Enhancing Satellite Hyperspectral Imagery for Archaeological Prospection through Pansharpening Techniques

Core Concepts

Pansharpening techniques can effectively enhance the spatial resolution of PRISMA satellite hyperspectral data, making it highly suitable for identifying sub-surface archaeological features.

Abstract

This research evaluates the use of pansharpened PRISMA satellite hyperspectral data for archaeological prospection. Three pansharpening methods (GSA, MTF-GLP, and HySure) were compared quantitatively and qualitatively over the archaeological landscape of Aquileia, Italy.
The key findings are:

Pansharpening significantly improved the spatial resolution of PRISMA hyperspectral data, enabling better detection and interpretation of both large-scale features (e.g., palaeochannels) and small-scale structures (e.g., buried roads, channels).
The GSA and HySure methods outperformed MTF-GLP in terms of visual quality, providing higher contrast and better outlining of archaeological and geomorphological features.
While quantitative metrics like UIQI, SAM, and ERGAS correlated with the qualitative assessment for the June 2022 image, they did not accurately reflect the practical utility of the pansharpened products for the August 2023 image.
The study highlights the need for developing more robust evaluation metrics that can better capture the visual potential of pansharpened hyperspectral data for archaeological applications.
The availability of higher resolution PRISMA hyperspectral data will enable further exploration of other geo-archaeological features and the exploitation of information in the SWIR portion of the spectrum.

Stats

The area surrounding the ancient Roman city of Aquileia in northeastern Italy was used as the case study.
Two PRISMA products covering the same area, acquired on June 4, 2022 and August 24, 2023, were analyzed.

Quotes

"Pansharpening techniques can effectively be used to enhance the spatial resolution of multispectral and hyperspectral products."
"The results suggest that the application of pansharpening techniques makes hyperspectral satellite imagery highly suitable, under certain conditions, to the identification of sub-surface archaeological features of small and large size."

Key Insights Distilled From

Pansharpening of PRISMA products for archaeological prospection

by Gregory Sech... at arxiv.org 04-09-2024

https://arxiv.org/pdf/2404.05447.pdf

Deeper Inquiries

How can the pansharpening techniques be further improved to better capture the visual potential of the data for archaeological applications?

To enhance the visual potential of pansharpening techniques for archaeological applications, several improvements can be considered:

Contextual Adaptation: Developing pansharpening algorithms that can adapt to the specific characteristics of archaeological sites, such as varying soil types, vegetation cover, and historical features, can improve the accuracy of feature enhancement.
Feature-Specific Optimization: Tailoring pansharpening methods to focus on enhancing specific archaeological features, such as ancient roads, structures, or buried artifacts, can lead to more targeted and effective results.
Integration of Machine Learning: Incorporating machine learning algorithms to learn from ground-truth archaeological data can help optimize pansharpening parameters for different types of features and environments.
Multi-Sensor Fusion: Combining PRISMA hyperspectral data with other remote sensing data sources, such as LiDAR or radar imagery, can provide complementary information for a more comprehensive view of archaeological sites during the pansharpening process.

What are the limitations of the current quantitative metrics in evaluating the performance of pansharpening methods for archaeological prospection, and how can they be addressed?

The limitations of current quantitative metrics in evaluating pansharpening methods for archaeological prospection include:

Lack of Archaeological Relevance: Existing metrics may not directly correlate with the visual interpretability of archaeological features, leading to discrepancies between quantitative assessments and practical utility.
Spectral-Spatial Complexity: Traditional metrics may not capture the complex interactions between spectral and spatial information crucial for archaeological feature detection.
Subjectivity in Evaluation: Quantitative metrics may not account for the subjective nature of archaeological interpretation, where human expertise plays a significant role.
To address these limitations, new evaluation metrics can be developed that:
Incorporate Archaeological Criteria: Integrate archaeological criteria into quantitative assessments to ensure that the metrics align with the goals of archaeological prospection.
Hybrid Metrics: Develop hybrid metrics that combine spectral and spatial information in a way that reflects the specific requirements of archaeological feature detection.
User-Centric Evaluation: Include user-centric evaluation methods where archaeologists and remote sensing experts collaborate to assess the effectiveness of pansharpening techniques based on practical interpretability.

What other remote sensing data sources, in combination with PRISMA hyperspectral imagery, could be leveraged to enhance the detection and interpretation of a wider range of archaeological features?

In combination with PRISMA hyperspectral imagery, the following remote sensing data sources can be leveraged to enhance the detection and interpretation of a wider range of archaeological features:

LiDAR Data: Light Detection and Ranging (LiDAR) data can provide high-resolution elevation information, allowing for the identification of subtle topographic features and buried structures that may not be visible in optical imagery alone.
Thermal Infrared Imaging: Thermal infrared data can help detect variations in surface temperature, revealing subsurface features like buried walls or structures that retain heat differently from the surrounding environment.
Ground-Penetrating Radar (GPR): GPR data can penetrate the ground to detect buried archaeological remains, offering insights into subsurface features without the need for excavation.
Multispectral Satellite Imagery: Combining PRISMA hyperspectral data with multispectral satellite imagery, such as Sentinel-2 data, can provide additional spectral bands for vegetation analysis, land cover classification, and change detection, enhancing the overall understanding of archaeological landscapes.