Multilayer perceptron classifier and shoreline extraction routine for Minnesota Point PlanetScope satellite imagery

A site-specific multilayer perceptron model was developed to classify PlanetScope satellite imagery of Minnesota Point, and the classifier was paired with a shoreline delineation routine to extract the position of the land-water interface. Satellite-dervived shorelines were used to measure changes in beach width associated with beach nourishment projects and changing water levels. The model is based on methods documented in Doherty, et al., 2022 (https://doi.org/10.1016/j.envsoft.2022.105512). This model archive includes the source code, the site-specific classifier, the labeled imagery used to train said classifier, and a subset of the input data needed to run the model. The input satellite imagery required to run the model are not included because the End User License Agreement (EULA) for PlanetScope imagery provided through the Commercial Smallsat Data Acquisition program prohibits making the imagery publicly available. The satellite-derived shorelines produced using this model are provided in a separate data release (Roland & Groten, 2024; https://doi:10.5066/P1W7SBUX).