasets in remote sensing platforms, and minimizes telemetric transmission costs in satellite-based Earth observation missions. However, current pansharpening algorithms often fail to fully preserve the spatial details of panchromatic images and the spectral integrity of multispectral images, resulting in spatial degradation and spectral distortions. This necessitates the development of advanced methodological approaches. The Detail-preserving pansharpening (DPAN) method introduced herein is a novel high-frequency detail injection algorithm that preserves detail during pansharpening. A key characteristic of DPAN is its simultaneous preservation of both spatial and spectral information, achieved through the optimization of structural parameters via Colony-Based Search Algorithm. In experimental validation, the performance of DPAN was benchmarked against 17 classical pansharpening methodologies and 3 deep neural network-based methods. Its capacity to preserve spatial and spectral information was objectively assessed using 9 image quality metrics. Statistical analysis confirmed that DPAN achieves superior performance compared to most of the conventional and deep learning-based pansharpening methods.