*1.1. Global Change*

Coastal landscapes have faced significant changes over billions of years, and their evolution is concomitant with major climatic upheavals. In its sixth and most recent report, the Intergovernmental Panel on Climate Change (IPCC) indicates that climate change is occurring more rapidly than originally predicted, with unprecedented increases in sea levels, heat waves, and the faster melting of polar ice caps [1]. Currently, mankind is trying

**Citation:** James, D.; Collin, A.; Mury, A.; Qin, R. Satellite–Derived Topography and Morphometry for VHR Coastal Habitat Mapping: The Pleiades–1 Tri–Stereo Enhancement. *Remote Sens.* **2022**, *14*, 219. https:// doi.org/10.3390/rs14010219

Academic Editors: Dar Roberts, Junshi Xia and Simona Niculescu

Received: 31 October 2021 Accepted: 23 December 2021 Published: 4 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

to cope with and adapt to rapid climate changes that influence ocean currents, winds, precipitations, temperatures, and strongly re–shaped landscapes [2].

#### *1.2. Landuse/Landcover Observation Techniques*

Observation techniques for tracking landscape changes are many and varied. At local–scale and very–high (VH) spatial resolution, unmanned aerial vehicles (UAV) are useful for the VH temporal resolution monitoring of coastal socio–ecosystems [3]. UAVs are cost–efficient and easily deployable for shoreline detection [4] and for the identification of seasonal variations in saltmarsh meadows [5]. They are, however, not well suited for monitoring areas at the landscape scale (several km2) due to not only the restrictions that are imposed by legislation but also the technical limitations that are enforced by the number of flight times that are permitted by the battery capacity. In addition, in coastal areas, the meteorological and marine conditions require a maximum time of presence on the site due to the tides (±one hour after low water slack).

Manned aerial vehicles (MAV) serve as a robust alternative that leverages passive sensors with a basic red–green–blue (RGB) spectrum and sometimes the infrared spectrum [6] or an active light detection and ranging (LiDAR) sensor [7]. The complete solution makes it difficult to plan missions, and sensors such as LiDAR are rather expensive.

#### *1.3. Spaceborne Acquisition and Stereoscopy*

Yet, the analysis of the environment at the landscape scale is made possible by satellites which have a VH spatial, a multispectral, and even a hyperspectral resolution for the best– equipped satellites [8].

A spaceborne solution exists to obtain multispectral VH resolution images that are 0.50 m and 0.30 m pixels in size, which are provided by Pleiades–1 or WorldView–3 and 4, respectively. Satellite–based multispectral VH resolution mapping of the coastal fringe has been successfully performed in studies focusing on tropical [9] or temperate environments [10].

Since 2000, some remote sensing satellites that are specialized in stereo acquisition have been launched into orbit around the Earth, such as the Worldview–1, –2, –3, –4 constellations, GeoEye–1 and –2, the Pleiades–1 and, –1B constellations, and the 2021– launched Pleiades Neo, whose images are not yet available for research at the time of this submission [11]. The operating principle of stereoscopy is to photograph an object or a landscape from two different angles in the same way as human vision is able to, with a specific overlap for determining the 3D information of the obtained images. Sometimes a third angle (nadir) of view can be available as a redundant observation to increase the accuracy when producing a digital elevation or surface model.

Satellite–based stereo topography has the capability of improving coastal mapping by improving the spectral discrimination of eco–geo–morphological objects [12]. However, when a tri–stereo product is used, do they augment/boost this coastal mapping? Do the morphometric parameters that are derived from the topography contribute to a better classification of coastal ecosystems than basic spectral information? This paper along with the experimental results that are presented in it will seek to answer these two questions.

#### **2. Materials and Methods**

#### *2.1. The Study Site*

The entire study site (76 km<sup>2</sup> terrestrial part) is located on the Emerald Coast in Brittany (France) along the Channel Sea (48.60◦ N, 2.00◦ W; Figure 1). It is characterized by a diversity of ecosystems that are shaped by the proximity of a megatidal sea and is one of the six areas with the highest tidal ranges in the world (up to 14 m) [13]. The Rance, a coastal river, ends its course in the bay of Saint Malo, dividing the area in two sub–sites. In terms of land cover, this study area is composed of temperate zone coastal vegetation, salt marshes, rocks, dunes, and fine sand beaches. The coastline is strongly indented, leaving multiple sandy beaches surrounded by rocky points and islets a little further offshore. Bays are also common on this coastal fringe and are featured by the presence of salt marsh meadows. In terms of urbanism, this Brittany coastal fringe is subject to strong anthropic pressures. The small fishing villages of the past have evolved into resort urban areas that now attract tourists in search of iodized air and marine landscapes.

**Figure 1.** Location of the study on the Emerald Coast (France).

As with all coastal areas in the world, the Emerald Coast is not exempted from coastal risks. Its highly populated coastline increases the vulnerability of lowland populations.

#### *2.2. Pleiades–1 Satellite Imageries*

The Pleiades–1A and 1B constellation multispectral satellites were launched on 16 December 2011 and 2 December 2012, respectively [14]. The Pleiades–1 constellation acquires images of the Earth daily and can cover up to 1,000,000 km2 per day. The radiometric spectrum of the sensor extends from 430 nm to 940 nm (B: 430–550 nm; G: 500–620 nm; R: 590–710 nm; and NIR: 740–940 nm).

Data collection is based on tri–stereo images from the Pleiades–1 satellite sensor (Table 1; Figure 2a). The satellite orbited over the study area on 28 November 2020 to collect three images at 11 h 26 min 14 s (UTC), then at 11 h 26 min 24 s (UTC), and finally at 11 h 26 min 32 s (UTC; Table 1). Each dataset of images contains panchromatic and multi–spectral images (R, G, B, NIR) that are 0.5 m and 2 m pixels in size, respectively (Table 1, Figure 2b,c). The images were delivered without initial geometric processing (primary level) and without radiometric processing.

**Table 1.** Pleiades–1 specificities of the tri–stereo acquisitions over the study site.


**Figure 2.** Viewing directions with respect to the target area of the tri–stereo Pleiades–1 images (**a**); natural–coloured nadiral imagery at 2 m pixel size (**b**), and panchromatic nadiral imagery at 0.5 m pixel size (**c**) of the study site.
