Identification of plastic debris on beaches by ground-based hyperspectral imaging

The presence of plastic debris in coastal areas represents a worldwide critical issue due to its negative impact on the ecosystems, with potential consequences on wildlife and human health. The use of hyperspectral sensors can help to monitor plastic pollution in coastal and marine environment. In this work a ground-based hyperspectral imaging (HSI) system was tested to identify and classify plastic debris directly at beach sites. Hyperspectral images of different plastic objects placed on a sandy beach (Torre Guaceto, southern Italy) were acquired in the near infrared range (NIR: 1000-1700 nm). A classification strategy was built based on partial least square-discriminant analysis (PLS-DA). The proposed ground-based HSI approach shows promising results with many advantages, being fast and accurate, allowing to characterize plastic debris along the coastlines.