Assessment of bioplastics degradation using near-infrared hyperspectral imaging and computer vision techniques

Monitoring the degradation of bioplastics under real-world conditions remains challenging due to the complex chemical, physical, and microbial processes involved. In this study, the biodegradation of cassava ( Manihot esculenta ) starch–cellulose films containing oregano ( Origanum vulgare ) essential oil (OEO) under different soil conditions were evaluated using near-infrared hyperspectral imaging (NIR-HSI) combined with chemometric analysis. Biodegradation assays were conducted over 24 days using soils of two pH levels (5.5 and 7.6), and time series hyperspectral images (TS-HSI) were acquired every two days. Multivariate Curve Resolution (MCR) enabled the identification of spectral features associated with starch, cellulose, essential oil, moisture, and degradation products, and allowed reconstruction of chemical maps depicting their spatial evolution. ANOVA–Simultaneous Component Analysis (ASCA) revealed that degradation time was the dominant factor affecting the spectral response, followed by film formulation and soil pH. The combined approach successfully captured chemical and structural transformations, including polymer breakdown, moisture uptake, and microbial activity. These results demonstrate that NIR-HSI is a powerful, non-destructive tool for monitoring the biodegradation of starch-based bioplastics and provides valuable insights for optimizing formulations and environmental applications.