| description:
|
<DIV STYLE="text-align:Left;"><DIV><P><SPAN>We used autoRIFT (Lei et al., 2021; Gardner et al., 2018), an open-source software package, to track dune migration between 2000 and 2023 using the Landsat panchromatic image catalog over the Riyadh and Mecca provinces. AutoRIFT is a feature-tracking algorithm that measures and geocodes inter-image displacements in optical or radar remote sensing imagery. It detects pixel offsets between two images captured at different times using the normalized cross-correlation (NCC) method (Gardner et al., 2018)—a statistical similarity measure that identifies optimal pixel shifts by comparing reference and template image patches. The NCC process evaluates the correlation between pixel intensity patterns and identifies the displacement corresponding to the peak correlation value using a Gaussian pyramid-based oversampling technique. The method employs a nested grid approach with progressively larger chip sizes to optimize resolution and signal-to-noise ratio. This sparse-to-dense search strategy enables the exclusion of low-coherence regions (Lei et al., 2021).</SPAN></P><P><SPAN>To implement this methodology in tracking dune migration over the Mecca and Riyadh provinces, we selected cloud-free Landsat image pairs (2000–2023) from similar dates and seasons to minimize illumination differences due to shadows, topography, and sun angle elevation. Landsat scenes from the same date, path, and UTM zone were mosaicked and processed to avoid reprojection-related offsets and ensure a seamless mosaic. The generated offset images in the X and Y directions for each image pair were merged to produce a consistent displacement mosaic for the entire study area.</SPAN></P><P STYLE="margin:0 0 11 0;"><SPAN STYLE="font-size:12pt">Lei, Y., Gardner, A. and Agram, P., 2021. Autonomous Repeat Image Feature Tracking (autoRIFT) and Its Application for Tracking Ice Displacement. Remote Sensing, 13(4), p.749.</SPAN></P><P STYLE="margin:0 0 11 0;"><SPAN /></P><P><SPAN /></P></DIV></DIV> |